Citrix virtual desktop handbook (5 x)

Citrix
Virtual Desktop Handbook
Version 5.x
An architect’s guide to desktop virtualization
Prepared by:
Worldwide Consulting Solutions

Citrix Virtual Desktop Handbook
TABLE OF CONTENTS
Section 1: Overview.................................................................................................... 5
Introduction ................................................................................................................ 4
Methodology............................................................................................................... 5
Assess...................................................................................................................... 5
Design ...................................................................................................................... 6
Deploy ...................................................................................................................... 6
Monitor ..................................................................................................................... 7
Project Plan ................................................................................................................ 8
Section 2: Assess ....................................................................................................... 4
Overview ................................................................................................................... 10
Business Drivers...................................................................................................... 10
Data Capture............................................................................................................. 11
Strategies............................................................................................................... 11
Manual................................................................................................................. 11
Survey ................................................................................................................. 12
Automated........................................................................................................... 12
Best Practices..................................................................................................... 13
User Data Capture................................................................................................. 14
Business Characteristics .................................................................................. 14
Technical Characteristics.................................................................................. 15
Application Data Capture ..................................................................................... 17
Technical Characteristics.................................................................................. 17
Capabilities Assessment ...................................................................................... 18
Best Practices..................................................................................................... 19
User Segmentation................................................................................................... 20
User Groups .......................................................................................................... 20
FlexCast ................................................................................................................. 20
FlexCast Model Comparison ............................................................................. 21
FlexCast Model Selection .................................................................................. 23
Best Practices..................................................................................................... 25
Application Assessment.......................................................................................... 26

Rationalization....................................................................................................... 26
Business Characteristics ..................................................................................... 27
Best Practices........................................................................................................ 29
Project Management ................................................................................................ 30
Roadmap................................................................................................................ 30
Project Team.......................................................................................................... 31
Business Roles................................................................................................... 31
Technical Roles.................................................................................................. 34
Section 3: Design...................................................................................................... 38
Overview ................................................................................................................... 39
User Layer................................................................................................................. 40
Endpoint Selection................................................................................................ 40
Decision: Endpoint Ownership ......................................................................... 41
Decision: Endpoint Lifecycle ............................................................................ 42
Decision: Endpoint Form Factor....................................................................... 43
Decision: Thin Client Selection......................................................................... 44
Receiver Selection ................................................................................................ 44
Decision: Receiver Type.................................................................................... 44
Decision: Initial Deployment ............................................................................. 47
Decision: Initial Configuration .......................................................................... 50
Decision: Updates.............................................................................................. 51
Resource Requirements ....................................................................................... 52
Decision: CPU..................................................................................................... 53
Decision: Memory .............................................................................................. 53
Decision: IOPS.................................................................................................... 54
Decision: Storage Space ................................................................................... 55
Access Layer ............................................................................................................ 56
Decision: Authentication Point ............................................................................ 56
Decision: Pre-Authentication Policy.................................................................... 57
Decision: Authentication Policy........................................................................... 58
Decision: Session Policy...................................................................................... 59
Decision: Session Profile ..................................................................................... 61
Decision: Preferred Data Center .......................................................................... 63

Decision: Access Bandwidth ............................................................................... 65
Desktop Layer........................................................................................................... 66
Personalization...................................................................................................... 66
User Profiles ....................................................................................................... 66
User Policies....................................................................................................... 71
Printing................................................................................................................ 76
Personal vDisk.................................................................................................... 84
Applications........................................................................................................... 86
Decision: Application Delivery Method ............................................................ 87
Images.................................................................................................................... 91
Decision: Operating System.............................................................................. 91
Decision: Delivery Method................................................................................. 92
Decision: Computer Policies............................................................................. 92
Decision: Desktop Catalogs.............................................................................. 93
Decision: Desktop Groups ................................................................................ 93
Control Layer............................................................................................................ 95
Access Controllers................................................................................................ 95
Remote Access Architecture............................................................................. 95
StoreFront........................................................................................................... 98
Desktop Controllers............................................................................................ 103
XenDesktop....................................................................................................... 103
XenApp.............................................................................................................. 106
XenClient Synchronizer ................................................................................... 109
Provisioning Services...................................................................................... 117
Infrastructure Controllers................................................................................... 133
Active Directory................................................................................................ 133
Database ........................................................................................................... 141
Hardware Layer ...................................................................................................... 152
Hypervisor ........................................................................................................... 152
Hyper-V ............................................................................................................. 152
Section 4: Deploy.................................................................................................... 163
Overview ................................................................................................................. 164
Section 5: Monitor................................................................................................... 165

Overview ................................................................................................................. 166
Support ................................................................................................................... 166
Structure .............................................................................................................. 166
Tools..................................................................................................................... 171
Delegated Administration................................................................................... 173
XenApp Delegated Rights................................................................................ 173
XenDesktop Delegated Rights ........................................................................ 175
Provisioning Services Delegated Rights........................................................ 175
XenServer Delegated Rights ........................................................................... 175
Testing .................................................................................................................... 176
Change Control ................................................................................................... 176
Testing Phases.................................................................................................... 179
Operations .............................................................................................................. 179
Daily...................................................................................................................... 180
Weekly.................................................................................................................. 181
Yearly ................................................................................................................... 182
Monitoring............................................................................................................... 184
Performance ........................................................................................................ 185
Services ............................................................................................................... 191
Events .................................................................................................................. 196
Growth.................................................................................................................. 197
Section 6: Appendix ............................................................................................... 198
Glossary and Abbreviations.................................................................................. 199
Baseline Policy Reference..................................................................................... 200
Citrix User Policy Settings ................................................................................. 201
Citrix Computer Policy Settings......................................................................... 210
Microsoft Windows Policy.................................................................................. 212
Folder Redirection Policy................................................................................... 215
Policy Quick Reference ......................................................................................... 217
User Policy Reference ........................................................................................ 217
Computer Policy Reference ............................................................................... 223
Revision History..................................................................................................... 227
Authors ................................................................................................................... 227

SECTION 1: OVERVIEW

Page 4
Introduction
If you read industry magazines, you would think desktop virtualization is an impossible technology to
deploy. So why do it? The simple answer is that Desktop Virtualization is designed to make your work life
more flexible and mobile. A flexible work life can mean different things to different people: any device, any
place, and any time. The “any-ness” of the work life means that a user becomes mobile. They can work
on any endpoint device, can work from any location, and can work at any time they choose as long as
they get their work done. Desktop virtualization provides flexibility without complicating the user‟s life or
putting corporate data at risk.
Take the following response from a desktop virtualization user:
Unfortunately, many organizations struggle or fail to achieve this level of success within their user
community. Why does one organization succeeds while another organization struggles?
If we compare the factors between success and failure between desktop virtualization and other
technology related projects, we see that there is little difference:
 Lack of justification: Without a solid business reason, desktop virtualization is simply a new way
to deliver a desktop. A business justification gives the project team a goal to strive towards.
 Lack of a methodology: Many people who try and struggle to deploy a desktop virtualization
solution do so because they jump right in without understanding or implementing the appropriate
prerequisites. A structured methodology provides the path for the project.
 Lack of confidence: For many who embark on a desktop virtualization project, there is a lack of
experience, which creates a lack of confidence in the design. Architects begin to second guess
themselves and the project stalls.
Our hope is that this handbook can alleviate the anxiety associated with a desktop virtualization design by
showing how to tackle the challenges in a manner that is technically sound, but also feasible and effective
for organizations facing deadlines and other organizational challenges.
The Citrix Virtual Desktop Handbook examines the project lifecycle for a desktop virtualization project.
The Handbook provides the methodology, experience and best practices needed to successfully design
your own desktop virtualization solution. However, this handbook is not the only resource to guide you
through your desktop virtualization journey. Citrix also provides Project Accelerator; an interactive online
tool creating customized sizing and design recommendations based on the methodology, best practices
and expert advice identified within this handbook.
Experience from the Field
As a remote employee for [company], I struggled every time I needed to access the company‟s
intranet, which forced me to VPN into the network. I also kept data on my local device because trying
to access it over my broadband connection was too slow. Some coworkers did the same and lost data
due to a virus, thankfully I was luckier.
Depending on my mood (and the weather), changing devices and locations was a challenge as I had
to have my applications and data copied to many different endpoints. I know this was unsecure, but I
didn‟t care because I was more concerned with flexibility.
Since moving to a virtual desktop, I‟m able to use any device. I‟m able to work from any location. And
best of all, I don‟t have to worry about copying my data and applications onto all of my personal
devices. I paid for these devices; I don‟t want work to clutter up my personal space.

Page 5
Methodology
The Citrix Virtual Desktop Handbook follows the Citrix Consulting Methodology. A proven methodology
that has been successfully employed across thousands of desktop virtualization projects. Each phase
includes guidance on the important questions to ask, what tools to use, and tips to help you succeed.
The Citrix Consulting Methodology consists of four phases:
Figure 1: Citrix Consulting Methodology
Assess
The Assess phase identifies information necessary for the design:
1. Business Drivers: Documents and prioritizes the motivation and key drivers behind the
desktop virtualization initiative so that the project team can focus their efforts on creating a
solution which meets the needs of the business.
2. Data Capture: Data is collected on users, applications, devices and infrastructure. This
information will be used to segment users, identify risks and determine current capabilities.
3. User Segmentation: Users are divided up into groups based on a common set of
requirements so that an appropriate FlexCast model can be assigned to each one without
compromising on performance or functionality.
4. Application Assessment: The application list is rationalized by removing legacy applications,
standardizing application versions, and removing non-business applications. The remaining
applications are then analyzed for compatibility issues.
5. Roadmap: Each user group is prioritized by comparing implementation time against business
impact. The results of this prioritization process are then used to update the project plan. The
project team is defined according to the skillsets required.
Assess Design Deploy Monitor

Page 6
Design
The design phase focuses on a five-layer approach to ensure than all key design decisions are
considered:
1. User Layer: The user layer documents the recommended end points and the required user
experience functionality.
2. Access Layer: This layer shows how the user layer will connect to their desktop, which is
hosted in the desktop layer of the architecture. Local users will often connect directly to
StoreFront while remote users often connect via a set of firewalls that protect the internal
environment. To bridge the firewalls, remote users will often connect with an SSL-VPN device
(like Citrix Access Gateway). Finally, for disconnected users using Citrix XenClient, they will
need to synchronize their local images with the backend store, requiring additional access
through the set of firewalls separating the internal/external environments.
3. Desktop Layer: This layer contains the user‟s virtual desktop, broken down into FlexCast
models. It is subdivided into three components: Image, Applications and
Personalization. Within each sub-layer, specifics are documented detailing the operating
system, assigned policies, and profile design and application requirements.
4. Control Layer: This layer is responsible for managing and maintaining all other layers. It
provides details on the controller requirements to support the entire solution. The Control
layer is broken up into Access Controllers, Desktop Controllers and Infrastructure Controllers.
The Access Controllers focuses on the hardware requirements to support the Access
Layer. The Desktop Controllers provide details on the components required to support the
desktop layer, which could include XenDesktop, XenApp, XenClient or Provisioning
Services. Finally, the Infrastructure Controllers are responsible for providing the underlying
resources to support each component. These resources can include databases, license
servers, and hypervisor controllers.
5. Hardware Layer: This layer is responsible for the physical devices required to support the
entire solution. It includes servers, processors, memory, storage devices. This layer is broken
into three groups focused on providing the necessary resources for specific parts of the entire
solution. One group of servers will support the XenApp (shared) components (if applicable). A
second group of servers will support the XenDesktop (VDI) components (if applicable). A final
group of servers will support the underlying infrastructure for the entire environment, which is
identified as the Control Layer.
Deploy
During the Deploy phase, FlexCast Infrastructure is installed and configured as described in the
design phase. All components of the infrastructure are thoroughly unit and regression tested before
users are provided with access to the environment.
A pilot is performed to ensure that all requirements are addressed in the deployment before it is rolled
out. In addition, the pilot helps determine scalability thresholds for the production environment. Key
success criteria should be identified for the pilot and tested by a subset of end users. Once the pilot is
complete, the solution is rolled out to production. Rollout includes technical assistance, deployment
work plans and end user and IT staff training.

Page 7
Monitor
The Monitor phase consists of four different activities:
1. Support: When problems arise, technical support is the first point of contact for issue
resolution. This section addresses the proper staffing, organization, training, and tools utilized
in effective support organizations.
2. Testing: Regular upgrades are required to ensure that the virtual desktop environment is up
to date. The change management processes ensures that all improvements are properly
approved, tested, and validated by appropriate parties. This section covers the proper
processes that ensure changes in production environments are deliberate, proven, and
accountable.
3. Operations: Maintenance, issue prevention and resolution are core responsibilities in running
a Citrix infrastructure. When the responsibilities and assignments are structured properly,
friction is kept to a minimum, reducing issues and their resolution times. This section
discusses routine operations that Citrix environments require for optimal performance.
4. Monitoring: After the virtual desktop environment has been designed and rolled out to
production, ongoing monitoring is required. Monitoring the new environment enables
administrators to address issues proactively. By having an in-depth understanding of current
and expected behavior of the various components, administrators are better equipped to
discover an issue before it impacts the user community. Furthermore the data tracked during
normal operations can be used for trending and capacity planning.

Page 8
Project Plan
A detailed, up-to-date project plan is vital to the success of the project. The project manager should use
the project plan to monitor costs, manage project staff, follow progress against milestones, and track key
dependencies such as hardware, storage, training, licenses and certificates so that they can be
addressed before they can become bottlenecks. The project plan should be regularly distributed between
project team members to ensure that everybody is aware of progress as well as remaining tasks.
At the start of the project, only plan for the Assess phase. You won‟t be able to plan for the Design or
Deploy phases yet because you‟ll need a better understanding of the FlexCast models, capabilities, user
groups and applications required. After the roadmap stage, update the project plan to include separate
design and deploy activities for each user group identified, ordered by priority. This will help to ensure
that the business receives maximum value from their investment as soon as possible and allow
infrastructure to be built out in manageable IT projects that minimize risk and maximize the chance for
project success.
A sample project plan, created in Microsoft Project 2010 format, accompanies this document. A high level
overview is shown below:
Figure 2: Sample Project Plan
Note: Some of the tasks in the project plan template have been scheduled to occur in parallel because it
is assumed that multiple project team members are available. All activities, resources and scheduling
within the sample project plan should be reviewed prior to starting your project.

Overview
Creating a desktop virtualization solution begins with doing a proper assessment. Architects that fail to
properly assess the current environment find that they require the assess information later on, forcing
them to backtrack, which can potentially stall and put the project at risk.
By gathering all of the information from the outset, the architect will gain an appreciation for the current
environment and be able to work from the beginning on properly aligning business and user requirements
with the overall solution.
The assess phase is a four-step, simple to follow process:
Figure 3: Assess Process
Business Drivers
The first step in your virtual desktop project should be to understand and prioritize the strategic
imperatives of the organization. This enables the project management team to define success criteria and
allows the design team to create a tailored and optimized architecture.
Requirements can be captured during meetings or by distributing questionnaires. Meetings are more time
consuming, but allow for follow-up questions to be asked and help to simplify the prioritization process. It
is important that this exercise be completed jointly by both business managers and IT decision makers
since both groups will have significantly different viewpoints. Take the following examples of what certain
organizations faced, which drove their selection of desktop virtualization.
These are just a few examples, but they demonstrate how organizations think about their priorities. Most
organization do not focus on technology, they focus on the needs of the user and of the organization.
These needs can be met with technical solutions but it is imperative the team understands the “Why” of
the project.
Justification
Gather the
right data
Group
users
Group
applications
Build the
right team
Finance: A large financial institution had a base of operations in the city designated as the host city for
an upcoming G8 summit. As these types of meetings historically include riots, protests and other
issues that can disrupt business and the safety of their employees, the financial organization needed
an alternative allowing their users to work from the safety of their homes.
Agriculture: Due to thin margins, an agriculture organization wanted to save money by extending the
life of desktop PCs while still being able to run the latest applications.
Healthcare: A large healthcare organization was in need of a solution to simplify application updates
as the main application required updates on a weekly basis. Due to the distributed nature of the
endpoint devices, the organization was in need of a better application delivery solution.

In addition to the three real world examples, the following table identifies a few other priorities often stated
from many organizations:
Requester Requirement
Business Managers Better IT Agility: Flexible desktop solution that is capable of accommodating periods of
change such as rapid growth or downsizing. For example, enabling the business to setup
project offices or temporary points of sale very rapidly without long delays or IT notification
periods.
Bring Your Own Device: Empower employees to choose their own devices to improve
productivity, collaboration and mobility.
Collaboration: With an increase in both globalization and mobility, team members are often
dispersed across multiple physical locations. Powerful collaboration capabilities are required
to ensure high levels of productivity, efficiency and quality.
Work From Anywhere: The business needs to support home workers in order to attract and
retain top talent, and / or travelling employees.
IT Decision Makers Better Desktop Management: Simplify the management of desktop infrastructure. IT is not
as proactive as they would like and spend too much time „fighting fires‟.
Increase Security: Data theft or the loss of devices which contain sensitive data (e.g.
laptops or tablets) is a big risk and preventive measures are a top priority.
Extend Desktop Hardware Lifecycle: Replacing workstations every three to five years in
order to keep up with the requirements of the operating system or the applications has been
very costly.
Table 1: Sample Business Drivers
The prioritization process should be completed in collaboration with the project team, business managers
and IT managers so that all views are considered.
Early on, organizations often need to estimate the hardware required to support the new solution. Citrix
Project Accelerator provides a fast and accurate way to estimate hardware and storage requirements.
Data Capture
The Data Capture process collects key information on users, devices, applications and infrastructure that
is required during subsequent phases of the project, including User Segmentation, Application
Assessment, Project Roadmap and Design.
Strategies
There are three main techniques that can be used to complete the data capture:
Manual
For small organizations, it is possible to capture user and application data by visiting each desktop
in person or by connecting remotely. Performance counters can be used to acquire raw user data
while the “Add / Remove Programs” applet in Windows XP or the “Program & Features” applet in
Windows 7 can both be used to provide a list of installed applications as well as information on how
often they are accessed.
Most medium and large-sized organizations use an Enterprise Software Deployment (ESD) tool
such as Microsoft System Center Configuration Manager (SCCM). Since the ESD solution is
typically the sole method of deploying application packages to workstations, the project team can
query the ESD configuration to identify which applications are assigned to which desktops.
However, most ESD solutions do not provide details on application performance requirements,
application usage metrics or user installed applications. In addition, significant effort is required to
extrapolate, analyze and document application and desktop information from the ESD tool.

The problem with the manual approach is that it is difficult to gain a good understanding of each
user and application‟s performance requirements over time. In addition, the manual approach is
very time intensive per desktop making it inappropriate for medium and large-sized organizations.
It can also be difficult to query application requirements of mobile users who may spend extended
periods of time away from the office.
Survey
A questionnaire can be created and distributed to each business manager to identify user and
application requirements within their departments. Survey questions may also be answered in a
face to face meeting or phone call. Surveys are far less time consuming for the project team than
the manual approach because the focus is now upon the department managers to identify the
applications necessary for their employees to complete their jobs. However, it is unlikely that every
manager within the organization will complete the questionnaire and a completion ratio of 70%
should be considered normal. The results from a survey are typically less accurate than the
manual or automated approach. Although less time consuming than the manual approach,
considerable effort is still required to extrapolate, analyze and document the results from the
returned questionnaires.
Automated
There are a variety of automated inventory tools available that allow for a wide range of information
to be collected, including applications installed, access frequency, and performance requirements.
This data is uploaded to a central database so that reports can be generated across the entire user
and desktop estate. To help reduce the cost and complexity of desktop virtualization projects,
Citrix has partnered with Lakeside Software to provide Project Accelerator users with a free 60-day
license of Lakeside FastTrack. FastTrack is a sophisticated inventory tool that has been developed
specifically based on Citrix methodologies, terminology and best practices. An automated
inventory tool is a great solution for medium and large-sized organizations, however the centralized
infrastructure and agent deployment effort is unlikely to be appropriate for very small organizations
due to the time required when compared to the manual method. In addition, some organizations
may not allow inventory agents to be installed across the desktop estate or may not have rights to
install the agent on user-owned desktops.
The advantages and disadvantages of each approach are summarized in the following table:
Approach Data for all
Characteristics
Results Returned Admin Time per
Desktop
User Involvement
Manual No ~80% Long Likely
Survey Yes ~70% Medium Yes
Automated No ~100% Short No
Table 2: Comparison of Data Capture Strategies
Although the automated method is accurate, fast and does not inconvenience employees, there are
a number of business characteristics that automated tools cannot identify, for example, what is the
criticality of the user, have the applications been packaged, what is the licensing model and how
frequently is the application updated? Therefore, the recommended approach is to use the
automated method to identify technical characteristics and a survey to identify the business
characteristics.

Best Practices
Don‟t forget to follow these top recommendations from Citrix Consulting based on years of
experience:
Citrix Consulting Tips for Success
1. Representative Users: If you don‟t have enough time, resources, or licenses to inventory
every desktop in your organization, make sure that you pick a representative subset of users.
For example, deploying an agent on every desktop in the HR department but missing out the
sales and finance departments will impact your results. Take time to ensure that you select a
representative subset of users from every department and role within the organization. And
remember, not all users within the same department are the same.
2. Check the Data: When using an automated inventory tool, regularly check that the data has
been successfully uploaded to the central database. There have been cases reported where
insufficient space has been allocated to the central database resulting in several weeks of
agent data being lost causing the project to be delayed.
3. Monitoring Period: It is extremely important that the automated inventory tool monitors the
desktops over a sufficient period of time. At the very least, monitor the desktops for a
minimum of one month; ideally monitor usage over periods of peak activity such as quarter
end so that you have a good idea of average and peak application performance
requirements. There may be situations where users only need access to a specific
application during quarter end and unless you are running the automated inventory tool at
this time you will not be aware of its usage.

User Data Capture
Key results from the user data gathering exercise should be documented in the User Assessment
worksheet within the Citrix Virtual Desktop Handbook - Assess Workbook:
Business Characteristics
Business characteristics must be gathered manually from the business managers and cannot be
determined using automated data capture tools.
 Identification:
o Username: Provide a unique identifier for the user being assessed.
o Department: Identify the department that the user belongs to, for example, HR,
Finance or Sales.
o Role: Capture the primary job role of the user, for example developer, accountant or
manager.
o Business Manager: Provide the name and contact details of the user‟s business
manager so that you can follow up with additional questions.
o Assigned User Group: Do not complete. This field will be completed during the User
Segmentation.
 Business:
o Primary Data Center: Each user will have a primary data center assigned that will be
used to host their virtual desktop, data, and application servers. Identify the data
center that the user should be assigned to rather than the data center they are
currently using. Users will be grouped based on their primary data center so that a
unique architectural design can be created for each one.
o Mobility: Mobility requirements are used to help determine the appropriate FlexCast
model for each user group. For example, if a user group sometimes connects
remotely, this eliminates the Streamed VHD option, which requires a high-speed local
network. There are four classifications available:
Mobility Requirement
Local Always uses the same physical desktop on an internal, high-speed, secured
network.
Roaming Local Connects from different locations on an internal, high-speed, secured
network.
Roaming Sometimes connects from external variable-speed, unsecure networks.
Offline Sometimes needs desktop access while disconnected from any network.
Table 3: Mobility Characteristics

o Personalization: Personalization requirements are used to help determine the
appropriate FlexCast model for each user group. For example, if a user group
requires complete personalization, a Hosted VDI desktop with Personal vDisk will be
recommended as the optimal solution. There are three classifications available:
Personalization Requirement
None User cannot modify any user or application settings (e.g. kiosk).
Basic User can modify user-level settings within desktops and apps.
Complete User can make any change, including installing applications.
Table 4: Personalization Characteristics
o Security: Security requirements are used to help determine the appropriate FlexCast
model and policies for each user group. For example, if a user group requires high
security, a Hosted-Shared Desktop, Pooled Desktop or On-Demand Apps FlexCast
model will be recommended as the optimal solution. There are four classifications
available:
Security
Level
Data Stays in
Data Center or is
Encrypted
Prohibit User
Installs
Allow Multi-User
Operating
Systems
MAC / IP
Address Auditing
Low No No Yes No
Medium Yes Yes Yes No
High Yes Yes No No
High & Audit Yes Yes No Yes
Table 5: Security Characteristics
o Criticality: Criticality will be used to determine the level of high-availability, load
balancing and fault tolerance measures required. There are three classifications
available:
Criticality Requirement
Low Desktops can be unavailable for one day or more.
Medium Desktops must not be down for more than four hours.
High Backup desktops must always be available in case the primary fails.
Table 6: Criticality Characteristics
Technical Characteristics
o Workload: Collecting user performance requirements will allow the desktop
virtualization infrastructure to be sized appropriately and an appropriate FlexCast
model to be selected. For example, if a user group requires OpenGL graphics, a
Hosted Shared desktop will not be recommended because OpenGL is not optimized
on a Hosted Shared desktop. There are four classifications available:

User Type Characteristics Graphics
Light 1-2 office productivity apps or kiosk DirectX or None
Normal 2-10 office productivity apps with light multimedia use DirectX or None
Heavy Multimedia, data processing or application
development
DirectX or None
Heavy & OpenGL OpenGL graphics requirements OpenGL
Table 7: Workload Characteristics
Note: Performance thresholds are not identified based on processor, memory or disk
utilization because these characteristics will change dramatically following the
application rationalization and desktop optimization process. In addition, it is likely
that the user‟s management tools and operating system will change during the
migration process. Instead, workload is gauged based on the number and type of
applications the user runs.
 User Environment: The user environment is a vital but often overlooked aspect of desktop
virtualization projects. These data points will be required during the design in order to plan
user migration and to schedule preparation work. For example reducing the size of the
profiles and home drives before they are moved.
o User Profile: The user profile holds all user settings related to the look and feel of
Windows and the applications as well as certain application configurations, (e.g. mail
signatures) or application data (e.g. favorites). The following information should be
collected for each user‟s profile:
 Profile Type: There are five classifications available for “Profile Type” in the
User Assessment Worksheet: Local, Roaming, Mandatory, 3
rd
Party or
Unknown.
 Profile Version: There are two classifications available for “Profile Version” in
the User Assessment Worksheet: Type 1 (Windows XP) or Type 2 (Windows
Vista / 7).
 Profile Location: Name of the file server used to store the user‟s home drive.
 Size (MB): Size of the user‟s profile.
o Home Drive: The home drive is typically used to store user and application data. The
following information should be collected for each user‟s home drive:
 Home Drive Location: Name of the file server used to store the user‟s home
drive.
 Size (MB): Size of the user‟s home drive (excluding profile).
 Client Hardware: Client hardware details are captured to decide whether devices should be
reused, rebuilt or replaced. For example, Hosted VDI and Hosted Shared users will not
require a high-performance device. The following information should be collected for each
user‟s client device:
o Number of CPUs/Cores & CPU Speed (GHz)
o Memory (GB)
o Disk Size (GB)
o Operating System
o Age of the System (years)
o Peripherals

Note: Although the Application Assessment will identify the peripherals required by each
application there may be some peripherals that are used directly by the operating system, for
example smartcard readers. The connectivity interface required should be identified because
it may be necessary to create a policy that maps the required interface from the endpoint
device to the virtual desktop.
 Local Resource Mappings: For reasons of security, resource mappings such as clipboard
and client drives are typically disabled unless there is a specific reason to enable them.
Therefore, the user assessment should identify all resource mappings required so that
appropriate policies can be created during the design. The majority of mappings will be
controlled at the user group or departmental level, however, there may be specific
requirements to restrict or enable mappings for certain users or user groups. The following
information should be collected for each user:
o Local Drives
o Printing
o Clipboard
o Audio
o COM Port
o USB
o LPT
Application Data Capture
Key results from the application data gathering exercise should be documented in the Application
Assessment worksheet within the Citrix Virtual Desktop Handbook - Assess Workbook.
 Identification:
o Application Name & Version: Multiple versions of the same application are likely to
be in use within the organization. Create a new row in the table for each version of
the application.
o Application Owner: If known, provide the name and contact details of the application
owner so that you can follow up with additional questions. If there is no application
owner assigned or you do not know the application owner, enter “Unknown.”
o Status: Identifies whether the application is being considered for deployment to the
new virtual desktop solution. For now, enter a value of “Active” for each application
identified during the inventory.
 Assignment:
o Number of Users: Provide the total number of users that have access to the
application.
o Departments: Document which departments require access to the application.
o User Groups: Access to an application may be restricted to specific user groups. For
example, access to the Citrix administration tools may be limited to the Citrix team
within the IT department.

o Individual Users: Some applications may be restricted to a small number of users
due to licensing or security restrictions. For example, access to a CAD application
may be restricted to a small number of employees within the Design Team.
 Workload: Collecting application workload requirements will allow the virtualization
infrastructure to be sized and an appropriate application delivery model to be selected. For
example, resource intensive applications will not be delivered via a Hosted Shared Desktop
because there is limited control over how the resources are shared between users. There are
three classifications available for “Workload” in the User Assessment Worksheet:
Workload Requirement
Resource
Intensive
If the application uses more than 1200MHz+ of available processor resources (50% of
a 2.4GHz CPU) or 1GB+ of RAM then it should be classified as “Resource Intensive”.
OpenGL If the application is graphically intensive, certain delivery models may be more
appropriate than others. For example, XenApp supports HDX 3D which offers GPU
acceleration for DirectX only while XenDesktop supports HDX 3D Pro which offers
GPU acceleration for both DirectX and OpenGL. Therefore, Open GL based
applications, should be classified as “OpenGL”.
None The application is not resource intensive and does not require OpenGL graphics.
Table 8: Application Workload Characteristics
 Business: Do not complete. This section of the worksheet will be completed during the
Application Assessment, once the application list has been completed.
 Compatibility: Do not complete. This section of the worksheet will be completed during the
Application Assessment, once the application list has been completed.
To save time and effort, the remaining items in the Application Assessment worksheet will be
completed during the Application Assessment once the rationalization process has been completed.
Capabilities Assessment
The information captured during the Capabilities Assessment will be used to achieve the following
objectives:
1. Identify Risks: Like traditional desktops, desktop virtualization is dependent on a wide range
of supporting technologies, including storage, networking, directory services and applications.
In many cases, issues reported with virtual desktops are a symptom rather than a cause. For
example, slow performance and disconnected sessions are more often caused by periods of
high-latency than a desktop specific issue. Therefore, a desktop virtualization project is an
excellent opportunity to review an organization‟s existing infrastructure to ensure that it
provides a good solid foundation upon which the VDI environment can be built. Any risks
identified should be reported to the project manager so that they can be addressed
appropriately.
2. Create Roadmap: The Capabilities Assessment provides the project team with a detailed
understanding of the existing environment so that they can estimate implementation time for
each user group and prioritize the implementation order.
3. Virtual Desktop Design: Provide the project team with detailed information on the current
state of the environment so that they can successfully integrate the new virtual desktop
solution. The Capabilities Assessment should also determine whether existing infrastructure
components can be leveraged or whether new infrastructure needs to be purchased, for
example shared storage and virtualization technologies.

Key results from the capabilities gathering exercise should be documented in the Citrix Desktop
Virtualization Handbook – Capabilities Assessment template. A list of questions has been provided
per infrastructure technology to highlight key information that needs to be collected. These questions
are based on the experiences of Citrix Consulting across numerous desktop virtualization projects and
should be answered by the appropriate technology architect for each section. Not all sections need to
be completed, for example if the organization does not already have a Citrix XenApp or XenDesktop
environment this section can be left blank. The length of the capabilities assessment will vary based
on the size and complexity of the environment but typically takes about three days to complete.
Best Practices
Performing a capabilities review helps set the foundation for the design by identifying experiences,
technologies and platforms to pursue. The review should first focus on whether the following areas
have already been implemented and if so, what is the current satisfaction – Access, Desktops,
Active Directory, XenApp/XenDesktop, Database, Virtualization, Storage, Network and Operations.
Don‟t forget to follow these top capabilities assessment recommendations from Citrix Consulting,
which are based on years of experience:
1. Discussions: Meet with the architects rather than sending them a questionnaire so that
additional questions can be raised, diagrams drawn and detailed explanations provided.
2. Schedule Meetings: It is advisable to schedule meetings with the architects well in advance
to ensure availability. Provide each architect with a copy of the questions that you plan to
ask them so that they can prepare appropriately. Also, when scheduling the meetings
request any relevant documentation for background reading as this may prompt additional
questions and discussions.
3. Documentation: Use this document to capture your discussions with the architects. The
document can then be circulated amongst the project team to ensure that everybody has the
same information.
4. Future Initiatives: Ask the architects whether they are planning any future initiatives, for
example upgrading to a new product version, or adjusting the architecture. This will ensure
that the project team is aware of all upcoming changes.
5. Identify Risks: It is important that all risks are identified as early as possible so that they can
be tracked and appropriate action taken. Risks should be graded in severity and if possible
remediation plans should be created stating an estimated time-frame and cost.

User Segmentation
Once the data capture is complete, you‟re ready to start dividing up the users into different groups based
on a common set of requirements. This allows a FlexCast model to be assigned to each user group
without compromising on performance or functionality.
User Groups
Users are often classified as “task workers”, “branch office workers”, “mobile workers” and the like.
Unfortunately, this classification is too broad to offer meaningful segmentation because many real
users can simultaneously be described as task workers, mobile workers, and branch office workers.
Instead, group users together that have the same requirements for:
1. Primary Data Center
2. Workload
3. Mobility
4. Personalization
5. Security
6. Criticality
The fastest and easiest way to identify your user groups within the User Assessment workbook is to
filter the results based on these key requirements. Once you have identified the users within one
group, transfer the relevant information to the User Segmentation worksheet within the Citrix Virtual
Desktop Handbook - Assess Workbook.
FlexCast
As with physical desktops, it is not possible to meet every user requirement with a single virtual
desktop type. Different types of user need different types of desktop. Some users may require
simplicity and standardization, while others may require high levels of performance and
personalization. Implementing a single desktop virtualization model across an entire organization will
inevitably lead to user frustration and reduced productivity.
Citrix FlexCast offers a complete set of application and desktop virtualization technologies that have
been combined into a single integrated solution. Because each FlexCast model has different
advantages and disadvantages, it is important that the right model is chosen for each user group
within the organization.
This chapter will identify the different FlexCast models available and provide guidance on how to
select the most appropriate one for each user group.
Utility Company: A large utility company collected data on every user in their organization. During
the user segmentation process it was realized that the organization‟s existing role definitions were
sufficiently well defined that all the users within a role shared the same requirements. This allowed
a significant amount of time to be saved by reviewing a select number of users per group.
Government: A government organization discovered that there was significant deviation between
user requirements within each role, particularly around security and criticality. As such, each user
needed to be carefully reviewed to ensure that they were grouped appropriately.

FlexCast Model Comparison
There are five FlexCast models available, the advantages and disadvantages of each model are
described below:
 Hosted Shared: With the Hosted Shared Desktop model, multiple user desktops are hosted
on a single server-based operating system. The Hosted Shared Desktop model provides a
low-cost, high density solution, however applications must be compatible with a multi-user
server based operating system. In addition, because multiple users are sharing a single
operating system, users are restricted from performing actions which may negatively affect
other users, for example installing applications, changing system settings and restarting the
operating system. There is also the potential that a single user could consume an unfair
share of resources which may negatively affect other users. The Hosted Shared FlexCast
model is provided by Citrix XenApp in combination with Microsoft Remote Desktop Services
(RDS).
 Hosted VDI: The Hosted VDI FlexCast model provides each user with a desktop operating
system. Hosted VDI desktops are less scalable than Hosted Shared desktops because each
user requires their own operating system. However, Hosted VDI desktops remove the
requirement that applications must be multi-user aware and support server based operating
systems. In addition, the Hosted VDI model provides administrators with a granular level of
control over the number of virtual processors and memory assigned to each desktop. The
Hosted VDI model is provided by Citrix XenDesktop, and offers the following sub-categories:
o Pooled-Random / Streamed: Desktops are based on a single master image and
provisioned using Citrix Machine Creation Services or Citrix Provisioning Services. Users
are dynamically connected to one of the desktops in the pool each time they logon.
Changes to the desktop image are lost upon reboot.
o Pooled-Static: Desktops are based on a single master image and provisioned using
Citrix Machine Creation Services or Citrix Provisioning Services. Users are allocated a
virtual desktop on first access. Once assigned, users will always be connected to the
same virtual desktop. Changes to the desktop image are lost upon reboot. If high
availability is required (medium-criticality), the personal vDisk must be stored on shared
storage.
o Pooled / Streamed with Personal vDisk: Desktops are based on a single master image
and provisioned using Citrix Machine Creation Services or Provisioning Services. Users
are allocated a virtual desktop on first access. Once assigned, users will always be
connected to the same virtual desktop. Changes to the desktop are stored in a personal
vDisk and retained between reboots. Desktops with a personal vDisk cannot be shared
between multiple users; each user requires their own desktop. If high availability is
required (medium-criticality), the personal vDisk must be stored on shared storage.
o Dedicated: Desktops are based on a single master image and provisioned using Citrix
Machine Creation Services. Users are allocated a virtual desktop on first access. Once
assigned, users will always be connected to the same virtual desktop. Any changes
made by the user persist for the life of the virtual desktop, even across reboots. Once a
dedicated desktop has been created it must be maintained manually or via 3
rd
party
desktop management tools. Changes to the master image are not reflected on existing
dedicated desktops.
o Existing: Virtual desktops which have not been created using Citrix Machine Creation
Services or Citrix Provisioning Services. For example, a manual build, from template,
cloned or using 3
rd
party tools. These desktops must be managed manually or with 3
rd
party desktop management tools.

o Physical / Remote PC: Physical Desktops that have already been deployed, including
Blade PCs and user workstations (Remote PC). These desktops must be managed
manually or with 3
rd
party desktop management tools. Physical desktops with a GPU
support HDX 3D pro, however Remote PC uses the standard ICA thinwire protocol and
does not allow offloading to the GPU to render video files or 3D apps.
 Streamed VHD: Desktops are based on a single master image and provisioned using
Provisioning Services. The Streamed VHD FlexCast model allows Windows 7, Vista, or XP
desktops to be run locally on the user‟s desktop computer. Streamed VHD is a great solution
for high-end workstations because it allows them to leverage local processing power.
Streamed VHD requires a LAN connection to be in place between the desktop and the
Provisioning Servers and changes to the desktops are lost upon reboot.
 Local VM: Windows 8, Windows 7, Vista, or XP desktops running locally within a hypervisor
on the user‟s laptop. The virtual desktop image is completely delivered to the hypervisor to
allow for offline connectivity. Citrix XenClient is used to provide the Local VM FlexCast model.
 On-Demand Apps: The On-Demand Apps FlexCast model does not provide users with a
virtual desktop; instead Windows applications are centralized in the data center, and instantly
delivered via a high-speed protocol (requires connection) or streamed (offline support) via
Citrix Application Streaming or Microsoft App-V.
The following table provides a summary of the different FlexCast models available:
FlexCast Model User
Installed
Apps
Image Delivery
Technology
Virtual / Physical Access Desktop to
User Ratio
Hosted Shared No Installed or PVS Physical / Virtual HDX 1 : Many
VDI: Pooled-Random No MCS Virtual HDX 1 : Many
VDI: Pooled-Static No MCS Virtual HDX 1 : 1
VDI: Pooled with PvD Yes MCS Virtual HDX 1 : 1
VDI: Dedicated Yes MCS Virtual HDX 1 : 1
VDI: Existing Yes Installed Virtual HDX 1 : 1
VDI: Physical / Remote PC Yes Installed Physical HDX 1 : 1
VDI: Streamed No PVS Physical / Virtual HDX 1 : Many
VDI: Streamed with PvD Yes PVS Physical / Virtual HDX/Local 1 : 1
Streamed VHD No PVS Physical HDX 1 : 1
Local VM Yes XC Virtual (XenClient) HDX 1 : 1
On Demand Apps No Installed or PVS Physical / Virtual Local 1 : Many
Table 9: FlexCast Model Comparison

FlexCast Model Selection
There are key technical differences between XenApp and XenDesktop. Both products provide the
user with an excellent user experience based on HDX. However, due to their underlying
architecture, many HDX features are slightly different, as shown in the following table:
Criteria XenDesktop 5.6 XenApp 6.5
HDX Broadcast No major differences
HDX MediaStream
(server rendered)
 24 fps (configurable to 30)
 Adaptive display
 Low latency audio path
 12 fps (configurable to 30)
HDX MediaStream
(Flash redirection &
Windows Media
Redirection)
No major differences
HDX Plug-n-Play  Generic USB support on the LAN in
addition to all optimized virtual
channels for device level
redirection.
 On a LAN environment, should be
able to support most USB devices.
 General support for Windows
portable USB devices (standard
keyboards, mice, printers,
smartcards, etc.).
 Optimized virtual channels for
device level redirection rather than
USB port level redirection,
potentially limiting availability for
certain USB devices.
HDX WAN No major differences
HDX Rich Graphics  HDX 3D Pro (not supported on
Remote PC) for deep GPU-based
compression and OpenGL/DirectX
acceleration (one GPU per user)
 Adaptive Display
 Microsoft RemoteFX support
 Aero Redirection
 DirectX acceleration (shared GPU)
HDX 3D Pro  GPU acceleration of DirectX and
OpenGL
 One user per GPU
 Deep-level compression
 3D SpaceMouse support
 GPU acceleration of DirectX only
 Each GPU can be shared by
multiple users
 Applications must be compatible
with RDS
HDX Smart Access No major differences
HDX RealTime  Low latency audio path
 UDP/RTP support for tolerance to
network congestion and packet loss
 DSCP and WMM packet tagging
 Softphone compatibility somewhat
limited due to use of Windows
Remote Desktop Services (formerly
Terminal Services)
HDX Adaptive
Orchestration
No major differences
Table 10: XenApp and XenDesktop Feature Comparison
Note: An overview of HDX technologies can be found on the Citrix HDX site.
Each user group in the User Segmentation worksheet should be compared against the following
table to determine which FlexCast Model should be assigned. Ensure that you update the
FlexCast value for each user group in the worksheet.

Segmentation
Characteristic
Hosted
Shared
VDI: Pooled-
Random /
Streamed
VDI:
Pooled-
Static
VDI: Pooled
/ Streamed
with PvD
VDI:
Dedicated
VDI:
Existing
VDI:
Physical
VDI:
Remote PC
Streamed
VHD
Local VM On
Demand
Apps
Workload
Light           
Normal           
Heavy           
Heavy + Open GL           
Mobility
Local           
Roaming Local           
Remote           
Offline           
Personalization
None           
Basic           
Complete           
Security
Low           
Medium           
High           
High + Audit           
Criticality
Low           
Medium           
High           
“”: Recommended. “”: Viable. ““: Not Recommended. “”: Streamed Apps Only. “” Viable with Microsoft Restrictions
Table 11: FlexCast Model Capability Comparison

Best Practices
experience:
1. Review the Capabilities Assessment: There are multiple FlexCast models available that
are an equal match, for example VDI: Pooled Random (Provisioning Services) and VDI:
Streamed (Machine Creation Services). In this situation, refer back to the results from the
Capabilities Assessment. If you already have skills or experience with one model over
another then this is a good reason for selecting it.
2. Lead with Hosted Shared/VDI: As you can see in the FlexCast Capability table above, the
Hosted VDI and Hosted Shared FlexCast models can be used in the majority of situations.
The Streamed VHD and Local VM FlexCast models should only be used on an exception
basis. By reducing the number of FlexCast models required, you will help to reduce
deployment time and simplify management.
3. Perfect Match. It may not be possible to select a FlexCast model which is a perfect match
for your user group, for example you can‟t provide users with a desktop that is highly secure
and offers complete personalization at the same time. In these situations, select the
FlexCast model which is the closest match.
4. Criticality. There are only three FlexCast models that meet the needs of a high criticality
user group (backup desktops available) – none of which allow for complete personalization.
If a high-criticality user group also requires the ability to personalize their desktop they could
be provided with a pool of backup desktops (Hosted Shared, pooled, streamed) in addition to
their primary desktop. Although these desktops would not include customizations made to
their primary desktop, they would allow users to access core applications such as mail,
Internet and Microsoft Office.

Application Assessment
At the completion of the user segmentation exercise, users will be divided up into groups based on a
common set of requirements. The next step is to provide users with the applications they need to perform
their jobs. There is a two-step process that will help guide you through the assessment of these
applications:
1. Rationalization: Help to simplify the application assessment by removing redundant applications
from the inventory that was captured during the data gathering chapter.
2. Business Characteristics: Work with application owners to define key application characteristics
that can‟t be determined using automated assessment tools or end-user questionnaires.
This chapter will provide guidance on how to complete each step and to document the results in the
Application Assessment worksheet within the Citrix Virtual Desktop Handbook - Assess Workbook.
It is important to remember that this is still the assessment phase of the project and that application
delivery solutions should not be selected until the design phase when all relevant information is available.
Rationalization
The number of applications identified during the inventory is often surprising, even for organizations
that believe they have a high-level of control over applications. To help reduce complexity as well as
overall time required, it‟s important to take the time to consolidate the list of applications. Start by
arranging an “Application Assessment” meeting with all relevant application owners.
Note: Consolidated applications should be identified within the Application Inventory Workbook by
selecting “Consolidated” in the Status column. Consolidated applications should not be removed from
the workbook so that the rationalization process can be reviewed within the organization.
The following guidelines will help to ensure that your application list is consolidated appropriately:
 Multiple Versions: Different versions of the same application may have been identified during
the inventory. There are various reasons for this, including an inconsistent patching or
upgrade process, decentralized application management, limited licenses and situations
where users require specific application versions for compatibility with other applications,
macros and document formats. Where possible, work with the application owners to reduce
the number of versions required. The best practice is to standardize on a single version of
each application, typically the latest.
 Business Applications: Applications which are not required by the business should be
removed from the application inventory to reduce resource requirements and to help simplify
the overall project. Non-business related applications are typically found in an application
inventory when users have been provided with the ability to install their own applications and
typically include games, communication clients, screen savers, peripheral software and media
players.
 Legacy Applications: The inventory may identify legacy applications that have since been
retired or that are no longer required within the business. These applications may not have
been removed from the desktops because there is no established process to do so or
because there are always more high-priority activities to complete. These applications should
be consolidated during the rationalization stage of the application assessment.

 Management Applications: The antivirus, application delivery, monitoring, inventory,
maintenance and backup applications will be completely re-designed across the organization
during the desktop virtualization project. These applications should also be consolidated
during this stage.
Once the application list has been rationalized, time should be taken to identify the business
characteristics that cannot be determined using automated data capture tools. The characteristics of
each application should be documented in the Application Assessment worksheet within the Citrix
Virtual Desktop Handbook - Assess Workbook.
 Update Frequency: The time and effort required to update an application will vary according
to the delivery model chosen. For example, it is faster to update an application installed into a
provisioned Hosted Shared desktop image than a VDI: Existing one. Therefore, each
application should be classified based on its update frequency so that an appropriate delivery
model can be selected during the design. There are four classifications available for “Update
Frequency” in the Application Assessment Workbook:
Update Frequency Requirement
Low One or two updates required per year.
Medium One or two updates required per quarter.
High One or two updates required per month.
Table 12: Update Frequency Characteristics
 Packaged: Packaged applications are installed in an automated fashion without intervention.
Application packaging reduces deployment time and ensures that applications are installed
consistently across the environment. If the design team requires additional packaging to be
completed, sufficient time should be incorporated into the Deploy phase to complete this
activity.
 Criticality: Application criticality will be considered during the design phase to ensure that the
delivery model selected incorporates a sufficient level of redundancy to meet business needs.
However, there may be situations where different user groups or departments require a
different level of criticality to others. For example, the criticality of Microsoft Office may be
high for the sales department and low for the helpdesk user group. If so, select the highest
criticality required. There are four classifications available for “Criticality” in the Application
Assessment Workbook: High, Medium, Low or Unknown.
Government: A government organization identified that there were 2,660 applications installed
across their desktop estate. Most of which were installed by users with local administrative rights.
By following the application rationalization recommendations above, it was possible to reduce the
number of applications required to 160.

 Technically Challenging: An application should be classified as technically challenging if it is
complex to set up, has extensive dependencies on other applications or requires a specialized
configuration, for example an Electronic Medical Records (EMR) application. Technically
challenging applications need to be identified during the application assessment because they
are not generally appropriate for installation in to a base desktop image or delivery by
application streaming. Delivering technically challenging applications as XenApp published
applications will help to reduce the complexity of the base desktop image. There are three
classifications available for “Technically Challenging” in the Application Assessment
Workbook: Yes, No or Unknown.
 Works Offline: Some user groups may require the ability to work offline. If so, it is important
that the design can determine which applications will work without a network connection and
which ones will not. Applications which require backend infrastructure such as web and
database servers are not typically available offline. There are three classifications available
for “Works Offline” in the Application Assessment Workbook: Yes, No or Unknown.
 Remote Application Servers: For optimal performance, application servers should be hosted
in close proximity to the virtual desktops because this helps to reduce latency and prevent
network capacity from becoming a bottleneck. However, traditional physical desktop
environments sometimes locate application servers in WAN sites so that they are in close
proximity to the desktops. The application assessment should determine whether any of the
application servers are currently hosted outside of the data center so that time is taken during
the design to determine whether it is possible to relocate these servers. There are three
classifications available for “Remote Application Servers” in the Application Assessment
Workbook: Yes, No or Unknown.
 Peripheral Connectivity. If applications require connectivity with peripheral devices, the
application assessment should determine which connectivity interface is required because it
may be necessary to create a policy that maps the required interface from the endpoint device
to the appropriate application. There are five classifications available for “Peripheral
Connectivity” in the Application Assessment Workbook: No, USB, TWAIN, LPT or COM.
 Restricted Access: Application access may need to be restricted due to insufficient licenses /
resources and to protect sensitive data / tools. For example, applications with a limited
number of licenses should not be installed in to a base image that is shared with unlicensed
users. The Application Assessment Workbook allows four options for “Restricted Access” to
be selected:
Restricted Access Requirement
No There are no security restrictions for the application and it can be accessed by any
user within the organization.
User Group The application may be installed on a multi-user operating system but only a
specific group of users should be provided with an icon.
Virtual Machine Application should only be installed on a virtual machine that is accessible by
authorized users.
Table 13: Restricted Access Characteristics

Best Practices
The application assessment portion can be a frightening process due to the sheer number of things to
take into account. However, most applications within an organization‟s complete inventory are fairly
standard. The key to doing an application assessment is to
 Focus on the inventory first
 Reduce the inventory as much as possible
 Classify remaining applications into defined categories
Don‟t forget to follow these top application assessment recommendations from Citrix Consulting, which
are based on years of experience:
1. Remember the Plugins: Plugins are available for various applications, including Microsoft
Internet Explorer, Microsoft Word and Microsoft Outlook. For example Citrix ShareFile and e-
mail archiving solutions are frequently implemented as plugins. To prevent critical functionality
being omitted, plugins should be treated as applications during the application assessment.
2. Application Dependencies: It is imperative that you understand all of the interactions between
your applications. For example, there may be an application which requires Microsoft Word and
Excel be installed on the same system so that reports can be viewed. When it comes to the
design phase of the project you will need to make sure that these applications are grouped
together appropriately.
3. Application Consolidation: It may be tempting to skip through the application consolidation
phase of the assessment, but time spent reducing the number of applications significantly
reduces complexity and time spent on the assessment and design phases of the project.
4. Application Owners: To ensure that you don‟t incorrectly categorize applications or remove an
important application from your inventory, work closely with the various application owners
during the Rationalization, Characteristics and Compatibility Steps.
5. Final Check: Once the consolidated list of applications has been finalized, complete with
characteristics and compatibility information it should be sent to the application owners for
review to ensure that it is correct and no critical applications or information has been omitted.

Project Management
Roadmap
Most companies don‟t have sufficient time or resources to migrate every user group in one go. As
such, it is important that the user groups identified are prioritized so that the business receives the
maximum value from their investment as soon as possible. To achieve this, you need to compare the
business impact and time to value of each group:
 Business Impact: Consider the impact that desktop virtualization will have on each user
group and rank them accordingly. It is important that you double back here, and use the
Business Drivers identified at the start of the project to make sure that you assign an
appropriate ranking. Don‟t just assign ratings based on how highly the business values each
user group; instead focus on which user groups offer the most benefit to the company after
virtualization. It‟s a subtle but important difference.
 Time to Value: For each user group, estimate how long it will take to implement the chosen
FlexCast model based on the findings from the Capabilities Assessment. For example, you
might find that a company already has a XenApp farm that can be leveraged to support those
user groups that require a Hosted Shared Desktop resulting in a low Time to Value.
Alternatively, the business might have no prior experience with XenClient resulting in a longer
time to value. Compare application sets, user requirements and user numbers when
differentiating between user groups that have been assigned the same FlexCast model.
Note: If there are limited skills available in-house to implement a chosen FlexCast model,
consider hiring external resources so that Time to Value can be reduced for the associated
user groups.
Representing this information in a graph provides an easy way to visualize the results:
Figure 4: User Group Prioritization
When it comes to assigning the implementation order, start from the top left hand corner of the chart
and work your way to the bottom right hand corner. This way you start with some quick wins that offer
a high-level of value to the company.
Once the project roadmap has been created, update the project plan so that it incorporates the
prioritized roadmap.
Utility Company: A large utility company realized that there would be a long Time to Value for user
groups that had been assigned with a Hosted VDI FlexCast mode, because they had no prior
experience or training with this technology. To address this concern, the utility company engaged
with Citrix Consulting who provided Consultants with previous experience of successfully deploying
XenDesktop.

Project Team
Desktop virtualization is a fundamental change that requires close collaboration between various
business and technical teams in order to be successful. For example, the virtualization and desktop
teams need to work together to ensure that the virtual desktop image meets user needs while also
being optimized for the data center. Failure to build a cohesive project team that consists of the right
roles and skillsets can negatively impact performance, availability, user experience and supportability
while also increasing costs and risk.
The following tables identify the business and technical roles required during an enterprise virtual
desktop deployment. Although the list may seem quite large, many of these roles are only required for
a short time and multiple roles may be performed by a single person. The project manager and Citrix
Architect are considered to be full time roles with other team members being brought in only when
required. The Project Manager role is key to ensuring that the right people are involved in the project
at the right time.
Business Roles
Role Description Example Responsibilities
Project Sponsor The Project Sponsor is a senior
company executive who recognizes
the benefits that desktop
virtualization will bring to the
business. The Project Sponsor
role is often performed by the Chief
Technology Officer (CTO).
Pre-project
 Promote desktop virtualization within
business
 Identify members of the steering committee
Secure funding
 Assess
 Identify and prioritize key business drivers
Project Manager The Project Manager directs the
project team and is responsible for
ensuring that project objectives are
completed on time and within
budget.
All Steps
 Define key project milestones
 Create and update project plan
 Track progress against plan
 Track expenditure against budget
 Maintain issue and risk register
 Manage scope changes
 Create weekly project reports
 Brief steering committee on progress
 Organize project workshops and meetings
 Ensure project teams are synchronized
 Ensure pre-requisites are in place
 Creates change control requests
Business
Manager
Depending on company structure
and size, Business Managers
oversee planning and performance
at a department, region or
company level. A Business
Manager will understand the
requirements necessary for their
employees to be successful.
Assess
 Assist with application consolidation project
 Provide details on connectivity requirements
of user group, including offline usage
 Provide details on risk tolerance of user
group
 Identify requirements for peripherals
Deploy
 Promote benefits of desktop virtualization
 Assist with coordinating the rollout

Business
Continuity
Manager
The Business Continuity Manager
ensures that an organization can
continue to function after a
disruptive event such as natural
disaster, crime or human/computer
error.
Assess
 Provide Citrix Architect with detailed
understanding of the current business
continuity plan
Design
 Update business continuity plan to
incorporate the new Citrix infrastructure
Deploy
 Test business continuity plan
Test Manager The Test Manager is responsible
for ensuring that the test and user
acceptance environments match
the production environment as
closely as possible. The Test
Manager helps to reduce risk by
ensuring that changes are fully
tested before being implemented in
production.
Assess
understanding of current testing
infrastructure and processes
Design
 Work with Citrix Architect to design an
appropriate testing infrastructure and test
plan for new Citrix environment
Deploy
 Ensure that testing design is implemented
correctly and new Citrix infrastructure is fully
tested before rollout
Application
Owners
An Application Owner is a subject
matter expert on specific
applications deployed within the
business. Application owners are
responsible for ensuring that
problems with the applications are
resolved and that
upgrades/updates are performed
without issue. Application owners
are also responsible for managing
support agreements with the
application vendors.
Assess
 Assist with application consolidation project
 Identify application licensing limitations
 Provide details on security restrictions
 Provide details on application dependencies
 Provide location of backend resources
Deploy
 Provide installation pre-requisites and install
guide
 Assist Citrix team with installing and testing
applications in VDI environment
Service Desk
Manager
The Service Desk Manager helps
to improve productivity and end-
user satisfaction by ensuring that
production issues are logged,
escalated and resolved in a timely
manner. The Service Desk
Manager is also responsible for
reporting on common issues, call
volumes and service desk
performance.
Assess
 Identify common issues with existing
environment
 Provide details on support tools currently
used
Design
 Assist Citrix Architect with designing a
delegated administration model
 Participate in operations and support design
workshops
 Work with Training Manager to identify
training requirements
Deploy
 Monitor helpdesk calls for rollout related
issues

Training
Manager
The Training Manager ensures that
support staff and end-users are
proficient with new areas of
technology. The Training Manager
also has responsibility for ensuring
that the training plan is up-to-date
and followed appropriately.
Assess
 Determine current skill set for support staff
and end users
Design
 Create training plan for support staff and
end users
Deploy
 Implement training plan for support staff and
end users
Communications
Manager
The Communication Manager is
responsible for disseminating key
information throughout the
organization.
Design
 Work with Project Manager to create
communications plan
Deploy
 Relay benefits of desktop virtualization
 Inform users of key migration dates
 Ensure expectations are set accordingly
Table 14: Business Roles

Technical Roles
Role Description Example Responsibilities
Citrix Desktop
Architect
The Citrix Architect will act as the
design authority for all Citrix
products and will liaise with other
architects to ensure that the Citrix
infrastructure is successfully
integrated into the organization.
Assess
 Work with Project Sponsor and Key
Stakeholders to identify and prioritize key
business drivers
 Oversee user segmentation and app.
assessment
 Map FlexCast models to user groups
 Perform capabilities assessment to
determine current state of readiness
 Identify areas of risk and provides remedial
actions
Design
 Create Citrix design which includes
hardware and storage estimates
 Coordinate with other architects to integrate
Citrix infrastructure into organization
 Work with Monitoring Architect to ensure
that Citrix environment is monitored
appropriately
 Create operations and support design
 Create implementation and rollout design
 Create test plan
Deploy
 Ensures that the Citrix environment is
implemented in accordance with design
 Verifies that implementation passes test
plan
 Ensures that the Citrix design is
implemented correctly
Active Directory
Architect
Design authority on Microsoft
Active Directory, including
Organizational Units (OU) and
Group Policy Objects.
Assess
understanding of current Active Directory
architecture
Design
 Work with the Citrix Architect to design OU
structure, group policies, permissions,
service accounts, etc. for new Citrix
environment
 Update Active Directory infrastructure
design to reflect centralization of user data
and accounts
Deploy
 Ensure that Active Directory design is

Virtualization
Architect
Design authority on server and
desktop virtualization using Citrix
XenServer, Microsoft Hyper-V or
VMware vSphere.
Assess
 Provides Citrix Architect with detailed
understanding of current virtualization
architecture
Design
 Works with Citrix Architect to design
hardware, networking, storage, high
availability, etc. for server and desktop
virtualization
that virtualization environment is monitored
appropriately
Deploy
 Ensures that the virtualization design is
Network
Architect
Design authority on networking,
including routing, VLANs, DHCP,
DNS, VPN and firewalls.
Assess
understanding of current networking
architecture
Design
 Work with Citrix Architect to design physical
network, virtual networks, routing, firewalls,
quality of service, remote access, network
optimization, etc. for new Citrix environment
that network is monitored appropriately
Deploy
 Ensure that network design is implemented
correctly
Desktop
Architect
Design authority on Microsoft
desktop operating systems,
including Windows XP and/or
Windows 7.
Assess
understanding of current desktop
environment
Design
 Work with Citrix Architect to design core
desktop virtual image, core applications,
desktop optimizations, etc. for new Citrix
environment
that the virtual desktops are monitored
appropriately
Deploy
 Ensure that desktop design is implemented
correctly

Storage
Architect
Design authority on storage
solutions, including direct-attached
storage, storage-attached networks
and network attached storage.
Assess
understanding of current shared storage
environment
Design
 Work with Citrix Architect to design storage
architecture, tiers, sizing, connectivity, etc.
for new Citrix environment
that storage is monitored appropriately
Deploy
 Ensure that storage design is implemented
correctly
Backup
Architect
Design authority on backup and
recovery, including virtual
machines, desktops, servers, user
data and databases.
Assess
understanding of current backup
architecture and processes
Design
 Work with Citrix Architect and Disaster
Recovery Architect to design backup
architecture, process, schedule, retention,
etc. for new Citrix environment
Deploy
 Ensure that backup design is implemented
correctly
Application
Packaging
Architect
Design authority on packaging
applications for deployment via the
Systems Management Team.
Assess
understanding of current application
packaging process and status
Deploy
 Ensure that all required applications are
packaged according to design
Monitoring
Architect
Design authority on monitoring,
including hardware, network,
servers, storage and security
appliances.
Assess
understanding of current monitoring
architecture and processes
Design
 Work with Citrix Architect to design
monitoring architecture, metrics, alerts, etc.
for new Citrix environment and supporting
infrastructure
Deploy
 Ensure that monitoring design is
 Provide regular reports on capacity and
trends during rollout

Systems
Management
Architect
Design authority on systems
management, including
server/desktop build process,
patching and automated application
installation.
Assess
 Provide Citrix Architect with a detailed
understanding of the current systems
management processes
Design
 Works with Citrix Architect to define
server/desktop build process, patching and
application delivery strategy for new Citrix
environment
Deploy
 Ensure that the systems management
design is implemented correctly
Security
Architect
Design authority on security,
including desktops, servers,
networks and VPNs.
Assess
understanding of current security policy
Design
 Work with Citrix Architect to design security
standards for new Citrix environment,
including authentication, encryption, port
numbers, firewall rules, etc.
Deploy
 Ensures that security design is implemented
correctly
Table 15: Technical Roles

Overview
Designing a desktop virtualization solution is simply a matter of following a proven process and aligning technical
decisions with organizational and user requirements. Without the standardized and proven process, architects
tend to randomly jump from topic to topic, which leads to confusion and mistakes. The recommended approach
focuses on working through five distinct layers:
Figure 5: Five-Layer Design Model
The top three layers are designed for each user group independently, which were identified during the User
Segmentation portion of the assess phase. These layers define the user‟s virtual desktop and how users access
their desktop. Upon completion of these three layers, the foundational layers (control and hardware) are
designed for the entire solution.
This process guides the design thinking in that decisions made higher up impact lower level design decisions.
User Layer
Hardware Layer
Desktop Layer
Control Layer
Access Layer
User Layer
Desktop Layer
Access Layer
User Layer
Desktop Layer
Access Layer
Design

User Layer
The top layer of the design methodology is the user layer, which is defined for each unique user group.
The user layer appropriately sets the overall direction for each user group‟s virtualized environment. This layer
incorporates the assessment criteria for business priorities and user group requirements in order to define
effective strategies for endpoints and Citrix Receiver. These design decisions impact the flexibility and
functionality for each user group.
Endpoint Selection
There are a variety of endpoints devices available, all with differing capabilities, including:
 Tablet based on Android or iOS
 Laptop
 Desktop PC
 Thin Client
 Smartphone
The user‟s primary endpoint device must align with the overall Business Drivers as well as each user‟s role
and associated requirements. In many circumstances, multiple endpoints may be suitable, each offering
differing capabilities.

Decision: Endpoint Ownership
In most organizations, endpoint devices will be corporate owned and managed. However, more and more
organizations are now introducing Bring Your Own Device (BYOD) programs to improve employee
satisfaction, reduce costs and to simplify device management. Even if BYOD is a business priority, it does
not mean that every user should be allowed to use a personal device in the corporate environment.
Certain user requirements, which were identified during the User Segmentation portion of the Assess
phase, can greatly impact the suitability of personal devices:
 Security: Users requiring high security might not be able to bring a personal device into the
secured environment for risk of data theft.
 Mobility: Users operating in a disconnected mode might not be able to use a personal device, as
the Local VM FlexCast option associated with this type of requirement can have specific hardware
requirements, or special maintenance requirements. Additionally, the local operating system may
be destroyed when the Local VM FlexCast option is utilized.
 Criticality: Users with a high criticality rating might require redundant endpoints in the event of
failure. This would require the user to have an alternative means for connecting in the event their
personal device fails, likely making these users poor candidates for a BYOD program.
 FlexCast Model: A personal device should not be recommended for user groups utilizing a client-
hosted FlexCast model like Streamed VHD, Local VM or Remote PC. These FlexCast models
typically require a specific hardware configuration or installation/configuration that will restrict
device selection.
The following diagram provides guidance on when user owned devices could be used:
Figure 6: BYOD or Corporate Device

Decision: Endpoint Lifecycle
Organizations may choose to repurpose devices in order to extend refresh cycles or to provide overflow
capacity for contract workers. Endpoints now offer more capabilities allowing them to have longer useful
lifespans. In many cases, these hardware capabilities vastly outstrip the needs of a typical user. When
coupled with the ability to virtualize application and desktop workloads, this provides new options to
administrators such as repurposing existing workstations. These options go well beyond the simple three-
year PC refresh cycle. However, the benefits of repurposing or reallocating a workstation should be
balanced against the following considerations.
 Minimum Standards: While cost factors of repurposing existing workstations may be compelling,
certain minimum standards should be met to guarantee a good user experience. At a minimum, it
is recommended that repurposed workstations have a 1GHz processor, 1GB of RAM, 16GB of free
disk space and a GPU that is capable of supporting HDX features.
 Business Drivers: Priorities underpin the success of any major project. Those organizations that
have prioritized reducing capital expenditure by means of prolonging the hardware refresh cycle
can benefit from repurposing hardware. Conversely, if an organization‟s Business Drivers include
reducing power consumption as part of an overall green initiative, purchasing newer endpoints may
be beneficial in order to take advantage of the latest generation of power management capabilities
available in the most modern devices.
 Workload: The type of work and FlexCast model for an end user can determine whether they are a
good candidate for a repurposed endpoint, or may be better served with a new device. If the work
performed by the individual involves locally installed applications, the individual may be best served
by a new endpoint that offers the most powerful and recently updated processor and graphics
architecture. However, if a user is largely performing tasks associated with virtualized applications
that do not involve the latest multimedia capabilities (such as webcams, VoIP, media redirection),
then a repurposed workstation should be a viable alternative.
The following planning matrix outlines considerations when repurposing existing hardware:
Endpoint Provisioning Criteria Repurpose Existing Procure New
Capital Restrained Environment 
Heavy Multimedia Usage 
High Failure Rate Among Existing Desktops 
High Number of Virtualized Applications 
Low Intensity Workload 
Outmoded Client Side Feature Set 
Power Consumption or Green Initiative(s) 
Desire to Prolong Hardware Refresh Cycle 
Table 16: Endpoint Procurement Criteria

Decision: Endpoint Form Factor
Capacities of endpoints have grown along with efficiencies offered in thin client form factors. Even mid-
range thin clients now have graphics capabilities that allow utilization of HDX features such as multi-
monitor support while offering management and power efficiency benefits. This expansion of capabilities
has given IT administrators more options and flexibility than ever before.
Most organizations will likely deploy a mixture of fully featured clients as well as thin clients. It is important
to focus on many of the attributes of the user group in order to best determine the type of endpoint that
should be deployed:
Endpoint Mobility FlexCast Business Drivers Security
Thin Clients  Local
 Roaming Local
 Hosted Shared
 Hosted VDI
 On Demand Apps
 Agility
 Better Desktop
Management
 Increased Security
High
Desktop
PCs
 Local
 Roaming Local
 Hosted Shared
 Hosted VDI
 Remote PC
 Streamed VHD
 On Demand Apps
N/A – Existing State Medium
Desktop
PCs with
XenClient
 Local
 Roaming Local
 Offline
 Local VM  Better Desktop
Management
High
Laptops  Local
 Roaming Local
 Remote
 Hosted Shared
 Hosted VDI
 Remote PC
 On Demand Apps
 BYOD
 Work from
Anywhere
Low
Laptops
with
XenClient
 Local
 Roaming Local
 Remote
 Offline
 Local VM  Better Desktop
Management
 Work from
Anywhere
High
Tablets  Local
 Roaming Local
 Remote
 Hosted Shared
 Hosted VDI
 On Demand Apps
 BYOD
 Work from
Anywhere
Low
Table 17: Endpoint Selection

Decision: Thin Client Selection
Thin client vendors now offer a range of operating system choices, including Windows Thin PC (based on
Windows 7), embedded versions of Windows (XP, Windows 7, and Windows 8), Linux variants, as well as
limited functionality clients that boot directly into a virtual desktop and offer zero operating system footprint.
The following factors should be considered during the selection of a thin client device:
 User workload: Windows Thin PC or limited functionality solutions such as Dell Wyse Zero clients
should be tailored to true task workers or knowledge workers with limited needs. More capable
thin devices such as Windows Embedded solutions can be provided for users that require more
graphic capabilities, stronger CPUs for compiling code and other intensive activities.
 In-house Expertise: Many organizations have management toolsets already in place to handle
thin client infrastructure (such as Retailers that may have kiosks). In-house expertise with a thin
client management infrastructure can determine the selection of thin client platform. It is
recommended that existing in-house expertise is leveraged, so long as the platform is capable of
supporting a virtual desktop infrastructure implementation, as outlined on the Citrix Ready site.
 Licensing Cost: There are licensing considerations for most platforms. Windows Thin PC and
embedded versions incur immediate license costs related to Microsoft licensing, whereas a custom
Linux solution may not. However, these costs must be closely balanced against additional add-on
licensing that may be required for Linux devices, which are built into Windows. For example,
various media codecs may require additional license expenditure in a Linux thin client context. For
more information, please refer to Microsoft's Partner Page.
Receiver Selection
Citrix Receiver and associated infrastructure provides a framework for access to enterprise applications,
desktops, and data. There is a Citrix Receiver available for nearly every mainstream platform, with small
differences in capabilities. By and large, most environments should choose the Citrix Receiver available for
the endpoint operating system platform. For organizations desiring specific functionality and more granular
control, there are more options available.
To fully understand the design decisions that are required in an enterprise environment, it is first important to
understand where Citrix Receiver fits into the technical landscape. At its root, Citrix Receiver allows
administrators to implement and deploy various features and capabilities, including the ability to access
enterprise applications, desktops, & data. The following section presents a series of design decisions
encountered when deploying Citrix Receiver in an enterprise.
Decision: Receiver Type
While most organizations should simply deploy the Citrix Receiver compatible with the endpoint, it is
important to recognize that there are certain differences in the feature set of various Citrix Receivers.
Functionality of Receiver can be enhanced if it is able to run plug-ins. In order to have the appropriate
Receiver available, it is important to review the features desired for the user group. The following table
should be referenced to help identify the appropriate Receiver type.
Large Systems Integrator: A large systems integrator recommended that a customer deploy a
single type of low-end, limited capability endpoint for all users. Upon deployment to production,
users immediately complained that they received a poor user experience when viewing multimedia
content over the WAN. At great cost, the systems integrator and customer re-assessed the
environment and chose to deploy endpoints that supported HDX MediaStream. The mistake
caused a schism between systems integrator and the customer, resulting in lost time, capital, and
the end of a business relationship that was fostered over many years. It is critical that the
endpoints assigned to each user group can support their requirements.

Feature
Citrix
Receiver
Citrix
Receiver
Enterprise
Citrix
Receiver
Mac
Citrix
Receiver
Windows 8 R/T
Citrix
Receiver
Linux
Citrix
Receiver
Android
Citrix
Receiver
iOS
Citrix
Receiver
BlackBerry
Citrix
HTML 5
Receiver
Java
Client
Online Plug-in
12.3
XenApp Applications: Accessing
published enterprise applications that
are centralized in the datacenter.
  
◦   
◦ ◦ ◦ 
XenDesktop Virtual Desktops:
Accessing full Windows XP, 7, or 8
virtual desktops.
        
◦ 
Email Based Discovery: Using DNS
configuration options, users simply
provide their email address in order to
fully configure their Citrix Receiver on
their endpoint.
      
Smart Card Support: Ability to use
physical form factor for additional
security during authentication.
          
Fast Access / Proximity Card
Support: A subset of smart cards that
are able to use proximity based
authentication for “touch-and-go”
functionality.
 
Flash Redirection: Using client
resources in order to fetch and render
flash content, reducing server load and
enhancing performance.
  
UDP Audio: Utilizing UDP as a
protocol for audio in order to improve
audio performance over most
networking configurations.
   
Prelaunch: Ability to have applications
launch immediately upon user login to
client, resulting in better user
satisfaction due to perceived
improvements in published application
launch time.
 
Aero Effects: Remoting Aero peek,
preview, & Flip3d effects in a Windows
7 or 8 desktop.
  
Drive Mapping: Local and remote
client drive access from within a
virtualized session.
    
◦ 
Clipboard Mapping: Local and remote
copy and paste mapping while utilizing
a virtualized session.
  
◦
HDX 3dPro Support: Ability to utilize
server side GPU to perform advanced
graphical interface operations in a
Windows virtual desktop.
 
◦ ◦ ◦ ◦
Plug-in Management via
Merchandising Server: Enhancing
client functionality on Windows
platforms by deploying software that
provides specific functionality.
    
ShareFile: Enterprise cloud storage
that can be configured to “follow” user
within Receiver.
    

Access Gateway Remote Access:
Secure access to enterprise resources
without a full VPN client.
      
Offline Applications: Streaming of
applications to endpoint that will then
execute in an isolation environment.
 
Single Sign-On: Centralized credential
management.
 
Start Menu & Desktop Launch:
Launching of published applications
from links that are created in the Start
Menu and Desktop.
  
Notification Area Launch: launching
of published applications from icons
created in the Windows Notification
Area.

Workspace Control: Quickly
disconnecting and re-establishing
connection to existing sessions across
roaming or multiple devices that saves
session initiation and creation time.
    
StoreFront Support: Support for Citrix
StoreFront for application and desktop
aggregation.
      
AppController / SaaS Support:
Support for Citrix AppController as part
of Citrix Mobile Solutions, which allows
Software as a Service (SaaS)
applications and more.
   
Desktop Lock Support: Lockdown
capability for thin devices to provide
extremely limited functionality.

Web Cam Redirection: Redirection of
local webcam to a virtual desktop.
   
◦ ◦
Client Drive Mapping. Makes any
directory mounted on a client device,
available to the user.
     
No Client Install Required: Ability to
access published applications,
desktops, and other functionality
without installing any local software
client.
 
ICA Proxy Capability. Capability to
perform limited security related
functiona (such as SSO pass-through)
without a full tunnel implementation.
       
“”: Functionality available. “◦”: Partial support. “” Requires Administrative Install “”: Applets Must Be Allowed “”: Available with Plug-in
Table 18: Receiver Feature Comparison

Decision: Initial Deployment
There are several deployment options available for delivering Citrix Receiver to an endpoint. Enterprises
must evaluate the endpoint platform and the capabilities of their organization before deciding on the most
appropriate approach. Although it is usually a best practice to have a full version of Receiver deployed to
an endpoint to provide the greatest level of functionality, it is important to consider fallback options such as
the Java Client and the HTML 5 Receiver for those situations where a Receiver client installation is simply
not possible.
The following mechanisms are commonly utilized to deploy and update Citrix Receiver:
 StoreFront: If Citrix StoreFront is available, administrators can deploy Citrix Receiver via a
Receiver for Web site. When deployed, a Receiver for Web site enables users to access
StoreFront stores through a webpage. If the Receiver for Web site detects that a user does not
have a compatible version of Citrix Receiver, the user is prompted to download and install Citrix
Receiver.
 Internal Download Site: Users may be prevented from downloading software from the Internet,
even if they have permission to install applications. Administrator can create an internal website for
supported Citrix Receivers. Templates for such a site are available from the Citrix Downloads site.
When a user accesses a deployed site, it will detect the operating system on the user‟s device and
guide the user to a local Citrix Receiver install suitable for their device.
 Web Interface: Web Interface can be configured to detect whether Citrix Receiver can be
deployed to the endpoint device. E-mail based account discovery is not available when Citrix
Receiver is deployed using Web Interface. E-mail based discovery is particularly useful for BYOD
scenarios.
 Windows Store: The Citrix Receiver for Windows 8/RT is available from the Windows 8 store.
This Receiver is available for ARM or Intel based Windows 8/RT devices only. It supports native
Windows 8 style and gestures.
 Mobile Device: Many mobile devices have unique methods of deploying applications. It simply
may not be possible to deploy Citrix Receiver via traditional methods, therefore the following
options are most likely to be selected for mobile devices:
o Markets & Stores: The latest supported Citrix Receiver is available for Android and iOS
on the deployment portal of choice for these platforms. For Android devices version 2.2
and higher, this will be the Android Play platform. For iOS devices, this will be the Apple
Store.
o Other Mobile Deployment Methods: Many mobile platforms offer proprietary methods of
deploying software. For example, it is possible to utilize BlackBerry Enterprise Server to
deploy the BlackBerry Citrix Receiver 2.2 to supported devices. This is accomplished by
assigning a software configuration. More details can be found in this BlackBerry
Publication.
Furniture Distributor: A furniture distributor maintains a configurator application for various
furniture options. The configurator application is accessed via a limited functionality kiosk that is
deployed at various furniture outlets, including small, independent retailers with little to no IT staff
present. The kiosks are completely locked down in many situations, to the point where even the
running of Java applications is limited. The kiosks do feature a modern browser (Google Chrome),
and therefore, are able to utilize the HTML 5 Receiver in order to provide access to the configurator
application.

 Merchandising Server: Citrix Merchandising server can be utilized to deploy Citrix Receiver. In
addition, Merchandising Server can be configured to update Citrix Receiver and all supported plug-
ins, or a sub-set depending on the needs of an organization. To achieve this, Merchandising
Server periodically communicates with the Citrix Update Service, which maintains a list of
components and their updates.
 Master Image: Most organizations support a limited number of master desktop images which are
deployed to each business owned desktop, laptop, server, or virtual desktop. A common
mechanism to ensure access to virtual desktops and applications is to include a supported version
of Citrix Receiver in the master image. Subsequent updates to Receiver are handled either by
utilizing the Citrix Receiver Updater Plug-in and Merchandising Server, enterprise software
deployment tools, or manually.
 Enterprise Software Deployment: Many organizations employ an enterprise software deployment
(ESD) tool. ESD tools can be used to deploy Citrix Receiver to managed endpoints. ESD tools
cannot be used to deploy Citrix Receiver to unmanaged endpoints, such as user owned devices or
partner devices.
 Group Policy: It is possible to deploy and configure Citrix Receiver via Group Policy. Sample
start-up scripts that deploy and remove Citrix Receiver and Receiver Enterprise are available on
Citrix XenApp and XenDesktop media:
Citrix Receiver and Plug-insWindowsReceiverStartup_Logon_Scripts
 Manual Install: All supported versions of Citrix Receiver are available from the Receiver Download
website. Upon landing on this site, client detection is performed and a platform and operating
system specific link is provided to allow users to download an appropriate edition of Citrix Receiver.
It is important to note that no configuration will be accomplished via this download, so e-mail based
discovery will need to be performed. This option is likely to be utilized in a BYOD environment.
Selecting the appropriate deployment method is based on the Receiver type selected. The following
table should be referenced to help identify the appropriate Receiver deployment options.

Deployment Option Build Into
Base Image
Enterprise
Software
Deployment
Active Directory
& Group Policy
Receiver
for Web
site
Web Interface
site
Internal
Downloads
Site
Windows
App
Store
Mobile
Market or
App Store
Merchandising
Server
Citrix Receiver for Windows    ◦ ◦ ◦ 
Citrix Receiver Enterprise for Windows    ◦ ◦ ◦
Citrix Receiver for Mac ◦ ◦ 
Citrix Receiver for Windows 8/RT ◦ ◦  
Citrix Receiver for Linux     ◦
Citrix Receiver for Android 
Citrix Receiver for iOS 
Citrix Receiver for BlackBerry  
HTML 5 Receiver 
Java Client 
Online Plug-in  
ShareFile Plug-Ins  
Offline Plug-In  
Single Sign-On Plugin  
Access Gateway Plug-in  
Repeater Acceleration Plug-in  
Desktop Lock 
“”: Recommended. “◦”: Available as an option. “”: Option for VDI hosts “”: Proprietary to BlackBerry “”: Possible, but not extensively tested
Table 19: Receiver Deployment Options

Decision: Initial Configuration
Citrix Receiver must be configured in order to provide access to enterprise resources. The method
of configuration varies by Receiver, the form factor of the device, and lastly the access method
(local or remote) that is involved. Several methods may be viable for an organization. The method
utilized is contingent on the resources (people, systems, time) available as well as larger
organizational initiatives such as BYOD programs. The following methods can be used to
configure Citrix Receiver:
 E-mail Based Discovery: The newest Receivers can be configured by simply entering a
company e-mail address. E-mail based Discovery requires Citrix StoreFront as well as an
SRV record which points to the FQDN of the StoreFront server.
Note: Any DNS platform should support e-mail based discovery, however only Windows
DNS has been explicitly tested.
For remote access, an Access Gateway must be utilized with the corresponding SRV
record in DNS. A valid server certificate on the Access Gateway appliance or StoreFront
server must be present in order to enable email-based account discovery.
 Group Policy: Microsoft Group Policy can be used to configure Citrix Receiver. Once
startup scripts have been deployed for Citrix Receiver, edit the corresponding script and
ensure there is a value for the following parameter: SERVER_LOCATION=Server_URL.
The default value is blank. Provide the URL of the server running Web Interface or
StoreFront. The URL must be in the format http://servername or https://servername.
 Provisioning File: For environments running StoreFront, it is possible to provide users
with a provisioning file that contains store information. Provisioning files are exported from
the StoreFront console. The file is saved with a “*.cr” extension and can then be placed on
a shared network resource, a Receiver for Web site, or other web based resource. The file
can then be launched from an endpoint, which automatically configures Citrix Receiver to
use the store.
 Auto-generated Setup URL: The One-Click-Provisioning tool can be used to create a
setup URL for mobile devices and Mac desktops. Administrators provide the requisite
information and the One-Click-Provisioning tool generates the URL. As the URL does not
contain username or password information, it can be emailed to mobile users or even
posted in a shared location. Upon clicking the URL, Citrix Receiver will incorporate the
data into its configuration. This tool is available for remote users, Web Interface, as well as
Services sites.
 Manually: If allowed, it is usually possible to configure Citrix Receiver manually. This
method should be reserved for administrators or users that have advanced knowledge.

Decision: Updates
Citrix Receiver and plug-ins are in active development. Accordingly, periodic updates are released
that provide enhanced functionality or address user issues. As with any actively developed
product, the latest version of these products should be deployed to the endpoints so that users
benefit from the latest functionality. There are multiple methods available to update Citrix Receiver
and, if applicable, associated plug-ins.
 Merchandising Server: For platforms that support the Citrix Receiver Updater Plug-In
(Citrix Receiver, Citrix Receiver Enterprise, and Citrix Receiver Mac), this is the most
preferred mechanism for keeping Citrix Receiver and its associated plug-ins up-to-date.
The Receiver Updater plug-in communicates with Merchandising Server and identifies
package updates for deployment.
 Enterprise Software Deployment: ESD tools provide a viable alternative to
Merchandising Server with Receiver Updater. Additional thought must be given to
updating unmanaged devices and endpoints outside of the corporate firewall, which a
solution that Merchandising Server can address.
 Manual Updates: When no automated solution is available, manual methods can be used
to update Citrix Receiver. Whether deployed on Receiver for Web site, Web Interface, an
internal Citrix Receiver site, or an external site, these options will require user involvement
in updating Citrix Receiver and the associated plug-ins. Due to the involved nature of
manual updates coupled with the opportunity for a user mistake, this option should only be
considered as a last resort.

Resource Requirements
The key to resource allocation is to successfully balance hardware costs against user experience.
Virtual desktops must offer similar levels of performance to the old physical desktops or productivity
and overall user satisfaction will be affected. Conversely, allocating excess resources is inefficient
and expensive for the business.
Processor, memory and storage requirements must be identified for each user group so that an
appropriate hardware specification can be selected and ordered. The resources required for the
Hosted VDI and Hosted Shared FlexCast models will be determined based on the workload
characteristic of each user group, identified during the User Assessment phase:
User Group Description
Light 1-2 office productivity apps or kiosk
Normal 2-10 office productivity apps with light multimedia use
Heavy Multimedia, data processing or application development
OpenGL Heavy performance profile + OpenGL Graphics requirement
Table 20: Workload Characteristics
The scalability estimates provided within this chapter are based on the following assumptions:
Component Assumption
Processor Speed Processor architecture and speed have a direct impact on the number of users that can be
supported. The estimates provided are based on a processor speed of 2.7 GHz and Intel
Westmere processor architecture.
Hyper-Threading Hyper-threading has the potential to improve user density per VM (XenApp only) or VM
density per host (XenApp and XenDesktop). The use of hyper-threading typically provides
a performance boost of between 20-30%, but may also have an adverse effect on
performance in rare cases. Therefore, it is critical to monitor and compare the
performance of workloads with Hyper-threading and without Hyper-threading. The
scalability estimates provided are based on physical cores rather than logical cores and do
not include the benefit from enabling hyper-threading.
Workloads It is assumed that all workloads include antivirus and standard monitoring tools.
Hypervisor
Overhead
The overhead for supporting the hypervisor has been accommodated by reducing the
estimated number of users per core rather than specifically reserving logical cores.
Memory
Utilization
Most users consume more RAM as the day progresses because:
1. Users have more applications open
2. Applications may fail to release memory properly when closed
3. Applications may have memory leaks
Therefore, memory usage should be calculated as a maximum. Even though memory
utilization increases consistently throughout the day, processor activity remains sporadic.
Table 21: Scalability Assumptions
Note: Although the estimates provided offer a good starting point, each implementation will have
different resource requirements and usage patterns. Therefore, either scalability testing or proper
monitoring during a pilot should be completed to generate results that are specific to the environment
and user group requirements.

Decision: CPU
The following table provides guidance on the number of light, normal and heavy users that can be
supported per physical core. The move from dual to quad sockets is not linear and has been
accounted for by a 15% drop in user density.
Workload Hosted VDI:
Windows 7
Hosted VDI:
Windows 7 with
Personal vDisk
Hosted VDI:
Windows XP
Hosted Shared:
2008 R2
Dual Socket
Light 13 11 15 18
Normal 10 8 11 12
Heavy & OpenGL 5 4 8 6
Quad Socket
Light 11 9 12 15
Normal 8 6 9 10
Heavy & OpenGL 4 3 7 5
Table 22: Processor Requirements by Workload
Personal vDisk technology allows users to perform advanced customizations, while still providing
administrators with the benefits of central image management. However, Personal vDisks incur a
processor performance overhead, which varies according to the extent that they are used. Testing
shows that user density can be reduced by 14% when Personal vDisks are used.
Note: Personal vDisks (version 5.6.10 or later) are not supported on Windows XP. Therefore, no
scalability data for this combination has been provided.
The following formula can be used to calculate the CPU resources required per user group:
Decision: Memory
The following table provides guidance on the memory required based on workload classifications:
Windows 7
Hosted VDI:
Windows 7 with
Personal vDisk
Hosted VDI:
Windows XP
Hosted Shared:
2008 R2
Light 1GB 1GB 768MB 340MB
Normal 2GB 2GB 1.5GB 500MB
Heavy & OpenGL 4GB 4GB 2GB 1000MB
Table 23: Memory Requirements by Workload
The following formula can be used to calculate the memory resources required per user group:

Decision: IOPS
The following table provides guidance on the IOPS required based on workload classifications:
Windows 7
Hosted VDI:
Windows 7 with
Personal vDisk
Hosted VDI:
Windows XP
Hosted Shared:
2008 R2 – per User
Installed
Light 8 - 6 3
Normal 16 - 13 6
Heavy & OpenGL 32+ - 26+ 12+
Provisioning Services
Light 5 4 4 2
Normal 10 9 8 4
Heavy & OpenGL 20+ 18 16+ 8+
Machine Creation Services
Light 8 7 6 -
Normal 16 14 13 -
Heavy & OpenGL 32+ 29+ 26+ -
Table 24: IOPS Requirements by Workload
Note: The IOPS estimate is an average of the steady state and is not a peak average that takes
the boot process, logons and logoffs into account.
The following formula can be used to calculate the logical IOPS required per user:
Functional IOPS are I/O operations that are visible from the virtual machine layer. The following
formula can be used to calculate the physical or raw IOPS required:
Note: The write penalty depends on the RAID level used by the storage system on which the
virtual machines are hosted. The following table outlines the key quantitative attributes of the most
commonly used RAID levels:
RAID Level Write Penalty
0 0
1 2
5 4
10 2
Table 25: Write Penalty

Decision: Storage Space
All writes made to a Provisioning Services virtual desktop are stored in a write cache while all
writes made to a Machine Creation Services virtual desktop are stored in a Difference Disk. Writes
include:
 User profile(s)
 Temporary application ad system files
 Windows, Citrix and application logfiles
Disk space requirements for the Write Cache and Difference Disk will depend on application usage
and user behavior. However, the following table provides a starting point for estimating disk space
requirements:
Windows 7
Hosted VDI:
Windows 7 with
Personal vDisk
Hosted VDI:
Windows XP
Hosted Shared:
2008 R2
Provisioning Services – Write Cache
1,2
Light 2GB 12GB
3
2GB
25GBNormal 3GB 15GB
3
5GB
Heavy & OpenGL 5GB 20GB
3
10GB
Machine Creation Services – Difference Disk
4, 5
Light 2GB 12GB
3
2GB -
Normal 5GB 15GB
3
5GB -
Heavy & OpenGL 10GB 20GB
3
10GB -
Table 26: Write Cache / Difference Disk Requirements by Workload
1
The write cache size is minimal following startup but will grow constantly during normal usage
even when deleting files. The write cache will be cleared on reboot, unless write cache persistence
has been configured.
2
In some situations, it may be necessary to redirect data to the write cache disk so that it is
preserved between reboots, for example EdgeSight database, Windows event logs and antivirus
definition files.
3
Includes PVS Write Cache or MCS Difference Disk and 10GB Personal vDisk (default size).
4
NFS based storage is recommended for MCS virtual desktops to ensure that they benefit from
thin provisioning capabilities. Without thin provisioning, each virtual desktop will require a
difference disk that is the same size as the master disk.
5
A new fresh Difference Disk is created each time a non-persistent MCS virtual desktop is
restarted. Once the virtual desktop has successfully started, MCS will remove the old Difference
Disk. The exact amount of additional disk space required depends on the number of concurrent
boot cycles. Based on project experience, a good starting point is to allocate 10% of additional
storage for non-persistent MCS virtual desktops.

Access Layer
The second layer of the design methodology is the access layer, which is defined for each user group.
Creating an appropriate design for the access layer is an important part of the desktop virtualization
process. This layer handles user validation through authentication and orchestrates access to all
components necessary to establish a secure virtual desktop connection.
The access layer design decisions are based on the mobility requirements of each user group as well as
the endpoint devices used.
Decision: Authentication Point
Before a user connects to a virtual resource, they must first authenticate. The place of authentication
is often determined by the user‟s operating location, which was defined in the User Data Capture.
There are four authentication points available:
 Web Interface: The Web Interface provides secure access to XenApp and XenDesktop
resources.
 StoreFront: Citrix StoreFront provides authentication and resource delivery services for Citrix
Receiver, enabling centralized enterprise stores to deliver desktops, applications, and other
resources.
 Secure Gateway (Web Interface): The Secure Gateway is a Windows-based application that
provides a secure Internet gateway between the Citrix virtual desktop infrastructure and user
devices. All data traversing the Internet between a remote workstation and the Secure
Gateway is encrypted using the Secure Sockets Layer (SSL) or Transport Layer Security
(TLS) protocol. It works in conjunction with Web Interface to provide authentication.
 Access Gateway: Access Gateway is an appliance providing secure application access and
granular application-level policy controls to applications and data while allowing users to work
from anywhere.
The following table lists preferred authentication points according to user group‟s mobility
requirements:
User Group‟s Mobility Requirement Preferred Authentication Point
Local StoreFront
Roaming Local StoreFront
Remote Access Gateway
Disconnected Access Gateway
Table 27: Preferred Authentication Point
Authentication for user groups with a mobility requirement of remote or disconnected may occur
directly on the Web Interface where required. For example, DMZ security policies may prohibit access
from the Access Gateway to the domain, which is required to support SmartCard client certificate
authentication. Access to the Web Interface for authentication may be delivered via a NetScaler
SSL_BRIDGE Virtual Server, which provides a conduit for https traffic. Typically, the Virtual Server
would be hosted alongside an Access Gateway on the same NetScaler configured to provide HDX
Proxy access to the Virtual Desktop environment. Never the less, outside of a special use case such
as this, best practice is to authenticate via an Access Gateway for external access.

Decision: Pre-Authentication Policy
An optional pre-authentication policy may be applied to user groups with Access Gateway as their
authentication point (this design decision is not applicable for non-Access Gateway authentication
points). Pre-authentication policies limit access to the environment based on whether the endpoint
meets certain criteria or not identified through Endpoint Analysis (EPA) Scans.
Pre-authentication access policies can be configured to test antivirus, firewall, operating system, or
even registry settings. These policies are used by XenApp and XenDesktop to control features such
as clipboard mapping, printer mapping and even the availability of specific applications. For example,
if a user device does not have antivirus installed, a filter can be set to hide sensitive applications.
The following figure provides an overview of how multiple policies can be used to customize the
features of a virtualization resource:
Figure 7: Simplified Smart Access Decision Logic
Retail: A small private retail company use EPA to scan for the presence of updated Anti-Virus
definitions prior to allowing access.
Financial: A medium financial enterprise use EPA scans of the Domain SID to verify that users
are members of the enterprise domain prior to allowing access.
Government: A large federal institution use EPA to scan endpoint devices to ensure that a
specific certificate (or set of certificates) has been installed on the device prior to allowing
access.

Decision: Authentication Policy
Once the authentication point has been identified, the type of authentication must be determined. The
following options are the primary methods available:
 Web Interface, Secure Gateway (Web Interface), or StoreFront: The StoreFront product is
early in its lifecycle while Web Interface is approaching the end of its lifecycle. Therefore, the
list below focuses on all methods available in StoreFront, which are also available in Web
Interface.
o User Name & Password: Requires users to logon directly to the site by entering a
username and password.
o Domain Pass-through: Allows pass-through of domain credentials from users'
devices. Users authenticate to their domain-joined Windows computers and are
automatically logged on when they access their stores.
o Access Gateway Pass-through: Allows pass-through authentication from Access
Gateway. Users authenticate to Access Gateway and are automatically logged on
when they access their stores.
For details of Web Interface authentication methods not available in StoreFront, please refer
to Citrix eDocs - Configuring Authentication for the Web Interface. Secure Gateway provides
secure access to Web Interface when hosted in a DMZ, yet relies on it to provide
authentication and thus provides the same methods.
 Access Gateway: The NetScaler Access Gateway supports several authentication methods.
The list below includes those primarily used in virtual desktop environments. Each may be
used individually, but are often combined to provide multi-factor authentication.
o LDAP: The Lightweight Directory Access Protocol (LDAP) is a TCP based network
protocol used to access directory information services such Microsoft Active Directory.
Access Gateway uses LDAP to authenticate users and extract their group
membership information.
o RADIUS (aka Token): Remote Authentication Dial In User Service (RADIUS) is a
UDP based network security protocol that provides authentication, authorization, and
accounting. A Network Access Server (Access Gateway in this case) forwards
credentials to a RADIUS server that can either check the credentials locally, or check
them against a directory service. The RADIUS server could then accept the
connection, reject the connection, or challenge and request a second form of
authentication such as a token.
o Client Certificate: Users logging on to an Access Gateway virtual server can also be
authenticated based on the attributes of a client certificate presented to the virtual
server. Client certificates are usually disseminated to users in the form of Smartcards
or Common Access Cards (CACs) that are read by a reader attach to each user‟s
device.
The authentication type for a user group is often determined based on security requirements as well
as the authentication point used. The following table helps define the appropriate solution for each
user group based on the level of security required:

Initial
Authentication
Point
Mobility Anonymous LDAP
Username &
Password
Pass-
through
LDAP +
Token
LDAP +
Smartcard
Token +
Smartcard
Web Interface Local &
Roaming
Local
Low Medium Medium High High High
StoreFront Local &
Roaming
Local
N/A Medium Medium N/A N/A N/A
Secure
Gateway
Remote N/A Low N/A High High High
Access
Gateway
Remote N/A Low N/A High High High
Table 28: Authentication Policy Guidance
Note: Anonymous authentication is only available with the Hosted Shared FlexCast model.
Decision: Session Policy
User groups with Access Gateway as their authentication point must have corresponding session
policies defined (this design decision is not applicable for non-Access Gateway authentication points).
Session policies are used to define the overall user experience.
Organizations create sessions policies based on their Receiver, identified during the design phase.
For the purpose of session policy assignment, devices are commonly grouped as either non-mobile
(such as Windows OS based), or mobile (such as iOS or Android). Therefore a decision whether to
provide support for mobile devices, non-mobile devices, or both should be made based on client
device requirements identified in the Assess section.
Healthcare: A hospital trust provides medical staff with access to Electronic Medical Records
(EMR). Due to the sensitivity of the content, the EMR software uses a high-security
authentication solution integrated into the application. Therefore, to reduce logon times and
authentication points, the application was delivered using XenApp anonymous authentication.
Retail: A small private retail company provides virtual desktop users with access to non-
sensitive data such as marketing catalogs and e-mail. They are not required to adhere to
security regulations that public companies were, such as Sarbanes Oxley. Therefore, LDAP
authentication has been implemented based on username and password.
Financial: A medium financial enterprise provides their virtual desktop users with access to
confidential data such as banking transaction records. They are governed by security
regulations such as the Statement on Accounting Standards (SAS) 70 and are required to utilize
multi-factor authentication for remote access users. LDAP authentication has been
implemented based on username and password along with RADIUS authentication using
tokens.
Government: A large federal institution provides virtual desktop users with access to highly
confidential data such as private citizens‟ personal records. They are subject to regulation by
Department of Defense (DOD) security standards. LDAP authentication has been implemented
based on username and password, along with Client Certificate authentication using CAC cards.

To identify devices session policies include expressions such as:
 Mobile Devices: The expression is set to “REQ.HTTP.HEADER User-Agent CONTAINS
CitrixReceiver” which is given a higher priority than the non-Mobile device policy to ensure
mobile devices are matched while non-Mobile devices are not.
 Non-Mobile Devices: The expression is set to “ns_true” which signifies that it should apply to
all traffic that is sent to it.
For more information, please refer to Citrix eDocs – Receiver and Plug-ins.
An alternative use of session policies is to apply endpoint analysis expressions. These session
policies are applied post authentication yet mimic the previously mentioned Pre-authentication
policies. Use of session policies is an option to provide a fallback scenario to endpoints that do not
meet full security requirements such a read-only access to certain applications.
Decision: Session Profile
Each session policy must have a corresponding session profile defined (this design decision is not
applicable for non-Access Gateway authentication points). The session profile defines details required
for the user group to gain access to the environment. There are two primary forms of session profiles
that determine the access method to the virtual desktop environment:
 SSLVPN: Users create a virtual private network and tunnel all traffic configured by IP
addresses through the internal network. The user‟s client device is able to access permitted
intranet resources as if it were on the internal network. This includes XenApp farms,
XenDesktop sites and any other internal traffic such as file shares or intranet web sites. This
is considered a potentially less secure access method since network ports and routes to
services outside of the virtual desktop infrastructure may be opened leaving the enterprise
susceptible to risks that may come with full VPN access. These risks may include denial of
service attacks, attempts at hacking internal servers, or any other form of malicious activity
that may be launched from malware, Trojan horses, or other viruses via an Internet based
client against vulnerable enterprise services via routes and ports.
Another decision to consider when SSLVPN is required is whether to enable split tunneling for
client network traffic. By enabling split tunneling, client network traffic directed to the intranet
by Receiver may be limited to routes and ports associated with specific services. By disabling
split tunneling, all client network traffic is directed to the intranet, therefore both traffic destined
for internal services as well as traffic destined for the external services (Internet) traverses the
corporate network. The advantage of enabling split tunneling is that exposure of the corporate
network is limited and network bandwidth is conserved. The advantage of disabling split
tunneling is that client traffic may be monitored or controlled through systems such as web
filters or intrusion detection systems.

Figure 8: SSL VPN
 HDX Proxy: With HDX Proxy, users connect to their virtual desktops and applications through
the Access Gateway without exposing internal addresses externally. In this configuration, the
Access Gateway acts as a micro VPN and only handles HDX traffic. Other types of traffic,
such as private mail or personal Internet traffic do not use the Access Gateway.
Based on the Endpoint and Receiver used, a decision must be made as to whether this
method is supported for each user group. HDX Proxy is considered a secure access method
for remote virtual desktop access since only traffic specific to the desktop session is allowed to
pass through to the corporate infrastructure. Most Receivers support HDX Proxy and it is the
preferred method:
Figure 9: HDX Proxy

Decision: Preferred Data Center
Enterprises often have multiple active data centers providing high availability for mission critical
applications. Some virtual desktops or applications may fall into that category while others may only
be accessed from a specific preferred data center. Therefore, the initial Access Gateway that a user
authenticates to in a multi-active data center environment may not be within the preferred data center
corresponding to the user‟s virtual desktop resources. Web Interface or StoreFront is able to
determine the location of the user‟s assigned resources and, direct the HDX session to those
resources.
There are static and dynamic methods available to direct HDX sessions to their virtual desktop
resources in their primary data center. The decision regarding which method to select should be
based on availability of technology to dynamically assign sites links such as Global Server Load
Balancing (GSLB) along with the network assessment of intranet and Internet bandwidth as well as
Quality of Service (QoS) capabilities.
 Static
o Direct: The user may be given a FQDN mapped to an A record that is dedicated to
the primary data center Access Gateway/(s) allowing them to access their virtual
desktop directly wherever they are in the world. This approach eliminates a layer of
complexity added with dynamic allocation. However, it eliminates fault tolerance
options such as the ability to access the virtual desktop through an alternative intranet
path when a primary data center outage is limited to the Internet access infrastructure.
 Dynamic
o Internet: For most dynamic environments, the initial data center selected for
authentication is the one “closest” to the user. Protocols such as GSLB dynamic
proximity calculate the least latency between the user‟s local DNS server and the
Access Gateway. Thereafter, the HDX session may be redirected through an Access
Gateway in the preferred data center by assignment in the Web Interface accordingly.
However, a significant portion of the HDX session would likely be forced to travel over
the “best effort” Internet without QoS guarantees.
For example, a user with a dedicated desktop in the United Stated, traveling in Europe
may be directed to an Access Gateway hosted in a European data center. However,
when the user launches their desktop, an HDX connection will be established to the
virtual desktop via an Access Gateway hosted in the preferred data center in the
United States.

Figure 10: Internet Connection
o Intranet: Alternatively, the HDX session over which their virtual desktop will be
delivered may be transported across the Enterprise WAN and proxied by the
same Access Gateway where they authenticated. The advantage of this
approach is that a significant portion of the HDX traffic would traverse the
Enterprise network where quality of service may be used to provide an optimal
user experience. Therefore, this is the recommended dynamic method provided
the network assessment identifies that sufficient bandwidth is available or that
the project allows for it to be provisioned, and that an appropriate QoS
infrastructure is in place or may be configured. Otherwise the Internet is the
alternative method.
Figure 11: Intranet Connection

Decision: Access Bandwidth
The last access layer decision is to determine the maximum concurrent network bandwidth required by
the virtual desktop user groups. This step will be a key factor in determining which NetScaler Access
Gateway platform is required during the control layer decision process.
The bandwidth utilized per user will depend on their computing requirements. For example, a user
intermittently entering data in an online ordering system would have a light requirement, while a user
that regularly uses Microsoft Office along with various intranet web sites would have a normal
requirement, and a user working on CAD drawings would have a heavy requirement.
Ideally, bandwidth required for specific user groups and use cases would be measured through
network analysis tools. However, in the absence of sufficient data the table below provides an
estimate of what may be expected.
User Group‟s Workload Primary Activity Average Bandwidth (Kbps)
Light Office 43
Normal Internet 85
Heavy Multimedia 174
OpenGL 3D Modeling 1,920
Table 29: Bandwidth Estimates per User
To calculate the total bandwidth for the user group, use the following formula:
Once the bandwidth required for each user group has been determined, calculate the Total
Concurrent Bandwidth required using the formula below:
Note: This calculation should be performed for each data center.
For additional information, please refer to the Citrix Knowledgebase Article: CTX126191 - XenDesktop
Planning Guide: User Bandwidth Requirements.

Desktop Layer
The third layer of the design methodology, and the final layer focused specifically on the user groups, is
the desktop layer.
The overall user acceptance of the solution is defined by the decisions made within the desktop layer.
Personalization, applications and overall desktop image design play a pivotal role in how well the desktop
is aligned with the user group‟s requirements, which were identified within the User Data Capture and
Application Data Capture sections of the assess phase.
Personalization
User Profiles
A user‟s profile plays a critical role in determining how successful the user experience is within a
virtual desktop or virtual application scenario. Even a well-designed virtual desktop solution can fail
if users are frustrated due to lengthy logon times or lost settings.
The user profile solution chosen must align with the personalization characteristics of the user
group captured during the Assess phase (Table 4: Personalization Characteristics) as well as the
FlexCast model selected.
Decision: Profile Type
This section provides an overview on the different profile types available and provides guidance
on the optimal user profile for each FlexCast model.
 Local Profiles: Local profiles are stored on each XenApp server or virtual desktop and
are initially created based on the default user profile. Therefore, a user accessing these
resources would create an independent profile on each system. Users are able to
retain changes to their local profile on each individual system, but changes are only
accessible for future sessions on that system. Local profiles require no configuration; if
a user logging into a XenApp server or virtual desktop does not have a profile path
administratively defined, a local profile is used by default.
 Roaming Profiles: Roaming profiles are stored in a centralized network repository for
each user. Roaming profiles differ from local profiles in that the information in the
profile (whether it is a printer, a registry setting, or a file stored in the Documents folder)
can be made available to user sessions accessed from all systems in the environment.
Configuring a user for a roaming profile requires an administrator to designate the
user‟s Profile Path (for virtual desktops) or Terminal Server Profile Path (for XenApp) to
a particular location on the network. The first time the user logs on to a XenApp server
or virtual desktop, the default user profile is used to create the user‟s roaming profile.
During logoff, the profile is copied to the administrator-specified network location.
 Mandatory Profiles: Mandatory profiles are typically stored in a central location for
many users. However, the user‟s changes are not retained at logoff. Configuring a
user for a mandatory profile requires an administrator to create a mandatory profile file
(NTUSER.MAN) from an existing roaming or local profile, and assign users‟ Terminal
Services profile path to the location where the file can be accessed.

 Hybrid Profiles: Hybrid Profiles are a profile type that combines a robust profile core (a
mandatory profile or a local default profile) with user specific registry keys or files that
are merged during logon. This technique enables administrators to tightly control which
changes are retained and to keep the user profiles small in size. Furthermore, hybrid
profiles address the “last write wins” issue using mature queuing techniques that
automatically detect and prevent simultaneous writes. Thus, user frustration that
results from making profile changes when accessing multiple XenApp servers or virtual
desktops is minimized or eliminated. In addition, they capture and record only the
changes within the profile, rather than writing the entire profile at logoff. A good
example of a hybrid profile solution is Citrix Profile Manager, which will be discussed in
detail within this chapter.
The following table compares the key features of each profile type:
Local Roaming Mandatory Hybrid
Central Management / Roams with user    
User settings are stored persistently    
Granular Configuration    
Logon Performance & Stability
enhancements
   
“”: Functionality available. “”: Optional. ““: Functionality not available.
Table 30: Profile Type Capability Comparison
In order to select the optimal profile type for each user group it is important to understand their
personalization requirements in addition to the FlexCast model which has been assigned.
The following table provides guidance on selecting the optimal user profile type:
User setting persistence required (Personalization Characteristic: Basic / Complete)
Hosted VDI – Random    
Hosted VDI – Dedicated / Static with PVD    
Hosted Shared    
XenClient    
User setting persistence not required or not desired (Personalization Characteristic: None)
Hosted VDI – Random    
Hosted VDI – Dedicated / Static with PVD    
Hosted Shared    
XenClient    
“”: Recommended. “”: Viable. ““: Not Recommended.
“”: Recommended for users who use a single virtual desktop only.
“”: Recommended for users who use more than one virtual desktop.
Table 31: Profile Type Selection

Decision: Folder Redirection
While redirecting profile folders to a network share (i.e. user documents folders) is a good
practice to minimize profile size, architects need to be aware that applications may frequently
read/write data to profile folders such as AppData, causing potential issues with file server
utilization and responsiveness. It is important to thoroughly test profile redirection before
implementation in production to avoid these issues. Therefore, it is important to research profile
read / write activities and to perform a pilot before moving to production. Microsoft Outlook is an
example of one application that regularly performs profile read activities as the user signature is
read from the user profile every time mail is created.
The following table should be referenced to help identify the appropriate folders to redirect:
Application Data    
Contacts    
Desktop    
Downloads    
Favorites    *
Links    
My Documents    *
My Music    
My Pictures    
My Videos    
Saved Games    
Searches    
“”: Recommended. “”: Optional. ““: Not recommended.
“”: Recommended for backup purposes.
“*”: Recommended for simplified backup / restore and data sharing between profiles.
Table 32: Folder Redirection Matrix
Decision: Folder Exclusion
Excluding folders from being persistently stored as part of a roaming of hybrid profile can help to
reduce profile size and logon times. By default Windows excludes the AppdataLocal and
AppdataLocalLow folders, including all subfolders, such as History, Temp, and Temporary
Internet Files. In addition the folders Downloads and Saved Games should also be excluded.
Decision: Profile Caching
Local caching of roaming or hybrid user profiles on a XenApp Server or virtual desktop is default
Windows behavior and can reduce login times and file server utilization / network traffic. With
profile caching, the system only has to download changes made to the profile. The downside of
profile caching is that it can consume significant amounts of local disk storage on multi-user
systems, such as a XenApp Server or a pooled VDI.

Citrix recommends not deleting locally cached profiles for the following scenarios:
 For optimizing logon times
o Hosted VDI – Dedicated / Existing / Physical
o Hosted VDI – Static with PVD
o Hosted VDI – Remote PC
o Local VM
 For optimizing logoff times
o Non-persistent virtual desktops and XenApp servers, with reboot on logoff or
daily reboot.
Citrix recommends deleting locally cached profiles on logoff for the following scenarios, to avoid
the proliferation of stale profiles:
 Hosted Shared provided by persistent XenApp servers.
 Hosted VDI – Pooled without immediate reboot on logoff.
Decision: Profile Permissions
For security reasons, administrators, by default, cannot access user profiles. While this level of
security may be required for organizations that deal with very sensitive data, it is unnecessary
for most environments and can complicate operations and maintenance. Therefore, Citrix
recommends enabling the “Add the Administrators security group to roaming user profiles”
policy setting. It is also recommended to align this decision with the security characteristics of
the user groups captured during the Assess phase.
Decision: Profile Path
Determining the network path for the user profiles is one of the most significant decisions during
a user profile design process. In general it is strongly recommended to leverage a redundant
and high performance file server or NAS device.
In addition, there are three topics that must be considered for the profile path:
 Performance: File server performance is a vital factor for user logon performance. For
large virtual desktop infrastructures, a single file server cluster may not be sufficient to
handle periods of peak activity. In order to distribute the load across multiple file
servers, the file server address and share name will need to be adjusted.
 Location: User profiles are transferred over the network by means of the SMB protocol,
which does not perform well on high-latency network connections. Furthermore, WAN
connections are typically bandwidth constrained, which can add additional delay to the
profile load process. Therefore, the file server should be located in close proximity to
the XenApp servers and virtual desktops to minimize logon times.
 Operating System Platforms: User profiles have a tight integration with the underlying
operating system and it is not supported to reuse a single user profile on different
operating systems or different platforms like 64-Bit (x64) and 32-Bit (x86). Windows
2008 / Vista introduced a new user profile structure (can be identified by .V2 profile
directory suffix), that makes older user profiles (i.e. Windows 2003 / XP based)
incompatible with new operating systems. In order to ensure that a separate profile is
used per platform, the profile directory has to be adapted.

There are two methods that can be used to address these challenges that are based on
Windows built-in technology:
 User Object: For every user object in Active Directory, an individual profile path, which
contains file server name and profile directory, can be specified. Since only a single
profile path can be specified per user object, it is not possible to ensure that a separate
profile is loaded for each operating system platform.
 Computer Group Policy or System Variables: The user profile path can also be
configured by means of computer specific group policies or system variables. This
enables administrators to ensure that a user profile is dedicated to the platform. Since
computer specific configurations affect all users of a system, all user profiles will be
written to the same file server. To load balance user profiles across multiple servers
dedicated XenApp Worker Groups or XenDesktop Desktop Groups have to be created
per file server.
Note: DFS-N combined with DFS-R is not supported for user profiles. For more information,
please refer to the Microsoft Knowledgebase Article KB2533009 - Information about Microsoft
support policy for a DFS-R and DFS-N deployment scenario.
When using Citrix Profile Manager a third option to address these challenges exists:
 User Object Attributes and variables: Citrix Profile Management enables the
administrator to configure the profile path by means of a computer group policy but in
addition to use attributes of the user object in Active Directory to specify the file server
dynamically. In order to achieve this, three steps are required:
1. Create a DNS Alias (e.g. Fileserver1) which refers to the actual file server
2. Populate an empty LDAP attribute of the user object (e.g. l or UID) with the DNS Alias
3. Configure Citrix Profile Manager by means of GPO to use a profile path which refers to
the LDAP attribute (e.g. If attribute UID is used the profile path becomes
#UlD#Profilesprofiledirectory)
In addition, Citrix Profile Manager automatically populates variables to specify the profile path
dynamically based on the operating system platform. Valid profile management variables are:
 !CTX_PROFILEVER! - Expands to v1 or v2 depending on the profile version.
 !CTX_OSBITNESS! - Expands to x86 or x64 depending on the bit-level of the operating
system.
 !CTX_OSNAME! - Expands to the short name of the operating system (e.g. Win7)
By combining both capabilities of Profile Manager, a fully dynamic user profile path can be
created, which can be load balanced across multiple file servers and ensure profiles of different
operating system platforms are not mixed. An example of a fully dynamic user profile path is
shown below:
#UlD#profiles$%USERNAME%.%USERDOMAIN%!CTX_OSNAME!!CTX_OSBITNESS!
Decision: Profile Streaming
Note: The following design decisions are specific to those environments that use Citrix Profile
Management (hybrid profile).
With user profile streaming, files and folders contained in a profile are fetched from the user
store (file server) to the local computer when a user accesses them. During the logon process,
Citrix Profile Management immediately reports that the profile load process has completed
reducing profile load time to almost zero.

Citrix recommends enabling profile streaming for all scenarios. If it is desired to keep a local
cached copy of the user profile, it is recommended to enable the “Always Cache” setting and
configure a size of 0. This ensures that the complete user profile is downloaded in the
background and enables the system to use this cached copy going forward.
Note: Profile Streaming is not required and does not work with the personal vDisk feature of
Citrix XenDesktop. Even if explicitly enabled by means of Group Policy, the Profile streaming
setting is disabled automatically.
Decision: Active Write Back
By enabling the Active Write Back feature, Citrix Profile Manager detects when an application
has written and closed a file, and copies the file back to the network copy of the profile during
idle periods. In scenarios where a single user leverages multiple virtual desktops or hosted
shared desktops simultaneously, this feature can be tremendously beneficial. However, Citrix
Profile Management does not copy any registry changes back to the network, except during an
ordered logoff. As such, there is a risk that the registry and files may get out of alignment on
provisioned systems, where locally cached profile information is wiped upon reboot. Therefore,
it is recommended to disable the Active Write Back functionality for non-persistent Provisioning
Services or Machine Creation Services scenarios.
Decision: Configuration Approach
Citrix Profile Manager can be configured by means of an “.ini” file as well as Microsoft Group
Policy. While both options offer the same configuration settings, it is recommended to configure
Profile Manager by means of Group Policy, since this enables the administrators to perform
configurations from a single point rather than individually on every system.
User Policies
Citrix policies provide the basis to configure and fine tune XenDesktop and XenApp environments,
allowing organizations to control connection, security and bandwidth settings based on various
combinations of users, devices or connection types.
When making policy decisions it is important to consider both Microsoft and Citrix policies to ensure
that all user experience, security and optimization settings are considered. For more details on
specific Windows related policies, please refer to the Microsoft whitepaper - Group Policy Settings
Reference for Windows and Windows Server, specifically settings related to Windows Server 2008
R2 and Windows 7.
Note: For a list of all Citrix-related policies, please refer to the Policy Quick Reference within the
Appendix.
Decision: Preferred Policy Engine
With current versions of XenDesktop and XenApp, organizations have the option to configure
Citrix policies via the Citrix administrative consoles; AppCenter for XenApp or Desktop Studio
for XenDesktop, or through Active Directory group policy using Citrix ADMX files, which extend
group policy and provide advanced filtering mechanisms.
Using Active Directory group policy allows organizations to manage both Windows policies and
Citrix policies in the same location, and minimizes the administrative tools required for policy
management. Group policies are automatically replicated across domain controllers, protecting
the information and simplifying policy application.

Citrix administrative consoles should be used if Citrix administrators do not have access to
Active Directory policies. Architects should select one of the above two methods as appropriate
for their organization‟s needs and use that method consistently to avoid confusion with multiple
Citrix policy locations.
Decision: Policy Integration
When configuring policies, organizations will often require both Active Directory policies and
Citrix policies to create a completely configured environment. With the use of both policy sets,
the resultant set of policies can become confusing to determine. In some cases, particularly
with respect to Windows Remote Desktop Services (RDS) and Citrix policies, similar
functionality can be configured in two different locations. For example, it is possible to enable
client drive mapping in Citrix policy and disable client drive mapping in RDS policy. The ability
to use the desired feature may be dependent upon the combination of RDS and Citrix policy. It
is important to understand that Citrix policies build upon functionality available in Remote
Desktop Services. If the required feature is explicitly disabled in RDS policy, Citrix policy will
not be able to affect a configuration as the underlying functionality has been disabled.
In order to avoid this confusion, it is recommended that RDS policies only be configured where
required and there is no corresponding policy in the XenDesktop or XenApp configuration, or
the configuration is specifically needed for RDS use within the organization. Configuring
policies at the highest common denominator will simplify the process of understanding resultant
set of policies and troubleshooting policy configurations.
Decision: Policy Filtering
Once policies have been created, they need to be applied to groups of users and/or computers
based on the required outcome. Policy filtering provides the ability to apply policies against the
requisite user or computer groups. With Active Directory based policies, a key decision is
whether to apply a policy to computers or users within site, domain or organizational unit (OU)
objects. Active Directory policies are broken down in to user configuration and computer
configuration. By default, the settings within the user configuration apply to users who reside
within the OU at logon, and settings within the computer configuration are applied to the
computer at system startup, and will affect all users who logon to the system. One challenge of
policy association with Active Directory and Citrix deployments revolves around three core
areas:
 Citrix Specific Computer Policies: Citrix XenApp servers and virtual desktops often
have computer policies that are created and deployed specifically for the XenDesktop
or XenApp environment. Applying these policies is easily accomplished by creating
separate OU structures for the XenApp servers and the virtual desktops. Specific
policies can then be created and confidently applied to only the computers within the
OU and below and nothing else. Based upon requirements, virtual desktops and
XenApp servers may be further subdivided within the OU structure based on server
roles, geographical locations or business units.

 Citrix Specific User Policies: When creating policies for XenDesktop and XenApp
there are a number of policies specific to user experience and security that are applied
based on the user‟s connection to the Citrix environment. However, the user‟s
accounts could be located anywhere within the Active Directory structure, creating
difficulty with simply applying user configuration based policies. It is not desirable to
apply the Citrix specific configurations at the domain level as the settings would be
applied to every system any user logged on to. Simply applying the user configuration
settings at the OU where the XenApp servers or virtual desktops are located will also
not work, as the user accounts are not located within that OU. The solution is to apply
a loopback policy, which is a computer configuration policy that forces the computer to
apply the assigned user configuration policy of the OU to any user who logs onto the
server or virtual desktop, regardless of the user‟s location within Active Directory.
Loopback Processing can be applied with either Merge or Replace settings. Using
Replace overwrites the entire user GPO with the policy from the XenApp or
XenDesktop OU. Merge will combine the user GPO with the GPO from the XenApp or
XenDesktop OU. As the computer GPOs are processed after the user GPOs when
merge is used, the Citrix related OU settings will have precedence and be applied in the
event of a conflict.
 Active Directory Policy Filtering: In more advanced cases, there may be a need to
apply a policy setting to a small subset of users (e.g. Citrix administrators). In this case,
Loopback Processing will not work as the policy is intended to be applied only to the
subset of users, not all users who log in to the system. Active Directory policy filtering
can be used to specify specific users or groups of users to which the policy is applied.
A policy can be created for a specific function, and then a policy filter can be set to
apply that policy only to a group of users such as Citrix administrators. Policy filtering is
accomplished using the Security properties of each target policy.
Citrix policies created using the Citrix administrative consoles in either XenDesktop or XenApp
have specific filter settings available, which may be used to address policy-filtering situations
that cannot be handled using group policy. Citrix policies may be applied using any combination
of the following filters:
Filter Name Filter Description Policy Scope
Access Control Applies a policy based on access control conditions
through which a client is connecting. For example,
users connecting through a Citrix Access Gateway
can have specific policies applied.
User policies
Branch Repeater Applies a policy based on whether or not a user
session was launched through Citrix Branch
Repeater.
User policies
Client IP Address Applies a policy based on the IPv4 or IPv6 address
of the user device used to connect the session.
Care must be taken with this filter if IPv4 address
ranges are used in order to avoid unexpected
results.
User policies
Client Name Applies a policy based on the name of the user
device used to connect the session.
User policies
Desktop Group Applies a policy based on the desktop group
membership of the desktop running the session
XenDesktop user or
machine policies
Desktop Type Applies a policy based on the type of machine
running the session. For example, different policies
can be set depending upon whether a desktop is
pooled, dedicated or streamed.
XenDesktop user or
machine policies

Organizational
Unit
Applies a policy based on the OU of the desktop
running the session.
XenDesktop user or
machine policies
Tag Applies a policy based on any tags applying to the
desktop running the session. Tags are strings that
can be added to virtual desktops in XenDesktop
environments that can be used to search for or limit
access to desktops.
XenDesktop user or
machine policies
User or Group Applies a policy based on the Active Directory group
membership of the user connecting to the session.
User policies
Worker Group Applies a policy based on the worker group
membership of the server hosting the session.
XenApp user or computer
policies
Table 33: Citrix Policy Filters
Decision: Policy Precedence
With the tree-based structure of Active Directory, policies can be created and enforced at any
level in the tree structure. As such, it is important to understand how the aggregation of
policies, known as policy precedence flows in order to understand how a resultant set of policies
is created. With Active Directory and Citrix policies, the precedence is as follows:
 Processed first/lowest precedence: Local server policies
 Processed second: Citrix policies created using the Citrix administrative consoles
 Processed third: Site level AD policies
 Processed fourth: Domain level AD policies
 OU based AD policies
o Processed fifth: Highest level OU in domain
o Processed sixth and subsequent: Next level OU in domain
o Processed last/highest precedence: Lowest level OU containing object
Policies from each level are aggregated into a final policy that is applied to the user or
computer. In most enterprise deployments, Citrix administrators do not have rights to change
policies outside their specific OUs, which will typically be the highest level for precedence. In
cases where exceptions are required, the application of policy settings from higher up the OU
tree can be managed using Block Inheritance and No Override settings. The Block Inheritance
setting stops the settings from higher-level OUs (lower precedence) from being incorporated
into the policy. However, if a higher-level OU policy is configured with No Override, the Block
Inheritance setting will not be applied. Given this, care must be taken in policy planning, and
available tools such as the Active Directory Resultant Set of Policy tool or the XenDesktop
policy planning feature should be used to validate the observed outcomes with the expected
outcomes.

Decision: Baseline Policy
A baseline policy should contain all common elements required to deliver a high definition
experience to the majority of users within the organization. A baseline policy creates the
foundation for user access, and any exceptions that may need to be created to address specific
access requirements for groups of users. It should be comprehensive to cover as many use
cases as possible and should have the lowest priority, for example 99 (a priority number of “1” is
the highest priority), in order to create the simplest policy structure possible and avoid difficulties
in determining the resultant set of policies. The unfiltered policy set provided by Citrix as the
default policy may be used to create the baseline policy as it is applied to all users and
connections. In the baseline configuration, Citrix policies should be enabled with default
settings in order to clearly identify the policies applied, and to avoid confusion should default
settings change over time.
A baseline policy configuration should also include Windows policies. Windows policies reflect
user specific settings that optimize the user experience and remove features that are not
required or desired in a XenDesktop or XenApp environment. For example, one common
feature turned off in these environments is Windows Update. In virtualized environments,
particularly where desktops and XenApp servers may be streamed and non-persistent,
Windows update creates processing and network overhead, and changes made by the update
process will not persist a restart of the virtual desktop or application server. Also in many
cases, organizations use Windows Software Update Service (WSUS) to control Windows
updates. In these cases, updates are applied to the master disk and made available by the IT
department on a scheduled basis.
In addition to the above considerations, an organizations final baseline policy may include
settings specifically created to address security requirements, common network conditions, or to
manage user device or user profile requirements. These areas need to be addressed both in
the default baseline policy configuration, as well as in any additional policy sets created to
address exceptions or additional needs.
 Security Policies: Security policies address policy decisions made to enforce
corporate security requirements on the XenDesktop or XenApp environments.
Requirements pertaining to data security and access can be controlled by the correct
application of security policy. Users can be allowed to read and write to local or
removable media, connect USB devices such as storage devices, smart phones, or
TWAIN compliant devices, or cut and paste from the local system based on security
requirements. Organizations can also enforce encryption and authentication
requirements through security related Citrix policies. Architects should consider the
most appropriate level of security and add the policy settings to the baseline policy set,
and then address security exceptions through additional policy sets.
 Connection-based Policies: Connection based policy considerations are used to
develop a policy solution that creates the best user experience based on the network
environment through which end-users access the network infrastructure. Latency and
available bandwidth will determine how to best provide access to audio and video over
the HDX connection, providing the best quality experience based on the available
resources. Image quality and compression, audio quality and video frame rates can be
adjusted based on the connection quality to utilize the bandwidth and network
performance appropriately. Multi-stream ICA features can be utilized in concert with
network Quality of Service (QoS) to provide an optimized experience for multimedia,
input and display and printing requirements. As with security policies, architects should
consider the appropriate base network configuration and add the settings to the initial
baseline configuration. Additional network requirements can be dealt with by creating
additional higher level policies to override baseline configurations.

 Device-based Policies: Device based policy configuration deals with the management
of specific device requirements such as tablets and smart phones within an
organization. Citrix has created a set of policies to optimize the experience of tablets
and smart phones when connecting to XenApp environments, allowing these devices to
use location services and to customize the user interface where appropriate.
Multimedia specific features, such as Windows Media and Flash redirection will
automatically drop back from client side redirection to server side rendering of media
content if the device does not support it; therefore no specific configuration is required
to address these features with tablets, or with other devices such as thin clients that
may not support these features. Another consideration for device based policy
configuration revolves around the security requirements for bring your own (BYO)
devices. These elements, such as the need to allow or prohibit local access to hard
drives or removable devices, should be addressed through security policy settings.
 Profile-based Policies: User profiles play a critical role in determining how successful
the user experience is within a virtual desktop or virtual application scenario. User
profile management can be a key player in mitigating the risks of lengthy logon times or
lost settings, providing a consistent user experience across multiple devices, and
providing users with their specific data and settings in a virtualized environment. With
Citrix Profile Management (UPM), policies control two important aspects of user
profiles; folder redirection, handled through AD group policy, and UPM settings through
Citrix policy. There is more to configuring UPM than simply turning the features on via
Citrix policy. Architects must consider the correct folder redirection configuration for
their environment, as well as configuring Citrix policy settings for folder exclusions from
the UPM environment. Settings for profile streaming and active write back must also be
carefully considered based on the size of the profile and whether the virtual desktops or
application servers are persistent or non-persistent respectively. Profile management
policies should be included in the baseline policy if they are to be applied across all
users in an organization.
Note: A Baseline Policy Reference is provided within the appendix that includes a baseline
recommendation for Citrix User Policy, Citrix Computer Policy, Microsoft Windows Policy and
Folder Redirection Policy.
Printing
Printing is a very important aspect of every Citrix infrastructure that is not very well understood and
a common cause of issues. Citrix XenApp and XenDesktop offer a variety of features that enable
administrators to successfully integrate printing into almost every scenario. Choosing the most
appropriate printing configuration for your organization helps to simplify administration and improve
user experience.
The printing architecture should align with the mobility characteristics of each user group, as
defined in Table 3: Mobility Characteristics:
This chapter focuses on those technologies and strategies required for printing in online scenarios,
which leverage XenApp and/or XenDesktop. For offline scenarios, which make use of XenClient,
standard Microsoft Windows printing techniques should be used.
Decision: Printer provisioning
For a computer to process a print command, it needs both the required printer object and a
printer driver. Because user sessions in a Citrix environment are hosted in a virtual workspace
instead of locally on the workstation, printers and their drivers do not have to be stored on the
local computer. Instead, during connection to a virtual resource, printers are created/restored
for the user on the virtual desktop. The process by which XenApp / XenDesktop makes printers
available in a session is known as provisioning.

Printer provisioning can be controlled by the administrator. The configuration applied affects the
printers available to users within a session and the printer performance. There are two types of
printer provisioning:
 Static: Static printers are provisioned once (manually or by means of a script) and
stored within the user profile. After that, they are always created in sessions with the
same properties and do not vary according to policies. This is typically used in small
environments.
 Dynamic: The printers that are available in a session are determined during session
initiation. As a result, they can vary according to changes in policy, user or network
location, or end-point device capabilities. When printers are provisioned dynamically,
the printers that appear in a session are not predetermined and stored. Rather, the
printers are assembled, based on policies, as the session is initiated. This process is
also called auto-creation. There are two different types of auto-creation:
o Auto-Creating Client Printers: At the start of a user session, XenApp /
XenDesktop auto-creates all printers available on the client device by default.
Locally attached printers (i.e. USB) as well as network-based printers (i.e. via
print server) can be connected. This process is also referred to as local printer
mapping. The Citrix policy setting “Auto-create client printers” enables
administrators to control what types of printers are provisioned to users or to
prevent auto-creation entirely. The policy setting allows administrators to auto
create all printers, all local printers (non-network) or the client‟s default printer
only.
Note: Users can also connect to client printers manually (if not explicitly
prohibited by policy) from within their session.
The Citrix Universal Printer should be considered in cases where only the
default printer will be created. The Citrix Universal Printer is a generic printer
created at the beginning of a session, and is not tied to a specific printing
device. It does not require the enumeration of the available client printers
during logon. This can greatly reduce bandwidth usage and decrease user
logon times. The Universal Printer can be used with almost any client-side
printing device.
Note: The Universal Printer leverages the Citrix Universal Print Driver, which is
compatible with Windows end-points only.
o Auto-Creating Session Printers: XenApp and XenDesktop provide
administrators with the ability to specify which network printers should be
created within each session. This process can be controlled by means of the
“Session Printers” policy. Network printers created with the Session Printers
setting can be assigned according to where the session was initiated from by
filtering on objects such as IP subnet or client name. This feature enables
administrators to control the assignment of network printers so that the most
appropriate printer is presented to the user, based on the location of the client
device (also known as Proximity Printing).

The following table provides guidance on selecting an optimal method for printer provisioning
based on user mobility:
Static Auto-Creating
Client Printers
Universal Printer Auto-Creating
Session Printers
User leverages thin clients only
Local    
Roaming Local    
Roaming    
User leverages fat clients only
Local    
Roaming Local    
Roaming    
User leverages thin and fat clients
Roaming Local   (hybrid)   (hybrid)
Roaming   (hybrid)   (hybrid)
“”: Recommended. “”: Viable. ““: Not Recommended. “ (hybrid)”: Mix of two methods recommended.
Table 34: Printer Provisioning Method Selection
As outlined within the table, the recommended printer provisioning strategy will vary depending
on the mobility requirements and endpoint device of each user group. Multiple strategies will be
required for most enterprise class environments.
Decision: Printer Property Retention
By default, XenApp and XenDesktop attempt to store printing properties, a combination of the
user‟s printing preferences and printing device-specific settings, on the client device, unless the
client does not support this operation, in which case the properties are stored in the user‟s
profile. To successfully store the printing properties on the client device a recent version of
Citrix Receiver (Citrix Presentation Server Client 9.x and higher) must be installed. For most
scenarios the default behavior is sufficient, but for scenarios listed below a reconfiguration of
the behavior may be required:
 Mandatory User Profile: In scenarios where mandatory user profiles are used, the
user‟s printer properties must be stored on the client device, in order to retain them.
 Multiple Farms/Sites: In scenarios where users connect to multiple XenApp /
XenDesktop environments in which separate user profiles are used, the only way to get
consistent printing behavior is to save the printer properties on the client device.
 Type of Workers: For roaming users, which use roaming or hybrid user profiles, the
user‟s printing preference should be stored within the user profile.
If none of these scenarios apply, Citrix recommends the default setting, which saves the printer
properties on the client device unless the client does not support this operation. This is the
easiest way to ensure consistent printing properties. The Printer Properties Retention policy
rule can be used to force properties to be saved on the client or within the user profile.

Decision: Print Job Routing
In a XenApp / XenDesktop environment, print jobs can take two paths to a printing device:
through the client or through a network print server. The default path will depend on the kind of
printer used:
 Locally Attached Printers: XenApp / XenDesktop routes jobs to locally attached
printers from the server, through the client, and then to the print device. The ICA
protocol compresses the print job traffic. When a printing device is attached locally to
the client device, print jobs must be routed through the client.
Figure 12: Locally attached printers
 Network-based Printers: By default, all print jobs destined for network printers route
from the server, across the network, and directly to the print server. However, if the
XenApp server or the virtual desktop and the print server are on different domains
(without trusts), the native printer driver is not available on the virtual desktop or
XenApp server, or direct network communication is not possible, print jobs are
automatically routed through the client.
Figure 13: Network-based printers – local print server
This fallback behavior, which can also be modified by means of the “Direct connections to print
servers” policy, can cause significant network overhead if not configured correctly. In a
scenario with a data center based print server where print traffic for a network-based printer
needs to be routed through the client, a print job may traverse the WAN three times.

Figure 14: Network-based printers – central printer server
Optimal printing performance can be achieved when the XenApp server / virtual desktop and
the print server are located within the same data center and the print device is located within
close proximity. In scenarios with remote branch offices the routing of the print jobs must be
planned carefully. While routing jobs through a network print server is ideal for fast local
networks, it is not optimal for WANs. The spooling of print jobs using the network printing
pathway method uses more bandwidth than using the client pathway; many packets are
exchanged between the host server / virtual desktop and the print server. Consequently, users
might experience latency while the print jobs are spooling over the WAN. Also, the print job
traffic from the server / virtual desktop to the print server is not compressed and is treated as
regular network traffic.
Therefore, Citrix recommends that print jobs are routed through client devices when printing
across a network with limited bandwidth ensuring that the ICA protocol compresses the print
jobs and enabling the administrator to limit the maximum consumable bandwidth. If this is not
possible (e.g. when a thin client is used without printing capabilities), Quality of Service should
be configured in order to prioritize ICA/HDX traffic and ensure a good in-session user
experience.
The following diagram visualizes the design decision tree related to print job routing:

Figure 15: Print job routing
Decision: Printer Driver Management
In order to print on a print device, it is required to install the respective printer driver. If the print
device is changed, a new driver or an updated version of the driver needs to be installed. In
large organizations with hundreds of printers, managing printer drivers can become complex.
Five strategies for managing printer drivers exist:
1. Automatic Installation of Printer Drivers: When connecting a printer within a user
session (auto-creating or manually), XenApp / XenDesktop will check if the required
printer driver is installed on the XenApp server or virtual desktop. If this is not true the
native printer driver will be installed automatically by default. In a hosted private
desktop (XenDesktop) or hosted shared desktop (XenApp) scenario where users can
easily roam between multiple end-points or locations (e.g. office and home office), the
print devices used by a user may change very frequently. This could potentially cause
hundreds of printer drivers to be installed on every hosted private or shared desktop,
since users may be load balanced to a different hosted resource every time they
connect. Within such a scenario, troubleshooting printing problems and maintaining
existing drivers can be very complicated.
2. Managing Printer Drivers via ESD or Single Image Management: It is a common
practice within enterprise class organizations to leverage existing Enterprise Software
Deployment (EDS) tools or single image management techniques, such as Provisioning
Services or Machine Creation Services to deploy printer drivers to XenApp servers /
virtual desktops. The advantage of this method is that the order of installation of the
drivers can be controlled. This method can be used in combination with Printer Driver
Mapping in order to keep the number of installed printer drivers to a minimum.
3. Printer Driver Mapping: When printer drivers installed on a XenApp server or virtual
desktop are the same as the client but the drivers are named differently (e.g. “HP
LaserJet 4L” versus “HP LaserJet 4”), XenApp and/or XenDesktop may not recognize
that the drivers are the same and users will have difficulty printing or printer auto-
creation may fail. Overriding or mapping the printer driver name for a substitute driver
can resolve this issue.
4. Printer Driver Replication (XenApp only): It is vital to maintain a consistent set of
drivers across all servers within a farm in order to simplify support and maintenance of
the environment and to ensure there is a consistent user experience. XenApp includes
a printer driver replication feature, which enables administrators to define a template
server and deploy installed printer drivers in an automated manner to other servers
within the same farm. This technology is only recommended for environments with a
low number of servers since the driver replication process can add significant overhead
to all XenApp servers in a large farm.
5. Universal Printer Driver: In order to simplify printing in XenApp / XenDesktop
environments, Citrix has developed the Citrix Universal Printer Driver (UPD). The UPD
is a device independent print driver, which has been designed to allow printing on any
printer. Consequently, it simplifies administration by reducing the number of native
drivers required or eliminates the need completely.
The Citrix UPD consists of two components:
a. Server Component: The Citrix Universal Printer driver is installed during the
XenApp server or XenDesktop VDA installation process. When a print job is
initiated the driver records the output of the application and sends it, without
any modification, to the end-point device via the ICA/HDX connection.
b. Client Component: The client component of the Citrix UPD is installed as part
of the Citrix Receiver installation. It fetches the incoming print stream for the

XenApp server / virtual desktop and forwards it to the local printing sub-system
were the print job is rendered using the device specific printer driver.
The Citrix UPD supports any print device, advanced printer functionality (e.g. stapling and
sorting) and does not limit resolution or color depth.
Due to its architecture and the way it integrates with the Windows Print Sub-System the
EMF/XPS Citrix Universal Drivers are compatible with Windows 32/64-Bit based end-
points only. For non-Windows clients the “HP Color LaserJet 2800 Series PS” driver is
provided. This driver is installed along with the Universal Print Driver and provides basic
printing functionality for a wide range of print devices. In contrast to the EMF/XPS UPD
this driver does not support advanced printer functions and limits the resolution of
printouts to 600dpi.
The following table provides guidance on selecting the optimal strategy for printer driver
management.
Automatic
Installation
ESD or Single
Image
Management
Printer Driver
Mapping
Printer Driver
Replication
(XenApp only)
Universal
Printer Driver
Printing infrastructure is very static. Very few new printer models are added each year.
Low number of
printer drivers
    
More than 10
different printer
drivers
   
Printing infrastructure is very dynamic. More than 10 new printer models are added each year.
Low number of
printer drivers
   
More than 10
different printer
drivers
   
“”: Recommended. “”: Viable. ““: Not Recommended.
“”: Viable Option in order to keep the number of native drivers to a minimum.
“”: Viable Option for scenarios with a very low number of XenApp servers only.
Table 35: Printer Driver Management Selection
Decision: Citrix Universal Print Server
The Citrix Universal Print Server (UPServer) extends XenApp and XenDesktop universal
printing support to network printing. This feature eliminates the need to install numerous
network printer drivers on XenApp servers and XenDesktop virtual desktops, and enables more
efficient network utilization. The new Citrix Universal Printer Driver supports direct network
printing to both Windows and non-Windows clients. The Universal Print Server feature is
comprised of the following components:
 Client Component: This component is installed on a XenApp server or XenDesktop
virtual desktop. It receives the EMF or XPS based print stream from the Citrix UPD and
forwards it to the print server. Both the print commands and print data are sent over the
wire using their own respective ports. Defaults are TCP 8080 for the print commands
and TCP 7229 for the print data.
 Server Component: The Citrix UPServer component is installed on a Windows based
print server. It retrieves the print data and forwards it to the respective printer by means
of the Citrix UPServer Virtual Port Monitor.

The Citrix Universal Print Server is highly recommended for remote print server scenarios.
Because the print job is transferred over the network in a compressed format, the utilization of
the network is minimized and the user experience is improved. In addition, the Universal Print
Server enables administrators to keep the number of print drivers installed on a XenApp server
and/or XenDesktop virtual desktop to a minimum, since all network printers will leverage the
Citrix Universal Printer Driver. This helps to reduce potential management and stability issues.
Best Practices
experience:
1. Kernel-Mode Drivers: Windows printer drivers are written in either user mode (also
called version 3 drivers) or kernel mode (also called version 2 drivers). In Windows
NT 4.0, drivers were moved into kernel mode to improve performance. However,
when a kernel-mode driver fails, it can crash an entire system, whereas the failure of
a user-mode driver causes only the current process to crash. Because of this
difference, and because performance enhancements were made in Windows 2000
and Windows Server 2003, native drivers on Windows 2000 and later run in user
mode. Windows Server 2003/2008 can still run kernel-mode drivers, although this is
not recommended due to the impact on system stability.
For some legacy printers only kernel mode (type 2) drivers exist. In such scenarios,
administrators should verify if printer driver mapping, the Citrix Universal Printer
Driver or the Universal Print Server can be leveraged in order to avoid installing the
legacy driver on a XenApp server / virtual desktop.
2. EMF Universal Driver vs. XPS Universal Driver: Although native XPS outperforms
native EMF, this is not true for the Citrix Universal Printer Drivers. In this case the
Citrix EMF Universal Printer Driver is superior to its XPS variant because of
advanced print stream optimization algorithms which are not available for XPS at
this point in time. Furthermore, the output of print jobs using the EMF UPD will be
started as soon as the first page has been transferred to the client. When using the
XPS UPD the output will be delayed until the full print job has been transferred.
Therefore, it is recommended that the Citrix EMF Universal Printer Driver be
leveraged whenever possible.
Personal vDisk
Unlike traditional VDI deployments involving pooled desktops, where users lose their
customizations and personal applications when the administrator alters the base VM, deployments
using personal vDisks retain those changes. This means administrators can easily and centrally
manage their base VMs while providing users with a customized and personalized desktop
experience. Personal vDisks provide this separation by redirecting all changes made on the user's
VM to a separate disk (the personal vDisk) attached to the user's VM. The content of the personal
vDisk is blended at runtime with the content from the base VM to provide a unified experience. In
this way, users can still access applications provisioned by their administrator in the base VM.

Personal vDisks are available in all versions of XenDesktop 5.6 using either Provisioning Services
or Machine Creation Services. Generally, the use of personal vDisks is evaluated when user
groups requires the ability to personalize their virtual desktop. This could include a need to use a
variety of departmental applications or general personalization that is beyond what is available in
the user profile. However, there is no defined limitation of Citrix Personal vDisk technology to
these specific use cases. It is entirely possible to utilize Personal vDisks in scenarios that may not
maximize flexibility or virtual desktop density, yet are entirely appropriate to the enterprise
environment.
The need and use of personal vDisk technology must align with the personalization characteristics
of the user group captured during the Assess phase as well as the FlexCast model selected.
Note: Citrix personal vDisk technology loads a set of Kernel mode drivers very early in the
Windows boot process. By and large, these Phase 1 drivers allow most applications to work in a
seamless manner with a personal vDisk attached. Exceptions to this are applications that operate
or install components prior to the loading of the Citrix personal vDisk drivers. Specifically, certain
device driver installers or Anti-virus software may not work properly if installed within the personal
vDisk.
Decision: Storage Location
Personal vDisks can be attached to any storage defined on the hypervisor host. If multiple
classes of storage are attached to the hypervisor host, the Personal vDisk can be attached to a
lower cost storage tier or a better performing tier in order to maximize end-user experience.
Decision: Size
While many factors play a part in deciding the size of a Personal vDisk, there are certain basic
requirements that should be followed:
 Minimum Size: 3GB
 Maximum Size: Undefined
Note: The Personal vDisk can be expanded but cannot be shrunk.
Beyond system requirements, the following major factors influence what size to allocate for a
user‟s Personal vDisk:
 Anticipated Growth: Personal vDisk size is allocated at a Desktop Catalog level, and
can be associated with higher CPU and memory resources for power users. Most
organizations should size personal vDisks well above the minimum requirement of 3GB
and should factor in each user group‟s workload requirement, as defined during the
assess phase. As a general rule of thumb, the following recommendations provide a
good starting point:
o Light User: 7GB
o Normal User: 10GB
o Heavy User: 15GB
 Storage Technologies: The use of storage technologies such as thin provisioning and
de-duplication can significantly reduce storage requirements for Personal vDisks
allowing larger sizes to be specified:
o Thin Provisioning: Administrator can use thin provisioning to present more
storage space to the virtual machines than is actually available on the storage
repository.
o De-Duplication: Storage requirements may be reduced through the use of
data de-duplication, whereby duplicate data is replaced with pointers to a single
copy of the original item.

 Folder Redirection or Cloud Data Service: In many smaller environments, utilizing
Personal vDisk as a profile management solution is common. However, the limitations
present in distributed profiles quickly present themselves when this option is selected.
The first step many organizations utilize towards a robust profile solution is to utilize
Microsoft Special Folder Redirection in order to redirect user data to a network share.
Other organizations employ a Cloud based data service such as Citrix ShareFile.
Personal vDisks can co-exist with either of these deployment options. Depending on
the implementation mechanism, the presence of either of these technologies can
significantly reduce the initial sizing of Personal vDisks for users.
Applications
Choosing an appropriate application delivery method helps improve scalability, management, and user
experience. Based on the outcome of the Application Assessment, there are several application
delivery methods that can be considered:
 Installed on Image: The application is part of the base desktop image. The install process
involves DLLs, EXEs and other files being copied to the image drive as well as registry
modifications.
 Installed on Personal vDisk: The install process is similar to Installed on Image application,
except that the application is installed into the user‟s Personal vDisk, which is physically
separate but logically connected to the base desktop image.
 Streamed: The application is profiled and delivered to the desktops across the network on-
demand. Application files and registry settings are placed in a container on the virtual
desktop (or in persistent storage associated with the virtual desktop), and are isolated from
the base operating system and each other, which helps to address compatibility issues.
 Hosted: The application is installed on a XenApp server, and users access the application
remotely using the Citrix HDX protocol.
This section outlines the design decisions pertaining to application delivery. While it is important to
understand the technical reasons for delivering an application in one way or another, it is also
important to consider that the most optimal way to deliver an application might not be the “best” way,
contingent on factors such as infrastructure cost and complexity, experience with a given solution, or
alignment with the overall design. The final decision on application delivery methods must be linked to
the Business Priorities determined during the Assess phase, and based on both technical
requirements/capabilities and business processes.
Citrix: Analysis of Citrix‟s internal deployment of XenDesktop virtual desktops featuring
Personal vDisk reveal that most vDisks are utilized at a rate of 36.74%. The average
size of a PvD is initially 10 GB and has only been expanded for very few power-users.
This size includes user profiles as well as any data, applications, and other aspects
accumulated in the course of eight months of usage. In this environment that is *NOT*
utilizing a profile management solution or roaming profiles, vDisks populate a few GB up
front with user profile data then are slowly accreting changes up to the aforementioned
level over time.

Decision: Application Delivery Method
Architects need to determine the most appropriate delivery methods for applications within the
environment, based on user requirements, application compatibility, and other factors determined
during the Application Assessment. It is unlikely that a single delivery method will meet all
requirements; multiple methods will likely be required to meet the needs of users within the
environment. Furthermore, the methods described above are viable for application delivery, but
they create different effects on the virtual desktop in terms of operations, resource impact and user
experience as shown in the following table. The number of different delivery methods used should
be kept to a minimum to avoid over complicating project execution and subsequent operations.
Installed Installed with PVD Streamed Hosted
Description Applications are
installed directly on
the OS Disk.
Applications are
installed on
Personal vDisk
(Hosted VDI Only).
Executed locally, but
not installed.
Streamed on first
use.
Executed remotely
on XenApp Server.
User Access Every user
authorized to access
the desktop image.
Individual users who
have installed the
application.
Only authorized
users.
Only authorized
users.
Updates Update to base OS
image.
Update to individual
applications.
Update to
application profile
Update to
application on
XenApp servers.
Limitations Operating System
dependencies (XP
vs. Windows 7).
Applications that
start early in the
Windows boot
phase (e.g.
Antivirus)
Applications with
device driver or
service
dependencies.
Compatibility with
multi-user windows,
Windows Server
Operating System.
Resource
Usage
Desktop Desktop Desktop XenApp Server
Table 36: Impact of Application Delivery Methods
In order to minimize delivery methods, architects should consider the organization‟s preference for
managing application delivery or multiple images in determining the best approach for application
delivery. Using multiple delivery methods decreases the number of required desktop images but
adds complexity. Taking these factors into consideration, the following table provides high-level
options for application delivery in a virtual desktop environment:
Base
Applications
Wide Use
Applications
Individual/
Departmental
Resource
Intensive
Technically
Challenging
Description Common
Apps/Utilities
used by all
users.
Applications
used by large
groups of users.
Applications
installed by
individuals or
department
managed.
Have heavy
system
requirements.
Complex,
extensive
dependencies,
or requires
specialized
configuration.
Example Microsoft Office,
Adobe Acrobat,
Antivirus.
Corporate
developed
applications.
Call Centre,
Sales
applications.
CAD/CAM,
Developer
Suite.
Healthcare
Management,
Business
Management
Applications.
Preferred
Delivery
Model
Installed on
Desktop.
Installed or
Streamed to
Desktop.
Installed on
Personal vDisk.
Installed or
Streamed to
Desktop.
Hosted on
XenApp.

Table 37: Potential Application Delivery Strategies
This is a starting strategy that must be modified to meet the specific implementation, based on
decisions made during the Application Assessment. For example, the following items will play a
crucial role in determining the proper mix of delivery techniques for a given implementation:
Compatibility
Desktop virtualization typically requires significant changes to be made to an organization‟s
application delivery strategy. For example, many organizations will take the opportunity to
upgrade their desktop operating system and to simplify management by reducing the number of
applications installed into the base image using techniques such as application streaming and
seamless applications. These are significant changes that require comprehensive compatibility
testing so that an appropriate delivery model can be selected for each application during the
design. Important compatibility requirements that may need to be verified include:
 Desktop Operating System: If the application is to be streamed or installed into a
Hosted VDI Desktop, the application must be compatible with the preferred desktop
operating system.
 Server Operating System: Some applications may be more appropriate for delivery
via a Hosted Shared Desktop or published application. In these situations, the
compatibility of the application must be verified against the chosen server operating
system, for example Microsoft Windows Server 2008 R2.
 Application Architecture: It is important to understand whether the application
includes 16-bit, 32-bit or 64-bit code so that an appropriate operating system can be
selected. 16-bit code cannot be executed on a 64-bit operating system such as
Microsoft Windows Server 2008 R2, Microsoft Windows XP x64, or Microsoft Windows
7 x64.
 Interoperability: Some applications may experience complications if they coexist on
the same operating system. Possible causes include shared registry hives, DLLs or INI
files as well as incompatible dependencies. Application interoperability issues should
be identified so that appropriate remediation steps can be taken or an alternative
delivery model selected.
 Application Streaming: The use of application streaming helps to simplify image
management by reducing the number of applications installed into the base image.
However, not all applications are suitable for streaming because they may install device
drivers, use COM+ or form part of the operating system. For an up-to-date list of
limitations, please refer to the documentation of your preferred application streaming
solution vendor.
There are three main techniques that can be used to perform the application compatibility
testing component of the application assessment:
 Manual: With manual application compatibility testing, each application is installed into
the target environment and manually tested. Manual testing is very time consuming
because it requires a full application test per delivery model. In addition, it is difficult to
test every aspect of the application and almost impossible to identify every application
interoperability conflict. As a result, most of the compatibility issues are identified by
production users rather than the testing process.

 Pre-Verified Applications: Most application vendors will supply detailed information on
which operating systems and delivery models their application supports. Please refer to
the application vendor‟s documentation for more information. In addition, the Citrix
Community Verified site maintains a list of applications which Citrix customers and
partners have verified as being compatible with XenApp, XenDesktop and application
streaming. Microsoft also maintains a spreadsheet of applications which have been
verified as being compatible with Microsoft Windows 7 by the software vendor or by the
Windows 7 Logo Program. Please refer to the Microsoft Windows 7 Application
Compatibility List for IT Professionals for more information.
The problem with pre-verified applications is that they are unlikely to provide
compatibility information for every application identified during the inventory and there
won‟t be any information available on applications which have been fully or partially
developed in-house. At best, there will be limited information available on application
interoperability issues.
 Automated Tools: Citrix AppDNA allows applications to be quickly and accurately
analyzed for compatibility across all relevant operating systems and delivery models
including Microsoft Windows XP, Microsoft Windows 7, Citrix XenApp, Microsoft App-V
and Citrix Application Streaming. Applications are imported into AppDNA where they
are analyzed against thousands of application compatibility rules and checked for
interoperability issues. When compatibility issues are identified, AppDNA provides an
explanation of the issue, possible remediation actions and an estimate of the time
required to resolve the issue.
The advantages and disadvantages of each approach are summarized in the following table:
Approach Time Required Cost Complexity Interoperability Accuracy
Manual High Low High Limited Low
Pre-Verified Medium Low Medium Limited Medium
Automated Tool Low Medium Low Yes High
Table 38: Application Compatibility Testing Approaches
Note: Manual compatibility testing may be chosen in preference to an automated solution such
as AppDNA due to licensing costs. However, manual compatibility testing has been shown to
be error prone, time consuming and more expensive than using an automated solution.
Regardless of the approach used, the compatibility testing results should be captured in the
“Compatibility” section of the Application Assessment Workbook so that they can be analyzed
during the design phase:
 Pre-Requisites: Many applications will depend on certain pre-requisites to function
correctly, for example, the Java Runtime Environment, .Net Framework or a database
driver. All essential pre-requisites should be captured during the application
assessment so that an appropriate image design can be created.
 Dependent Apps: Applications may need to interact with each other to provide the
users with a seamless experience. For example, applications that present information
in a PDF format require a suitable PDF viewer to be available. Capturing application
dependencies will help to ensure that an appropriate image design can be created.
 16-bit Code: The application assessment should determine whether the applications
include any 16-bit code because they cannot be supported on a 64-bit operating
system. There are three classifications available for “16-bit Code” in the Application
Assessment Workbook: Yes, No or Unknown.

 Windows XP: Specify whether the application passed compatibility testing for Microsoft
Windows XP. There are six classifications available for “Windows XP” in the
Application Assessment Workbook: Yes – No Remediation, Yes – Low Remediation,
Yes – Medium Remediation, Yes – High Remediation, No or Unknown.
 Windows 7: Specify whether the application passed compatibility testing for Microsoft
Windows 7. There are six classifications available for “Windows 7” in the Application
Assessment Workbook: Yes – No Remediation, Yes – Low Remediation, Yes – Medium
Remediation, Yes – High Remediation, No or Unknown.
 XenApp 6.5: Specify whether the application passed compatibility testing for Citrix
XenApp 6.5. There are six classifications available for “XenApp 6.5” in the Application
Assessment Workbook: Yes – No Remediation, Yes – Low Remediation, Yes – Medium
Remediation, Yes – High Remediation, No or Unknown.
 Application Streaming: Specify whether the application passed compatibility testing
for application streaming. There are five classifications available for “Application
Streaming” in the Application Assessment Workbook: Yes, No, Citrix Streaming Only,
App-V Only or Unknown.
User Segmentation
It is unlikely that all users require all applications. Certain applications may be required by
departments or a small numbers of users within/across departments. Architects need to
determine the optimal way to deliver applications to users, considering the delivery methods
available, and the management requirements for various applications. If a large number of
applications will be delivered to a subset of users, a separate image may be required. If a small
number of users require an application, it can be delivered using an alternative delivery method
such as installed on a Personal vDisk or streamed.
 IT Experience: If IT has experience or infrastructure already in place for a given
application delivery model, then it may make sense to have preference for that model.
For example, if an organization has experience using Microsoft App-V for streaming
applications, then streamed application delivery may be preferential vs. hosted
providing applications can be delivered using both technologies.
 Management Requirements: Application Delivery strategy may depend on who owns
and is responsible for a given application. If an application is owned by the corporation,
then centralized IT can maintain the application, and it can be published or streamed. If
however the application is owned and managed by a department, and management
responsibility cannot be given to centralized IT, then the application may have to be
installed in a Personal vDisk and managed by the department as a unique entity.
 Update Frequency: Time and effort required to update an application will impact the
selection of a delivery model. If an application must be updated frequently, then
architects will want to choose a delivery model which minimizes the number of
instances of the application that need to be updated, as well as the complexity of the
update process. Hosted applications can be updated in a few locations, and the
process will be simple. Applications on a Personal vDisk will need to be managed in
(potentially) many locations. Streamed applications may have a single source, but a
relatively complex update process, including re-packaging and testing. These factors
will impact the delivery decision.

 Licensing: If an application is available to an organization at no cost, or in a site
licensing model, then it can be distributed to users even if they may not use the
application. While this may add some slight complexity to the image, it can simplify the
delivery process, reduce the number of images required and the overall complexity of
the solution. If application licensing is a concern, Architects will need to use a delivery
model that conforms to the vendor-licensing model and allows for policy based control
over access to the application.
 Resource Use: Applications that make heavy use of system resources are better
suited for execution on the virtual desktop as it allows for more granular control over
assigned resources and isolation between users is maximized. Consider a delivery
method such as installed, streamed or installed on Personal vDisk for these
applications.
 Technically Challenging: If an application is difficult to setup, requires a specialized
configuration, or has significant dependencies on other applications, it should be
considered technically challenging. Delivering technically challenging applications by
hosting on XenApp can reduce the complexity of the base image, but consideration
needs to be made for dependencies upon other applications which may already exist as
part of the base image set (e.g. Office applications, Acrobat, E-mail, etc.), and the
number of users accessing the application.
Images
When determining the strategy for the desktop image, architects need to consider the needs of the
users and their applications as well as the delivery methods required to provide images to the
desktops. Considerations will include the operating system version and whether the users require a
32 or 64 bit environment, the image delivery strategy, appropriate policies and how the users and
images are aligned into different sets of desktop groups and catalogs.
Decision: Operating System
When considering the operating system, it is critical to understand the application and user
requirements that define the user‟s operating environment. An operating system upgrade to a new
OS such as Windows 7 may be a driver for a virtualization project, but architects need to consider
the outcome of the user segmentation and application assessment during the assess phase when
defining the image. Similarly, the decision to use a 32 or 64 bit version of an operating system is
also related to these assessment phases as the availability or compatibility of applications with a 64
bit operating system will play a role in the decision.
If a 64 bit operating system is used, architects will need to consider the additional overhead of the
operating system versus its 32 bit variant. As an example, the table below lists the differences
between the uses of a 32 bit vs. a 64 bit version of Windows 7 on a virtualized infrastructure.
Communications: A large telephone company delivers the majority of its applications
using the installed on desktop model, with images focused on user segmentation.
Applications that are used by only a few users are installed on personal vDisk to allow
customization. Applications that are used by all users, but need isolation from the pooled
desktop image are delivered through application streaming.
Energy: An energy company installs applications on the base image for the majority of
users and streams departmental applications as required.
Financial: A banking customer maintains and deploys multiple desktop images containing
user group focused applications as required by various departments.

Consideration Windows 7 X86 (32 Bit) Windows 7 x64 (64 Bit)
Minimum Memory
Requirement (per
VM)
1GB 2GB
Minimum Disk
Requirement
16 GB 20 GB
PVS Stream boot
data
166 MB 210 MB
PVS Stream
bandwidth
83,000 Mb 105,000 Mb
Table 39: Differences between 32 bit vs. 64 bit
Decision: Delivery Method
When considering the various FlexCast delivery methods available to deliver virtual desktops, there
is no single best answer for all organizations. The decision on delivery methods should be
predicated on the results of the user segmentation exercise during the assess phase, and should
also consider the following.
Scope
Scope deals with the number of different methods that may be used within an organization.
User segmentation may indicate that multiple methods should be used in order to meet the
needs of various departments, or may determine that a single method will suffice. The scope of
desktop types required will determine the complexity and number of delivery methods used.
Manageability
Each delivery method employed in an environment brings unique and specific manageability
challenges, both from a perspective of maintaining the image and growing the environment.
From an image maintenance perspective, architects need to be concerned with the processes
required to update the master image or individual desktop images, and how to keep the images
in sync across the various physical devices or hypervisor pools. Requirements to avoid or
recover from image corruption should also be considered. With respect to expansion, architects
should consider the scalability requirements for each component of the infrastructure required to
deliver the desktops, and should consider when to scale up or scale out components such as
storage, hypervisor pools and infrastructure components (e.g. Provisioning Servers) when
designing the delivery method.
Infrastructure Requirements
Finally, the infrastructure requirements of the delivery method need to be considered. These
can be hypervisor resources such as CPU, memory and storage, or can be infrastructure
components such as Provisioning Servers. Architects need to consider the overall
infrastructure requirements as they pertain to the various FlexCast models and how they will
scale as the size of the environment grows. These elements will be covered in detail in the
Control and Hardware sections of this document.
Decision: Computer Policies
Citrix policies provide the basis to configure and fine tune your XenDesktop and XenApp
environments, allowing organizations to control connection, security and bandwidth settings based
on various combinations of users, devices or connection types. Correctly defining an initial
baseline policy and assigning additional policies based on security requirements and specific
access scenarios can be important in delivering a high definition user experience.

When making policy decisions it is important to consider both Microsoft Windows and Citrix policies
as components within each of these policy configurations have an impact on user experience and
environment optimization. For more details on specific Windows related policies, refer to the Group
Policy Settings Reference for Windows and Windows Server, specifically settings related to
Windows Server 2008 R2 and Windows 7.
Developing a Computer Policy solution for the overall architecture includes the same set of design
decisions defined within the User Policy section. These include:
 Preferred Policy Engine
 Policy Integration
 Policy Filtering
 Policy Precedence
 Baseline Policy
Note: A Baseline Policy Reference is provided within the appendix that includes a baseline
recommendation for Citrix User Policy, Citrix Computer Policy, Microsoft Windows Policy and
Folder Redirection Policy.
Note: For a list of all Citrix-related policies, please refer to the Policy Quick Reference within the
Appendix.
Decision: Desktop Catalogs
Desktop catalogs specify virtual machines or physical devices that host user desktops, the Active
Directory computer accounts associated with the desktops and, in some cases, the master VM that
is used to create the user desktops.
As a catalog is simply a definition of the virtual desktop resources, a single catalog can span
across multiple hosts or hypervisor pools and associated resources such as storage. If a catalog
spans across multiple hosts, it is important to ensure that the hosts have the appropriate templates
and required VM‟s available for cloning and imaging, depending upon the FlexCast model of the
desktop being delivered. Architects will also need to consider the methods used to ensure that the
base desktop image is updated across all the hosts within the catalog, as required.
Generally, in order to simplify management of the environment, each catalog created should be of
the same machine type (e.g. Pooled Random). Although this is not a product based restriction,
constraining catalogs by machine type will allow for simplified management and an easier to
understand catalog structure.
Decision: Desktop Groups
Desktop groups provide the linkage between users and catalogs. Assigning desktop groups to
users can be performed 1:1 or 1 to many, and desktop groups can span multiple catalogs. This
process allows architects to better align their desktop allocations with unique user requirements.
For example, if a set of developers require both a corporate pooled desktop for day to day
operations, and a set of dedicated desktops for development and testing, these desktops can be
assigned within a single desktop group. To define the desktop allocation, architects need to
understand their user requirements through analysis during the assessment phase, after which the
results can be used to develop the desktop allocation solution. Decisions for desktop group
assignment include the following considerations:
 Catalogs can span multiple hypervisor pools
 Users can be assigned to multiple catalogs
 A catalog can be divided between any number of user groups where certain groups are
allocated a percentage of the maximum catalog size

 Groups can only be created if there are unused machines in the catalog you select
 A machine cannot be used in more than one desktop group
 Groups can be created from multiple catalogs with the same machine type
 Mixed desktop groups cannot be created from catalogs with multiple machine types
Desktop groups also allow an organization to setup delegated administration allowing specific
administrators to manage a group of virtual desktops. For example, a desktop group can be setup
for executive use, with a defined set of delegated administrators to provide VIP support of the
executive desktops.

Control Layer
The control layer is the fourth layer of the design methodology. Every major design decision made for all
user groups in the upper three layers are used as a whole to help design the control components of the
overall solution.
The design decisions for each user group are met by incorporating the correct control layer components,
which includes access controllers, desktop controllers and infrastructure controllers. Determining
capacity, configuration, topology and redundancy for the respective components creates a control
environment capable of supporting the user requirements.
Access Controllers
The access controller sub-layer is responsible for providing the infrastructure components to support
the access layer requirements for each user group. Access controllers often include some combination
of Web Interface, StoreFront and NetScaler Access Gateway.
Remote Access Architecture
If there are user groups that require „remote‟ or „offline‟ mobility, appropriate remote access
architecture must be designed.
Decision: Topology
Selection of the network topology is central to planning the remote access architecture to
ensure that it can support the necessary functionality, performance, and security. The design of
the remote access architecture should be completed in collaboration with the security team to
ensure adherence to corporate security requirements. There are three primary topologies to
consider, each of which provides increasing levels of security:
 1-Arm (Normal Security): With a 1-arm topology, the Access Gateway utilizes one
physical or logical bonded interface, with associated VLAN and IP subnet, to transport
both frontend traffic for users, and backend traffic for the virtual desktop infrastructure
servers and services.
Figure 16: 1-Arm Topology
 2-Arm (High Security): With a 2-arm topology, the Access Gateway utilizes two or
more physical or logical bonded interfaces, with associated VLANS and IP subnets.
Transport of the frontend traffic for users is directed to one of these interfaces. The
front-end traffic is isolated from backend traffic, between the virtual desktop

infrastructure servers and services, which is directed to a second interface. This allows
the use of separate demilitarized zones (DMZs) to isolate frontend and backend traffic
flows along with granular firewall control and monitoring.
Figure 17: 2-Arm Topology
 Double-Hop DMZ (Very High Security): The Double-Hop DMZ shares characteristics
of the 2-arm topology, yet utilizes two separate Access Gateways. Some organizations
use three physical or logical firewalls to protect their internal networks. The three
firewalls divide the DMZ into two stages to provide an extra layer of security for the
internal network.
The Access Gateway in the first DMZ handles client connections and performs the
security functions of an SSL VPN. This Access Gateway encrypts the client
connections, determines how users are authenticated, and controls access to the
servers in the internal network.
The Access Gateway in the second DMZ serves as a proxy device. This Access
Gateway enables the ICA traffic to traverse the second DMZ to complete client
connections to the server farm. Communications between the Access Gateway in the
first DMZ and the Secure Ticket Authority (STA) in the internal network are also proxied
through the Access Gateway in the second DMZ.

Figure 18: Double-Hop DMZ Topology
Decision: Platform
As discussed in the Access Layer, any design decisions pertaining to Remote Access will be
made with the NetScaler Access Gateway in mind. In order to identify an appropriate NetScaler
platform to meet project requirements, the key resource constraints must be identified. Since all
Remote Access traffic will be secured using the secure sockets layer (SSL), transported by
Hypertext Transfer Protocol (HTTP) in the form of HTTPs, there are two resource metrics that
should be targeted:
 SSL Throughput: The SSL Throughput is the Gigabits of SSL traffic that may be
processed per second (Gbps).
 SSL Transactions per Second (TPS): The TPS metric identifies how many times per
second an Application Delivery Controller (ADC) may execute an SSL transaction. The
capacity varies primarily by the key length required. TPS capacity is primarily a
consideration during the negotiation phase when SSL is first setup, and it is less of a
factor in the bulk encryption / decryption phase, which is the majority of the session life.
While TPS is an important metric to monitor, field experience has shown that SSL
Throughput is the most significant factor in identifying the appropriate NetScaler Access
Gateway.
For more information, please refer to the Citrix whitepaper - Best Practices for implementing
2048-bit SSL.
The SSL bandwidth overhead average is often considered negligible relative to the volume of
virtual desktop traffic and is not typically accounted for as part of required SSL Throughput.
However making provisions for SSL bandwidth will help ensure the total throughput estimated is
sufficient. The fixed bandwidth added to packet headers can vary according to the encryption
algorithms used, and the overall percentage of bandwidth may vary widely according to packet
size. Ideally, the overhead should be measured during a proof of concept or pilot. However, in
the absence of such data incrementing the workload bandwidth by 2% is a reasonable rule of
thumb. Therefore, to determine the SSL throughput required by a NetScaler platform, multiply
the maximum concurrent bandwidth for a data center by 1.02:

For example, assuming 128Mbps maximum concurrent bandwidth, the appropriate NetScaler
model can be determined as follows:
The SSL throughput value should be compared to the throughput capabilities of various
NetScaler platforms to determine the most appropriate one for the environment. There are
three main platform groups available, each of which provides broad scalability options.
 VPX: A NetScaler VPX device provides the same full functionality as hardware
NetScaler. However, NetScaler VPXs can leverage „off the shelf‟ servers for hosting
and are suitable for small to medium sized environments.
 MDX: A NetScaler MDX is the hardware version of the NetScaler devices. The MDX
device is more powerful than the virtual NetScaler and can support network
optimizations for larger scale enterprise deployments.
 SDX: A NetScaler SDX is a blend between the virtual and physical NetScaler devices.
An SDX machine is a physical device capable of hosting multiple virtual NetScaler
devices. This consolidation of devices aids with reducing required shelf space and
device consolidation. NetScaler SDXs are suitable for handling network
communications for large enterprise deployments.
SSL Throughput capabilities of the NetScaler platforms may be found in the Citrix NetScaler
Data Sheet. Therefore, based on the example calculation above, a NetScaler MPX 5500
appliance would be sufficient to handle the required load. However, actually scalability will
depend on security requirements. NetScaler SSL Throughput decreases with the use of
increasingly complex encryption algorithms, and longer key lengths. Also, this calculation
represents a single primary NetScaler. At a minimum, N+1 redundancy is recommended which
would call for an additional NetScaler of the identical platform and model.
Note: The Citrix NetScaler Data Sheet typically represents throughput capabilities under
optimal conditions for performance. However, performance is directly affected by security
requirements. For example, if the RC4 encryption algorithm and a 1k key length are used, a
VPX platform may be able to handle more than 500 HDX proxy connections. However, if a
3DES encryption algorithm and 2k key length are used (which is becoming more common), the
throughput may be halved.
StoreFront
Citrix StoreFront, which is the successor to Citrix Web Interface, authenticates users to
XenDesktop sites, XenApp farms, App Controller (SaaS Apps), and VDI-in-a-Box enumerating and
aggregating available desktops and applications into stores that users access through Citrix
Receiver for Android, iOS, Linux, Windows, Win8/RT or Receiver for Web sites. StoreFront is an
integral component of XenDesktop 7 but can be used with XenApp and XenDesktop 5.5 and up
deployments. It has been built on a modern, more flexible and powerful framework than Web
Interface that enables StoreFront to provide next generation features, as outlined in eDocs – About
StoreFront.

The following diagram depicts a typical StoreFront infrastructure for environments without XenMobile:
Please refer to CTX138635 - Citrix Reference Architecture for XenMobile 8.5 for further information about XenMobile deployments.

Decision: Web Interface or StoreFront
Web Interface and StoreFront are two different solutions, whose feature sets overlap in many
areas, but also offer a variety of distinct features. Therefore it is very important for
organizations to review the capabilities of each product against their requirements. In general, it
is strongly recommended to build new solutions based on StoreFront, since new features will
not be added to Web Interface and end of life has been announced for Web Interface.
Furthermore it is important to understand that Web Interface does not support XenDesktop 7 or
later. Details on Web Interface lifecycle milestones are available from the Citrix website –
Lifecycle Milestones.
While StoreFront goes beyond Web Interface in many areas, StoreFront 2.0 does not support
all features of Web Interface. The following tables outlines the Web Interface features that are
not currently available in StoreFront:
Area Feature
Deployment Options Web Interface on NetScaler (StoreFront is deployable as an application
behind NetScaler but runs on separate servers)
Authentication Delegated Kerberos Authentication
Active Directory Federation Services (ADFS) 1.0 integration
Account self-service (SSPR) (reset/unlock with security questions)
Smart card authentication via browser (Native Receivers required)
Domain pass through authentication via browser (Native Receiver for
Windows required)
Support for Novell NDS
Anonymous authentication
Other Messaging (user notifications)
Settings per location (IP Subnet)
Client proxy settings configuration
Offline Apps (Users cannot access offline applications or App-V
sequences through Receiver for Web sites. Native Receiver is
required)
Compact/Low graphics Mode and embedding
Table 40: Web Interface features currently not supported by StoreFront 2.0
Decision: High Availability
If the server hosting StoreFront or the respective web service is unavailable, users will not be
able to launch new virtual desktops, published applications or manage their subscriptions.
Therefore at least two StoreFront servers should be deployed to prevent this component from
becoming a single point of failure. Options include:
 DNS Round Robin: Provides rudimentary load balancing across multiple servers
without performing any checks on availability. If one server fails, only a portion of user
requests will be affected.
 Windows NLB: A Windows service capable of performing rudimentary checks to verify
the server is available but cannot determine the status of individual services. This can
cause users to be forwarded to StoreFront servers which are not able to process new
requests

 Citrix NetScaler: An intelligent appliance, which is capable of verifying the availability
of the StoreFront service and actively load balance user requests appropriately.
Decision: Security – Inbound Traffic
Communications from the web browser or Receiver and StoreFront server include user
credentials, resource sets, and session initialization files. This traffic is typically routed over
networks outside the datacenter boundaries or on completely untrusted connections (such as
the Internet). Therefore Citrix strongly recommends that this traffic is encrypted using SSL.
Note: By default, Citrix Receiver requires SSL to connect to StoreFront. This means email-
based account discovery or the manual configuration of a StoreFront store in Receiver will not
work unless a valid and trusted SSL certificate has been implemented on the StoreFront server
and/or the respective external load balancer Please refer to eDocs – Configure and install
Receiver for further information.
Decision: Security – Backend Traffic
User credentials are sent between StoreFront and the XenApp Controllers, XenDesktop
Controllers and the App Controller virtual appliance. For example, in a typical session with a
XenDesktop Controller, the StoreFront server passes user credentials to the Citrix XML Service
for user authentication and the Citrix XML Service returns resource set information. A TCP/IP
connection and the Citrix XML protocol is used to pass the information between the StoreFront
server and the XenDesktop site. The XML protocol uses clear text to exchange all data, with the
exception of passwords, which are transmitted using obfuscation.
Recommendation: For Citrix environments with high security requirements, encrypt StoreFront
to XenApp, XenDesktop and App Controller communications. For further guidance on how to
encrypt this traffic, please refer to
 eDocs – Use the SSL Relay (XenApp only).
 CTX130213 - How to Configure SSL on XenDesktop 5 Controller to Secure XML Traffic
(XenDesktop only)
 eDocs – Use SSL on XenDesktop 7 Controllers (XenDesktop only)
Please refer to eDocs – Secure your StoreFront environment for further information
Delivery Controllers
To provide users with desktops and applications, StoreFront must be configured with the IP
address or DNS name of at least one Controller in each XenDesktop site and/or XenApp farm.
For fault tolerance, multiple Controllers should be entered for each site and/or farm specified.
StoreFront will automatically failover to the second server in the list in case the first server
becomes unavailable (active/passive). For large infrastructures or environments with a high
logon load an active distribution of the user load (active/active) is recommended. This can be
achieved by means of an industry proven load balancer with built-in XML monitors and session
persistency, such as Citrix NetScaler.
Decision: Beacons
Citrix Receiver uses beacon points (web sites) to identify whether a user is connected to an
internal or external network. Internal users are connected directly to resources while external
users are connected via Citrix NetScaler Gateway. Citrix Receiver continuously monitors the
status of network connections (e.g. link up / link down or change of the default gateway). When
a status change is detected, Citrix Receiver will first check that the internal beacon points can
be accessed before moving on to check the accessibility of external beacon points. StoreFront
provides Citrix Receiver with the http(s) addresses of the beacon points during the initial
connection process and provides updates as necessary.

Specifying at least two highly available external beacons that can be resolved from public
networks so that Citrix Receiver can determine whether users are located behind an Internet
paywall, such as in a hotel or Internet café.
Decision: Auto Provisioned Apps (Keywords)
StoreFront displays applications differently to Web Interface. Instead of having all accessible
applications appear on the home screen, first time users are invited to choose (subscribe) to the
applications they want to regularly use after they logon. Before a user can launch an
application, they must first choose which applications should be placed on their home screen.
This approach, deemed “Self-Service” apps, allows users to restrict the applications that they
see on their home screen to the ones that they use on a regular basis. The applications chosen
by every user for each store are recorded by the subscription store service so that they can be
displayed on the Receiver home screen from any device that the user connects from (Follow me
Apps).
To avoid users from having a blank screen when they first logon, it is recommended that
administrators automatically subscribe users to a few core applications. To do this,
add KEYWORDS:Auto to the application or desktop description in XenApp or XenDesktop.
Another option that can be used to organize applications is KEYWORDS:Featured. Unlike the
Auto keyword which places certain apps on the home screen, the Featured keyword only places
apps in the Featured category (as shown below).
The app will also appear in another category if a Client Application folder has been specified.
In addition the string KEYWORDS:prefer="application" can be used to specify that the locally
installed version of an application should be used in preference to the equivalent delivered
instance if both are available.
For further information please refer to eDocs – Optimize the user experience.
Decision: Scalability (preliminary)
The number of Citrix Receiver users supported by a single StoreFront server depends on the
hardware specifications and on the level of user activity. At the current point in time, scalability
testing for StoreFront 2.0 has not been completed. Early testing results indicate that, a single
StoreFront 2.0 server with twin 2 GHz quad-core CPUs and 8 GB RAM supports up to 25,000

user connections per hour in a light usage scenario (users log on, enumerate their resources,
and access existing subscribed resources) or up to 6000 user connections per hour in an
intensive usage scenario (users log on, enumerate their resources, and then subscribe and
unsubscribe to a resource.)
For the optimum user experience, Citrix recommends that not more than 10 XenDesktop,
XenApp, App Controller, and VDI-in-a-Box deployments are aggregated in a single store.
Desktop Controllers
The desktop controller sub-layer is responsible for providing the infrastructure components to support
the desktop layer requirements for each user group. Desktop controllers often include some
combination of XenApp data collectors, XenDesktop controllers, Provisioning Services servers and
XenClient Synchronizers.
XenDesktop
A Citrix XenDesktop site groups desktops and controllers together to form a single architectural
and management entity. All persistent and dynamic data for the site, including desktop
assignments and session state, is stored in a central site database.
Site-wide functions are equally spread across all XenDesktop Controllers within a site. While it is
possible to assign certain functions to specific controllers, it is not recommended as XenDesktop is
self-optimizing and manual configurations can interfere with the automatic load balancing and
failover mechanisms of XenDesktop.
Decision: Number of Sites
There are several factors that must be considered when determining the number of XenDesktop
sites required:
 Network: XenDesktop Controllers constantly access the site database and regularly
communicate with desktops. XenDesktop may exhibit unexpected behavior if
significant levels of latency exist between the XenDesktop Controllers, SQL Database
and virtual desktops or the site communications are interrupted. Therefore, all site
components - controllers, desktops, virtualization hosts and database servers, should
be connected by a redundant and high-speed network.
The following two scenarios demonstrate how the network can affect XenDesktop site
design:
1. Three data centers connected by a redundant but high-latency network -
three sites:
2. Two data centers connected by a fast, redundant network and a third data
center connected by a high latency, redundant network – two sites:

Note: Implementing a single XenDesktop site spanning multiple data centers
connected by fast, redundant links helps to reduce management overhead
and supporting infrastructure. However, creating separate sites per data
center is a safer alternative.
 Risk Tolerance: Multiple XenDesktop sites can be created to minimize the impact from
a site-wide outage. For example, corruption of the XenDesktop site database could
affect site-wide availability. For many organizations, the decreased risk from
implementing multiple sites outweighs the additional management overhead and
supporting infrastructure required.
 Security: Although delegated administration is available, high-security organizations
may require complete separation between environments to demonstrate compliance
with specific service level agreements.
In general, the number of XenDesktop sites should be kept to a minimum to reduce architectural
complexity and administrative effort.
Decision: XenDesktop Controller Sizing
The XenDesktop Controllers authenticate users, enumerate resources, direct user launch
requests and control desktop startups, shutdowns and registrations.
XenDesktop Controller scalability is based on CPU utilization. The more processor cores
available, the more virtual desktops a controller can support. Each desktop startup, registration,
enumeration and launch request impacts the controller‟s processor. As the storm increases in
intensity, the CPU utilization of the controller will increase. If the CPU reaches a critical
threshold, roughly 80%, the site will need to either scale up or scale out.
Finance: A large financial institution hosts 10,000 desktops from a single data
center. To reduce risk, it was decided that no site should exceed 5,000 desktops.
Therefore, despite the desktops being connected by a fast and redundant
network, two sites were created.
Retail: A retail organization required complete separation for employees
responsible for managing financial data. To meet this requirement, two separate
sites were created within the same data center – one for the financial employees
and a second for all other employees.

Adding additional CPU cores to a XenDesktop controller will lower the overall CPU utilization,
thus allowing for greater numbers of desktops supported by a single controller. This is really
only feasible when dealing with virtualized controllers as adding virtual CPUs is fairly easy and
straightforward. The other alternative is to add another controller into the site configuration.
The controller would have the same configuration as other controllers, and the load would be
evenly distributed across all controllers, thus helping to reduce the overall load on each single
controller.
Testing has shown that a single XenDesktop Controller using the following configuration can
support more than 5,000 desktops.
Component Specification
Processor 4 vCPU
Memory 4GB RAM
Network Bonded virtual NIC
Host Storage 40GB shared storage
Operating System Windows Server 2008 R2
Windows Roles / Features Web Services (IIS)
.Net Framework 3.5.1
Windows Process Activation Service
XenDesktop 5,5, 5.6
Table 41: XenDesktop Controller Specification for 5K Desktops
Controllers are constantly updating the SQL database with site status allowing a Controller to go
offline without impacting user activities, assuming that scalability demands are met. At least
one additional XenDesktop Controller should be assigned to each site (N+1 redundancy) to
ensure availability should a single controller fail.
The following formula can be used to calculate the number of XenDesktop Controllers required:
Decision: Host Connection Configuration
Host connections allow controllers within a site to access the hypervisor(s) hosting the virtual
desktops. Separate host connections will be required for each XenServer pool, Hyper-V
SCVMM server or VMware vCenter server. For redundancy, XenServer host connections
should have at least one HA server configured and SCVMM servers and vCenter servers
should be highly available.
Host connections define the storage repositories and guest networks that will be used by the
virtual machines. Both Provisioning Services and Machine Creation Services use these
parameters during the desktop creation process to determine which guest network and storage
Government: A government organization was sizing the XenDesktop Controllers
required to support a 3,700 desktop site. Based on the testing performed by Citrix,
they deployed two dedicated virtual XenDesktop Controllers, each of which was
assigned 4 vCPUs and 4GB of RAM. This configuration ensures that they have
sufficient capacity to support all 3,700 users in the event that one XenDesktop
Controller fails.

location should be used. Each host connection can define multiple storage repositories and a
single guest network. Therefore, separate host connections are necessary for each guest
network required.
Power operations, such as startup, shutdown and suspend are very resource intensive on the
hypervisor and the storage subsystem in particular. Each host connection includes a number of
advanced options that help to prevent the Controllers from overloading the associated
hypervisor with power operations. The values for the maximum number of active actions and
the maximum number of new actions per minute should be adjusted according to the
capabilities of the hypervisor pool. Initial values should equal the number of hosts multiplied by
two. Higher values should only be enabled after comprehensive scalability testing has been
completed as they could negatively impact performance and user perception.
Note: Other host connection advanced options should only be configured under the guidance of
Citrix support.
Decision: XML Service Encryption
In a typical session, the Web Interface / StoreFront server passes credentials to the Citrix XML
Service on a XenDesktop Controller. The Citrix XML protocol uses clear text to exchange all
data, with the exception of passwords, which are transmitted using obfuscation.
If the traffic between the Web Interface / Storefront servers and the XenDesktop Controllers can
be intercepted it will be vulnerable to the following attacks:
 Attackers can intercept the XML traffic and steal resource set information and tickets.
 Attackers with the ability to crack the obfuscation can obtain user credentials.
 Attackers can impersonate the XenDesktop Controller and intercept authentication
requests.
For most organizations, the Citrix XML traffic will be isolated on a dedicated physical or virtual
data center network making interception unlikely. However, for safety consider encrypting this
traffic using one of the following methods:
 Use the SSL Relay as a security intermediary between the Web Interface / StoreFront
server and the XenDesktop Controllers. The SSL Relay performs host authentication
and data encryption.
 Use the HTTPS protocol to send Web Interface data over a secure HTTP connection
using SSL if IIS is installed on the XenDesktop Controller.
XenApp
Decision: Farm Topology
A Citrix XenApp farm forms a single architectural and management entity with all persistent data
for the farm, including applications assignments and server properties, being stored within the
central database, while all dynamic data, including session, user and connection information,
being stored in memory on the zone‟s primary controller (data collector). Farm-wide functions
are consolidated into a centralized, primary data collector, which acts as the master for the
entire zone.
A XenApp farm can be further subdivided into multiple zones where each zone is managed by a
single, primary controller and all members within the zone communicate with only their
respective controller. In order to maintain consistency across the entire farm, each zone‟s
primary controller synchronizes the dynamic data with each other in a star topology. As the
number of zones increase, the amount of traffic between zones increases exponentially.

There are several factors that must be considered when determining the XenApp topology
(farms and zones):
 Network: Designing a single farm with multiple zones to support multiple sites typically
assumes there is sufficient network bandwidth available to support IMA traffic, database
traffic and inter-zone communications. However, it is important to take into
consideration the quality of the network connection. If the network connection exhibits
very high latency and/or high error rate, a single farm design may not be appropriate.
The network quality issue is a typical problem with satellite links and links to
international sites.
 Geography: If the existing infrastructure is regionally (global or nationwide) deployed,
and the management of the infrastructure is decentralized (i.e., due to political and
organizational boundaries), multiple farms may eliminate the complication of
coordinating and scheduling changes to the farm, prevent the risk of a global outage or
downtime, minimize the scope of troubleshooting farm-wide issues, and improve the
overall performance of the farm.
 Risk Tolerance: An outage in a single large (or global) farm may impact all users while
an outage in a regional farm will only affect users in a region. With multiple smaller (or
regional) farms, the potential for a large-scale outage is substantially minimized. During
major farm upgrades or maintenance, multiple farms can be an effective mechanism for
pursuing a phased implementation; the first farm can be used as a production pilot prior
to wide-scale implementation.
 Security: Although delegated administration is available, high-security organizations
may require complete separation between environments to demonstrate compliance
with specific service level agreements.
In general, the number of XenApp farms and zones should be kept to a minimum to reduce
architectural complexity and administrative effort. In many situations, organizations deploy a
single farm/single zone for each geographic location.
Note: For additional information on XenApp zones, please refer to Designing Zones for XenApp
Deployment.
Decision: XenApp Controller Sizing
The primary XenApp controller authenticates users, enumerates resources and directs user
launch requests.
XenApp controller scalability is based on the time it takes for the controller to respond to a
request. Adding additional processor cores or memory has only a slight impact to overall
scalability of the zone. Each desktop/application enumeration and launch request impacts the
controller‟s performance. As the number of requests increases in intensity, the time it takes for
the controller to respond to the requests also increases.
In order to increase performance, the primary XenApp controller should be dedicated for
environments with more than 10 XenApp servers, and a dedicated backup XenApp controller
should be allocated for environments with more than 20 XenApp servers. Each server should
have the following specifications.
Processor 4 vCPU
Memory 4GB RAM

Windows Roles / Features .NET Framework 3.5.1
Remote Desktop Services Role
Windows Application Server Role
XenApp 6.5 (Controller role)
Decision: XML Broker Sizing
The XML Brokering process of the primary XenApp controller can be allocated to another set of
dedicated servers to improve performance of the primary XenApp controller. For environments
beyond 2,000 users, or for farms where periods of heavy logon traffic is anticipated, Citrix
recommends dedicated virtual servers for XML Broker components with the following
specifications:
Processor 4 vCPU
Memory 4GB RAM
Windows Roles / Features .NET Framework 3.5.1
Remote Desktop Services Role
Windows Application Server Role
XenApp 6.5 (XML Broker role)
The following formula can be used to calculate the number of XML Brokers required:
Decision: XML Service Encryption
In a typical session, the Web Interface / StoreFront server passes credentials to the Citrix XML
Service on a XenApp Controller. The Citrix XML protocol uses clear text to exchange all data,
with the exception of passwords, which are transmitted using obfuscation.
If the traffic between the Web Interface / Storefront servers and the XenApp Controllers can be
intercepted it will be vulnerable to the following attacks:
 Attackers can intercept the XML traffic and steal resource set information and tickets.
 Attackers with the ability to crack the obfuscation can obtain user credentials.
 Attackers can impersonate the XenApp Controller and intercept authentication
requests.
For most organizations, the Citrix XML traffic will be isolated on a dedicated physical or virtual
data center network making interception unlikely. However, for safety consider encrypting this
traffic using one of the following methods:

 Use the SSL Relay as a security intermediary between the Web Interface / StoreFront
server and the XenApp Controllers. The SSL Relay performs host authentication and
data encryption.
 Use the HTTPS protocol to send Web Interface data over a secure HTTP connection
using SSL if IIS is installed on the XenApp Controller.
XenClient Synchronizer
Citrix XenClient is a client hypervisor that runs on bare metal. XenClient allows users to run
multiple local virtual desktops simultaneously, side-by-side and in complete isolation. XenClient
provides control over hard-to-manage laptops and provides users with access to their virtual
desktop anywhere, anytime, even while disconnected from the network.
XenClient consists of two main components – „Synchronizers‟ and „Engines‟. At the control layer,
the Synchronizer provides the delivery of images and policies operations. The Engine refers to the
XenClient hypervisor that is installed on a physical endpoint computer so that it can host virtual
machine (VM) images.
Decision: Architecture
The XenClient Synchronizer can be hosted on a physical or virtual server. However, the
Synchronizer is only compatible with Hyper-V when authoring VM images. Although Hyper-V
must be used during the VM creation process, XenServer and VMware can also be used to host
the Synchronizer server. The decision on whether the Synchronizer should be virtualized or not
will depend on the existing capabilities of the environment:
 Physical Server: If a Hyper-V deployment is not present, it is recommended that the
Synchronizer role be hosted on a dedicated physical server. In this scenario, the
Hyper-V role is added to the same physical server that hosts the Synchronizer role.
Utilizing a dedicated physical server will ensure that other services do not impact the
performance of the Synchronizer role. In addition, time spent creating and updating
images will improve since the data is transferred locally rather than over the network.
Along with the Hyper-V role, the SQL database can also be installed on the same
physical server if an existing SQL server is not available. A downside to using a
physical server is that it does not offer the same high availability features as a virtual
deployment with shared storage. In a virtual deployment, XenClient would be able to
utilize the hypervisor‟s failover features if the host supporting the Synchronizer failed.
Figure 19: Synchronizer on Physical Hardware
 Virtual Machine: If a Hyper-V server is already in use within the organization, it is
recommended that the XenClient Synchronizer utilize that instance. In this scenario,
Synchronizer would be installed as a VM on the primary hypervisor. This hypervisor

can be Hyper-V, XenServer, or VMware but a Hyper-V server must be used for image
authoring. Installing the Synchronizer on a VM will help to reduce hardware
requirements and allow the Synchronizer server to benefit from existing high availability
features.
Figure 20: Virtual Synchronizer
Decision: Processor Specification
The XenClient Synchronizer can be processor intensive during certain processes. The
following list details the major factors that will increase processor usage on the Synchronizer as
well as recommendations to reduce their impact:
 Creating & Updating Images: The image creation and update process utilizes Hyper-
V to perform the necessary actions to the images. If Hyper-V is installed on the same
physical server as the Synchronizer, sufficient processor and memory should be pre-
allocated to prevent contention with the Synchronizer service.
 Publishing Images: The publishing process, which takes approximately 20-40 minutes
per image, is a very disk intensive process that causes some processor spikes.
Publishing occurs in a serialized fashion, only allowing one VM at a time to be
published. Therefore, images should be published during off peak hours. Since the
image publishing process takes a relatively short amount of time, this is just a
precaution.
 Deploying Images: The deployment process is the most processor intensive activity
due to the network traffic generated, ranging from 20 to 30GB per VM. Although the
initial deployment of images will be intensive, subsequent updates will be much smaller
and will be downloaded in the background. For this reason, it is important to allocate
resources based on normal business activity rather than the initial deployment of VMs.
The Synchronizer should be allocated three physical cores for every 1Gbps LAN connection.
An additional physical core should be available for the Synchronizer server process. At a
minimum, it is recommended that two CPU cores be assigned to the Synchronizer.
Decision: Memory Specification
The default recommendation is for 8GB of RAM to be assigned to the Synchronizer server.
Additional memory will greatly increase performance by reducing disk I/O when distributing
images to clients. If additional memory is available, Windows will attempt to cache images in
the System Cache. To improve performance by offloading work from the storage subsystem,
sufficient memory should be available to cache the compressed copy of the VM images. The
compressed version of a image is approximately 50% smaller than the original image size.
Decision: Network Specification

The Synchronizer performs bandwidth intensive operations when distributing VM images to
endpoint computers. These operations will utilize all available bandwidth until the operation is
complete.
The following items should be considered to overcome this potential bottleneck:
 Capacity Planning: The type of VM deployed (Custom vs. Shared) has a direct impact
on the future bandwidth load of the Synchronizer. If Custom VMs are chosen, they will
only be deployed a single time. Since incremental updates for Custom VMs cannot be
deployed, these endpoints will only contact the Synchronizer server for policy updates.
In contrast, Shared VMs can be patched causing increased bandwidth since the
incremental image updates are distributed to XenClient Engines.
 Quality of Service: A Quality of Service (QoS) policy can be used to effectively
manage bandwidth utilization between the XenClient Engines and the Synchronizer
server. The QoS policy should be defined based on the destination IP of the
Synchronizer server. A port based policy cannot be used since the Synchronizer listens
on 443 which is also used for secure website traffic.
 Bandwidth Policies: Policies can be created on the Synchronizer to limit the
bandwidth and number of connections that a Synchronizer can utilize. This will help to
ensure that the Synchronizer does not get overloaded with a large number of requests
at the same time. While bandwidth policies are important, they do not replace QoS
policies. Bandwidth policies enforce specified limits even when additional bandwidth is
available.
 Network Rule: XenClient policies are available which restrict the networks that can be
used to download and upload images. This policy is established by specifying the DNS
hostname for the networks that will allow downloads and uploads. Although this limits
flexibility, it can help to relieve congestion on networks that are known to have limited
bandwidth.
Decision: Synchronizer High Availability
The Synchronizer, which is an integral component of the XenClient infrastructure, does not have
any built-in high availability features. If the Synchronizer server were to fail, users would still be
able to operate normally although there would be some limitations. Users would not be able to
receive images and policy updates. Administrators would be even more restricted, losing the
ability to manage any of the endpoints. For the Synchronizer, true high availability comes from
a strong data backup strategy. The following items should be considered when designing a
resilient Synchronizer server:
 Virtual Machine HA: If the Synchronizer is deployed as a VM, the high availability
features of the hypervisor can be utilized. In this scenario, the Synchronizer is installed
on a hypervisor with a shared storage solution. This allows the Synchronizer VM to
failover in the event of a host failure.
 Data Backup: To guard against failures to shared storage, master VM images should
be periodically backed up or archived to a separate storage repository. This will allow
for the database to be easily recovered in the event of failure.
Decision: SQL Database
The Synchronizer relies on Microsoft SQL, which stores the entire list of endpoints, images, and
policies assigned to each computer. If the database is unavailable, all user and administration
functions will be unavailable. While the SQL database is very important, it does not consume a
large amount of resources. The following recommendations should be considered when
designing the SQL database for the Synchronizer:

 Server Location: While the SQL database can be installed locally on the Synchronizer
server, a highly available SQL environment is recommended. The SQL server should
be allocated at least 2GB of RAM on the system hosting the instance and at least
500MB of reserved disk space.
 Backups: Automated backups of the XenClient database should be made daily. While
downtime of the Synchronizer can be tolerated, the loss of configuration data would
cause irreparable damage.
Decision: Remote Synchronizer Servers
When the XenClient user base is distributed across different physical locations, a decision must
be made on whether additional Central Synchronizers or a Remote Synchronizer should be
deployed. A Remote Synchronizer is a subordinate of the Central Synchronizer that can be
placed at locations physically separated from the Central Synchronizer. Remote Synchronizers
allow users at these locations to retrieve and backup images locally instead of traversing the
WAN. The pros and cons of each deployment option are discussed below:
 Deploy Additional Central Synchronizers: Deploying additional Central Synchronizer
servers will create a new XenClient environment with a new database. Since a Remote
Synchronizer server requires database access, this may introduce reliability issues
when reading or writing data across the WAN. A Central server would allow the SQL
database to be hosted locally ensuring the reliability of the connection. The downside
of deploying additional Central servers is that they require separate databases and
cannot be managed from the same console. These limitations increase management
overhead since each Central server must be managed independently.
 Deploy a Remote Synchronizer: A Remote Synchronizer server acts as a subordinate
of the Central Synchronizer and is managed from the same management console. The
Remote Synchronizer shares the same SQL database as the Central Synchronizer
allowing the same policies and configurations to be used. A Remote Synchronizer
helps reduce bandwidth between the Engine and Central Synchronizer server by
intelligently caching content once a user has initially requested it. The downside to
using a Remote Synchronizer is that it must be able to reach the Central Synchronizer,
SQL database and Active Directory environment over the network from where it is
located. This may introduce some issues for high latency and low bandwidth sites.
Figure 21: Remote Synchronizer Architecture
If it is decided that a Remote Synchronizer should be deployed, the following items must be
considered:

 Storage: Enterprise level redundant storage must also be available at the locations
where the remote servers are located. This becomes even more important if user data
is being backed up to the remote server. While VM image data can be re-downloaded
from the central server, user data is only stored on the server that the users are
registered with.
 Management: Locating servers at remote locations can introduce issues for the IT staff
managing the remote hardware. The remote office where the Synchronizer is placed
may not have sufficient hosting resources or staff on hand that are able to resolve
potential issues. This adds to the complexity of the environment and introduces
additional risk areas.
 Firewall Configuration: Since remote servers share the same Microsoft SQL
database, the port used by the database server (1433) must be open on all firewalls
between the remote server and the central server.
 Active Directory: If configured, each remote server must have connectivity to an Active
Directory domain controller so that user passwords can be verified and computer
objects created.
Decision: Storage Repository Size and Location
The storage repository is the location where user data and VM images are stored. The
repository size is directly dependent on the number of images, their size and number of users in
the environment. Administrators can control the amount of data used to store the VM images
since they create and update them. However, administrators have less control over user data
that is backed up to the server. This data can grow at a fast rate depending on the amount of
data modified and the number of backups retained.
User Data Backups
User data consists of all data stored on the U: drive of the XenClient endpoint, which includes
the user profile and files stored in the My Documents folder. The table below provides an
estimate on the storage space required for a 500-user environment. The user disk, which is
100GB, has a compression ratio of 90% by default, reducing the storage space to 10GB per
user. The table is modeled on a „moderate user‟ that changes approximately 7.5% of their data
every time a backup occurs:
Financial: A mid-sized financial institution needs to provide multiple versions of their
financial software running on different operating systems to hundreds of remote
locations. These remote locations have very limited bandwidth so image updates from
XenClient are delivered from a Remote Synchronizer located at each site. This allows
for a one-to–many-deployment, since updates are deployed to a single server at each
site and then distributed locally to each computer.
Personal Fitness: A large personal fitness company was in need of a solution to isolate
general computing work from sensitive user data such as credit card transitions. In
order to become PCI compliant, these workloads must be separated and the physical
space to locate multiple computers at each desk was unavailable. XenClient has been
deployed to hundreds of fitness locations throughout the United States that do not house
any server equipment. Since Custom images have been deployed, small incremental
image updates cannot be pushed out to computers. For this reason, the fitness company
decided not to deploy Remote Synchronizer servers at each remote location. Their
current package management system will be used to maintain the Custom images.

Initial Backup
One
Backup
Two
Backup
Three
Backup
Four
Backup
Five
Backup
Six
Backup
Seven
Storage Consumed on Synchronizer (Per User)
Differential 0% 8% 15% 23% 30% 38% 45% 53%
Differential
Size (GB)
10 0.8 1.5 2.3 3.0 3.8 4.5 5.3
Cumulative
Size (GB)
10 10.8 12.3 14.5 17.5 21.3 25.8 31
Storage Consumed on Synchronizer (All Users)
Total (GB) 5000 5375 6125 7250 8750 10625 12875 15500
Total (TB) 4.88 5.25 5.98 7.08 8.54 10.38 12.57 15.14
Table 44: User Data Backups
The general recommendation is to keep at least two backup copies of user data. This will allow
for a moderate level of resiliency while significantly reducing storage requirements. It is
recommended that user data be hosted on RAID-10 storage to offer fast reads/writes and
redundancy.
Virtual Machine Images
XenClient can also require a large amount of storage for VM image data. Each VM is actually
stored twice, once compressed for image deployment and a second time that is uncompressed
for image updates. The compressed copy is a collection of TarGzip files, which are downloaded
to clients and remote servers. The compressed image typically requires 50% less storage than
the uncompressed image. Additional storage space is also consumed each time a VM image is
updated. For every update, additional differencing disk files will be created on the server.
These files form a chain linked back to the original image that allow it to be reverted back to
previous versions.
The following recommendations should be considered when designing storage space for virtual
images:
 Expansion: VM images grow in size each time a new version of the image is created.
Allocate sufficient space for each VM to grow approximately 50% of its original size.
 Performance: It is recommended that VM images be located on RAID-5 storage due to
its ability to offer redundancy and fast reads.
 Reclaim Space: A rollup operation should be performed once old versions of a VM
image become outdated. This process combines the differencing disk files of a VM
eliminating previous revisions up to the version selected. It is recommended that a
revision plan be established to help administrators decide how many versions should be
kept before a rollup occurs. As a general recommendation, at least two revisions of
each VM should be kept in case issues arise with the latest version.
 Daily Backup Rollup: The Daily Backup Rollup process consolidates VM backups that
have exceeded their retention limits. These limits are set through the backup policy
assigned to the VM. Rolling up old backups reduces disk space usage on the server. It
is recommended that this option be enabled and set to run during off peak hours.
Decision: External Access

To allow users to receive image and policy updates when they are outside the corporate
network, the Synchronizer must be accessible over the Internet. The following list provides
some of the benefits from enabling external access to the Synchronizer:
 VM Deployment: When accessible over the Internet, the Synchronizer will allow mobile
users located off the corporate network to receive VM images and policy updates. It is
recommended that remote access be configured for all organizations with mobile users.
 Lockout Policy: A lockout policy is set to enhance security by setting how long a
XenClient computer can operate before it must contact the Synchronizer server. If this
limit is reached, a user will be locked out of their computer, although their data would
still be intact. This policy exists to ensure that users receive the latest VM image and
policy updates. It is recommended that this policy be used in high security
environments to ensure that all Engines adhere to the latest image and policy updates
deployed. If this policy is enabled for remote users, the Synchronizer must be available
externally.
 Kill Switch: The kill switch option is available to provide a failsafe against stolen or
compromised XenClient computers. Making the Synchronizer available externally
enables endpoint computers to be terminated when they connect to the Internet. The
endpoint will receive the kill command once the policy update interval time has been
reached.
Note: As the kill switch option requires that the Engine have Internet connectivity, it
should be used in combination with the lockout policy.
If the Synchronizer has been deployed externally, aliases should be configured so that it can be
reached using the same hostname regardless of whether the Engine is internal or external.
When deploying the Synchronizer for external users, two options are available when exposing
the Synchronizer to external traffic:
 Behind Firewall: The Synchronizer server sits in the DMZ behind a firewall with only
the appropriate port (443) accessible externally.
 Behind Citrix NetScaler: Citrix NetScaler provides additional security and scalability
when external users connect to the Synchronizer. In this model, the NetScaler is
located in the DMZ and the Synchronizer is located inside the corporate network.
NetScaler acts as a „man in the middle‟ to ward off TCP/IP oriented exposures such as
denial of service (DOS) attacks. There are two deployment modes that can be chosen
when using a NetScaler:
o SSL Offload: Reduces the workload on the Synchronizer allowing it to achieve
improved single server scalability. The Synchronizer certificates are placed on
the NetScaler device so that it can offload the SSL processing from the
Synchronizer.
o SSL Bridging: SSL traffic is passed directly from the NetScaler to the
Synchronizer. Although SSL Bridging reduced the scalability of the
Synchronizer, it does not require that the Synchronizer certificates be placed on
the NetScaler.
For added security and scalability, it is recommended that Internet facing Synchronizers be
placed behind a NetScaler with SSL Offload configured.
For example, a central and a remote synchronizer have been deployed behind redundant
NetScaler devices. A remote synchronizer was deployed because the organization has two
locations from which a large number of XenClient users are located. The remote synchronizer
will allow users at „Data Center – Regional‟ to receive VM images from their local network,
helping to save bandwidth. The central server has been virtualized while the remote server is a
physical machine since no virtualization environment is available at „Data Center – Regional‟.

As the organization already utilizes NetScaler for load balancing web servers, using them to
protect the XenClient environment is an added benefit. The NetScaler will help to guard against
network level attacks such as Denial of Service (DOS) attacks. Additional features such as ACL
lists, surge protection and rate limiting can also be implemented to help prevent the
Synchronizer server from being overloaded by malicious traffic.
Figure 22: Central (virtual) and Remote Synchronizer (Physical) Servers
Decision: Active Directory Integration
The XenClient Synchronizer can integrate with Active Directory to join images to the domain,
authenticate users to their desktop, and assign administrator privileges. The Synchronizer
server itself does not need to join the domain. However, a XenClient service account should be
created with the correct privileges to create computer objects in the designated organizational
unit. The following recommendations should be considered when designing the integration
between the Synchronizer and Active Directory:
 Pre-Join Images to Domain: XenClient images can be deployed pre-joined to the
domain. When this option is chosen, XenClient automatically creates the computer
objects and generates unique hostnames for each VM. The hostnames can be
customized automatically using the operating system type, username, and random
number variables. While the hostnames can be tailored, only a single naming
convention is specified for every VM in the environment. This constraint may not
comply with organizational standards for naming to include location or user group in the
hostname.
 Administrator Login: It is recommended that administrator accounts to the
Synchronizer be assigned to active directory user accounts helping to streamline
management.
 User Synchronization: The Synchronizer has the ability to schedule automatic
updates to reflect any changes in Active Directory users and groups. It is

recommended that this feature be enabled to ensure Synchronizer has the correct user
group assignments.
 Active Directory SSL: It is recommended that SSL be enabled when connecting to the
Active Directory server for authentication. This will ensure that all username and
password data is encrypted.
Citrix Provisioning Services (PVS) uses streaming technology to simplify the deployment of virtual
and physical desktops. Computers are provisioned and re-provisioned in real-time from a single
shared-disk image. In doing so, administrators can completely eliminate the need to manage and
patch individual systems. Instead, all image management is performed on the master image.
Decision: Number of Farms
A Provisioning Services farm represents the top level of the Provisioning Services infrastructure.
All Provisioning Servers in a farm share the same SQL database and Citrix License server.
There are several factors that must be considered when determining the number of Provisioning
Services farms required:
 Network: Provisioning Servers are constantly communicating with the farm database to
retrieve system configuration settings. Therefore, separate farms should be created for
each physical location where target devices reside, unless they are connected to the
database server by a fast and robust connection.
 Criticality: Organizations need to determine how much risk they are willing to tolerate.
Although unlikely, a Provisioning Services farm outage could potentially affect the entire
organization. For organizations that require very high levels of availability, multiple
farms may be warranted.
 Administration: Organizations may need to maintain the separation of administrative
duties at a departmental, regional or countrywide basis. Additional Provisioning
Services farms will add some complexity to the management of the environment.
However, this overhead is typically limited to initial configuration, desktop creation and
image updates.
Decision: Number of Sites
Each Provisioning Services farm contains one or more sites. A site is a logical entity that
contains Provisioning Servers, vDisk pools, and target device collections. Multiple sites share
the same database; however target devices can only failover to other Provisioning Servers
within the same site. Additional sites may need to be created in the following scenarios:
 Networking: Sites can be used to control the Provisioning Services streaming traffic.
For example, a site can be created per site, rack, or blade chassis to ensure that
streaming traffic stays local and does not become a bottleneck.
 Organization: Another practical reason for building multiple sites is due to
organizational changes. For example, two companies may have recently merged
through acquisition, but need to keep resources separate while integration takes place.
Configuring the organization to use separate sites is one way to keep the businesses
separate but managed centrally through the Provisioning Services console.
Only create additional sites if the business requirements warrant it. A single site per farm is
easier to manage and requires no additional configuration.
Decision: Database Specification
The Provisioning Services database stores all system configuration settings that exist within a
farm. Provisioning Services 6.x supports the Express, Standard and Enterprise editions of SQL

2005, 2008 R2, and 2012. The decision on which SQL edition to use comes down to the level
of fault tolerance required. The following table compares SQL 2008 R2 features, relevant to
Provisioning Services, across editions:
Feature SQL Server Express SQL Server Standard SQL Server Enterprise
Max Number of CPUs 1 4 8
Max Memory Used 1GB 64GB 2TB
Max Database Size 10GB 524PB 524PB
Supports Mirroring Witness Only Yes (Single Thread) Yes
Supports Clustering No Yes (2 Nodes Only) Yes
Table 45: Database Features
The Provisioning Services farm database will rarely grow beyond 250MB in size, even with
large environments. When comparing the SQL Server editions, SQL Server Express is better
suited for testing or proof of concept environments since it lacks fault tolerance functionality,
and has a 1GB memory limitation. This holds true for SQL 2005 and 2012 editions as well.
The following formula can be used to estimate the size of the Provisioning Services database:
Decision: Database Offline Mode
If connectivity to the Provisioning Services database is lost, target devices and Provisioning
Servers in the affected farm will fail to boot. Target devices that have already connected to a
Provisioning Server will be unaffected.
The offline database support feature allows Provisioning Services to remain operational in the
event that connectivity to the database is lost. A snapshot is taken of the database at startup
and kept up-to-date by the Stream Service. If the connection to the database is lost, the Stream
Service uses the latest snapshot available to obtain details on the farm configuration. Once the
database connection is re-established, the Stream Process synchronizes changes back to the
database.
Note: Offline database support is disabled by default and is only recommended for use with a
stable farm running in production. It is not recommended when running an evaluation
environment or when reconfiguring farm components „on the fly‟. Only a farm administrator can
set this option.
Decision: Service Accounts
The Stream and SOAP services interact with many different components within the Provisioning
Services architecture, as shown in the following diagram:

PVS Admin
Console
Command
Line
Interface
PowerShell
Interface
Domain
Controller
SQL DB
Virtual
Desktop
Provisioning
Services
Network
Storage
Communications:
Green: SOAP Service interaction
Red: Active Directory interaction
Blue: Stream Service interaction
Figure 23: Provisioning Services Architecture
The Stream and SOAP services can be run in one of two different contexts:
 Specified User Account: High security organizations should use a specified user
account that is closely controlled and monitored. If the password of this account
expires, Provisioning Services will not function until the password has been changed or
reset.
Note: If the Provisioning Services store will be placed on a Windows file share, the
Stream and SOAP services must be configured to run using a specified user account,
which must be granted full permission to all files in the store. If the Stream and SOAP
services are not configured as user accounts, Provisioning Services will be unable to
access vDisk files on the remote server.
 Local System Account: Running the Stream and SOAP services under the context of
the Network Service account is easy to configure, and requires no additional
management.
Decision: Device Collections
Device collections provide the ability to create and manage logical groups of target devices.
Creating device collections simplifies device management by allowing actions to be performed
at the collection level rather than the target device level.
Device collections can represent physical locations, subnet ranges, chassis or different
departments within an organization. Collections can also be used to logically separate
production target devices from test and maintenance ones.
Consider creating device collections based on vDisk assignment so that the status of all target
devices assigned to a particular vDisk can be quickly identified.
Decision: Provisioning Server Memory
The Windows operating system hosting Provisioning Services partially caches the vDisks in
memory (system cache) reducing the number of reads required from storage. Reading from
storage is significantly slower than reading from memory. Therefore, Provisioning Servers
should be allocated sufficient memory to maximize the benefit from this caching process.
The following formula can be used to determine the optimal amount of memory that should be
allocated to a Provisioning Server:

Note: A good starting point for determining the minimum amount of server memory is to allocate
2GB of RAM per active desktop vDisk and 10GB of RAM per active server vDisk. For most
Provisioning Services implementations, the amount of RAM allocated to each Provisioning
Server will fall between 8 and 32GB.
Decision: Scale Up or Out
As the farm grows, administrators will need to decide whether to add more resources to the
Provisioning Servers or to add more Provisioning Servers to the farm.
There are a number of environmental factors that need to be considered when determining
whether the Provisioning Servers should be scaled up or scaled out:
 Redundancy: Spreading user load across additional less-powerful servers helps
reduce the number of users affected from a single Provisioning Server failure. If the
business is unable to accept the loss of a single high-specification server, consider
scaling out.
 Failover Times: The more target devices connected to a single Provisioning Server,
the longer it will take for them to failover in the event that the server fails. Citrix internal
testing has identified a failover time of approximately eight minutes for 1,500 target
devices with Provisioning Services 5.6 SP2.
 Data Center Capacity: The data center may have limited space, power and/or cooling
available. In this situation, consider scaling up.
 Hardware Costs: Initially, it may be more cost effective to scale up. However, there
will be a point where scaling out actually becomes more cost effective. A cost analysis
should be performed to make that determination.
 Hosting Costs: There may be hosting and/or maintenance costs based on the number
of physical servers used. If so, consider scaling up to reduce the long-term cost of
these overheads.
Decision: vDisk Store Placement
A farm can contain one or more vDisk stores. There are two primary options available for
hosting the vDisk store:
 Local Storage or Direct Attached Storage (DAS): DAS can be local or block level
storage that is directly accessible to the Provisioning Server, for example SATA, SCSI,
iSCSI, and Fibre Channel. Local vDisk stores can only be accessed by one
Provisioning Server at a time. Therefore, vDisks must be replicated, manually or
automatically, between relevant vDisk stores whenever a vDisk is created or changed.
Note: Placing the vDisk store on local storage is not a single point of failure because
Provisioning Services Load Balancing allows target devices to automatically failover to
other Provisioning Servers in the site should the local vDisk store fail.
 Network Attached Storage (NAS): NAS is a file based storage solution that is
accessed using a network redirector; NAS protocols include CIFs and NFS. Unlike
DAS, NAS vDisk Stores can be accessed by multiple Provisioning Servers at the same
time. NAS ensures vDisk consistency between servers in the farm since there is no
need for vDisk replication.
Note: Provisioning Servers before version 6.0 required registry changes to enable
opportunistic locks and SMB support in order to take advantage of the System Cache in
Windows for caching the contents of the vDisk.

The following RegKeys must be configured on every Provisioning Server (the sample
below outlines the RegKeys for W2K8R2):
o HKLMSYSTEMCurrentControlSetservicesLanmanWorkstationParameters
“EnableOplocks” = dword:0×00000001
o HKLMSYSTEMCurrentControlSetservicesmrxsmbParameters
“OplocksDisabled” = dword:0×00000000
“CscEnabled” = dword:0×00000001
o HKLMSYSTEMCurrentControlSetservicesLanmanServerParameters
“autodisconnect” = dword:0x0000ffff
“Smb2″ = dword:0×00000001
Note: Network Attached Storage can become a single point of failure. If the network
share becomes unavailable, all target devices streamed from the disk will be
unavailable as well.
Decision: Estimating Store Size
A vDisk consists of a VHD base image file, a properties file (.pvd), and may also contain a chain
of referenced VHD differencing disks (.avhd). Every time a vDisk is updated using the
Provisioning Services versioning method, a new differencing disk file is created.
When estimating the size of the vDisk store, you will need to consider:
 Total Size of vDisks: Add together the size of the planned vDisk images that will be
delivered.
 Maximum # of Versions in vDisk Chain: vDisk versioning simplifies vDisk update and
management tasks, providing a more flexible and robust approach to managing vDisks.
Differencing disks are created to capture the changes made to the base disk image,
leaving the original base disk unchanged. Each differencing disk that is associated with
a base disk represents a different version.
Note: When too many differencing disks are chained from the base image disk,
Provisioning Services performance begins to degrade. Therefore it is recommended
that the number of differencing disks should not exceed more than five to seven
versions deep.
 vDisk Version Change %: The size of the differencing disks will vary according to the
changes made to the vDisk. The more significant the change is to the vDisk, the larger
the differencing disk that will be created. If it is anticipated that new applications will be
added to the vDisk then plan for differencing disks that could be large in size.
The following formula can be used as a guide when estimating the vDisk store:
Consider the following example:
You plan to deploy three vDisk images:
 Windows 7 (x64) image = 40GB
 Windows 7 (x86) image = 35GB
 Windows XP image = 20GB

Each vDisk will not contain more than 5 differencing disks. You anticipate a differencing
disk approximately 20% the size of the master vDisk image. The estimated size required
for the Provisioning Services store will be:
Decision: RAID Levels
The RAID level of a storage sub-system can have a direct impact on the performance of
associated applications and workloads. RAID levels 0, 1, and 10 are optimized for writes while
RAID levels 5 and 6 are optimized for reads. For Provisioning Services, the different RAID
levels can be categorized as follows:
RAID Capacity Fault Tolerance Read
Performance
Write Performance Minimum #
of Disks
0 100% None High High
(Write Penalty 0)
2
1 50% Single-drive failure High Medium
(Write Penalty 2)
2
5 67%-94% Single-drive failure High Low
(Write Penalty 4)
3
10 50% Single-drive failure
in each sub array
High Medium
(Write Penalty 2)
4
Table 46: RAID Levels
Capacity, read performance, and write performance depend on the number of drives in the
array. In general, the more drives, the better the performance.
Provisioning Services is read intensive for vDisks and write intensive for the Write Cache:
 Write Cache: RAID 0, 1, and 10 are optimal choices since they have low write
penalties.
 vDisk Store: RAID 0, 1, 5 and 10 are optimal because they allow read operations to be
spread across all disks within the RAID set simultaneously.
RAID 0 is not a good choice to use since it offers no redundancy. If RAID 0 were to be chosen
for the vDisk Store and a disk were to fail, Provisioning Services would stop streaming the disk.
If possible, place the vDisks on RAID 5 storage and the Write Cache on RAID 10 storage. If the
vDisks and Write Cache are placed on the same storage, then select RAID 10 since it offers the
best fault tolerance and read/write performance.
Decision: Write Cache Placement
If data is written to the Provisioning Server vDisk in caching mode, the data is not written back
to the base vDisk. Instead, it is written to a Write Cache file in one of the locations specified
below:
 Cache on Device Hard Drive: The write cache is placed on the target device‟s hard
drive. This option frees up resources on the Provisioning Server since it does not have
to process write requests. While not as fast as RAM, this option provides a fast
response time because the read/write to and from the write cache is local.
 Cache on Device Hard Drive Persisted: The option is the same as cache on device
hard drive, except that the cache persists between reboots. This feature is
experimental and only supported on Windows 2008 R2 and Windows 7.

 Cache in Device RAM: The write cache is placed in the RAM of the target device.
Although Cache in Device RAM offers high levels of performance it is expensive to
implement and will negatively affect stability if RAM is exhausted.
 Cache on Provisioning Server Disk: The Write Cache is stored on the Provisioning
Server as a temporary file. All writes are handled by the Provisioning Server, which can
increase disk IO and network traffic. This configuration is easy to setup and requires no
additional resources, however performance of the target devices is affected since all
read and write requests to the cache must cross the network. Placing the write cache
on the Provisioning Server is generally only recommended when testing and not for
production.
Note: For additional security, the Provisioning Server can be configured to encrypt write
cache files. Since the write-cache file does exist on the server between reboots, the
data can be encrypted in the event the hard drive is stolen.
 Cache on Provisioning Server Disk Persisted: This option is the same as cache on
the Provisioning Server disk, except changes are saved between reboots. The benefit
of this option is that target devices are able to save changes made to the vDisk image
and have them persist after a reboot. Any changes made to the vDisk will force the
cache file to be marked invalid. For example, if the vDisk is set to Private Image Mode,
all associated cache files are marked invalid. Any of the following changes listed below
will also cause the cache file to be marked invalid.
o Placing a vDisk in Maintenance mode
o Mapping the drive from the console
o Changing the location of the write cache file
o Using Automatic update
The write cache uniquely identifies the target device by including the target device‟s MAC
address and disk identifier. A target device can be assigned to multiple vDisks and therefore
have multiple cache files associated to it.
Placing the write cache on the target device hard drive is recommended for most
implementations because it offers fast performance without consuming expensive RAM and
does not reduce the scalability of the Provisioning Servers.
Decision: Write Cache Size
There are many variables that go into determining the proper size for the Write Cache drive.
Much of it is based on estimation and typical user profile sizes. Selecting a size that‟s too large
will result in wasted storage, while selecting a size that‟s too small can negatively affect stability.
The Write Cache should be large enough to store the following data:
 Temporary application data
 User profile temp files
Gas Distributor: A gas distribution company is designing a Provisioning Services
solution to support five hundred virtual desktops. The design has the following
requirements: it must be able to support a wide variety of desktop users, it must offer
fast performance, and any problems with one desktop should not impact the other
desktops. To meet these requirements, the gas distribution company went with a design
that places the write cache on the target device local storage.

 Windows Event Logs
 Citrix Logs
 Antivirus pattern files
 Citrix Application Streaming/Microsoft App-V Cache
Note: User profiles with folder redirection enabled can help reduce the size of the write cache.
The size of the write cache is highly dependent on the applications that will be installed or
streamed to the vDisk, and the size of the user‟s profile. A good starting point for estimating the
size of the write cache is 2GB. For most virtual desktop implementations the size of the Write
Cache is presented in the following table:
User Type Write Cache Size
Light 2GB
Normal 5GB
Heavy 10GB
Table 47: Average Write Cache Sizes
In addition, the Windows pagefile is also written to the same drive as the write cache. Therefore:
For example, if planning for 1.5 GB pagefile for a Windows 7 vDisk, a safe estimation for the
write cache drive would be:
Since most hypervisors can‟t create a disk size that precise, round up to the nearest whole
number, which is 4 GB in this case.
Decision: vDisk Format
Provisioning Services supports the use of fixed-size or dynamic vDisks:
 Fixed-size Disk: For vDisks in Private mode, fixed-size prevents disk fragmentation of
the disk hosting the vDisk, and offers improved write performance over dynamic disks.
However in Private mode, vDisk access is limited to one target device, which is only
practical when performing maintenance on the vDisk.
 Dynamic Disk: Dynamic disks require less storage space than fixed-size disks, but
offers significantly lower write performance. Although vDisks in Shared mode do not
perform writes to the vDisk, the time required to complete vDisk merge operations will
increase with dynamic disks.
If the Provisioning Servers have been provided with sufficient RAM to cache the contents of the
vDisk in its System Cache, when the vDisk is in Shared mode, there are fewer read IOPS that
get serviced by the underlying physical disk subsystem of the Provisioning Server. Since most
reads will be to the System Cache in RAM, there is no significant change in performance when
utilizing fixed-size or dynamic disks. There is a cost savings in storage when using dynamic
disks, since disk space is used more efficiently. Therefore dynamic disks are recommended in
this scenario.
Decision: vDisk Size
The size of a vDisk depends on the operating system used and the number of applications
installed into the vDisk. The following table provides a starting point for estimating vDisk sizes:

Operating System Microsoft minimum
disk requirement
Estimated vDisk Size
Windows XP 1.5GB 15GB
Windows Vista 15GB 40GB
Windows 7 16GB (32-bit), 20GB
(64-bit)
40GB
Table 48: vDisk Size Estimates
If the size selected for the vDisk is too small or too large, the vDisk can be resized later on in
Provisioning Services. To accurately estimate the size of the vDisk, follow these guidelines:
 Target Device Space: Identify the amount of disk space currently in use by the
computer that will serve as the master target device.
 Application Sizing: As applications are installed, the amount of space in use
increases. If applications are to be installed into the vDisk image after capture, then the
vDisk should be sized appropriately. Plan for approximately 25% growth to the initial
footprint of the image.
Organizations will need to perform their own internal tests to determine a proper size for all
applications that will be installed on the vDisk, plus some additional room for future growth.
Decision: vDisk Replication
vDisks hosted on Direct Attached Storage must be replicated between vDisk Stores whenever a
vDisk is created or changed. Provisioning Services supports the replication of vDisks from
stores that are local to the Provisioning Server as well as replication across multiple sites that
use shared storage. The replication of vDisks can be performed manually or automatically:
 Manual: Manual replication is simple, but can be time consuming, depending on the
number of vDisks and vDisk stores. If an error occurs during the replication process,
administrators can catch them straight away and take the appropriate steps to resolve
them.
 Automated: For large environments, automated replication is faster than the manual
method due to the number of vDisks and vDisk Stores required. Some automated
tools, such as Microsoft DFS-R, support bandwidth throttling and Cross File Remote
Differential Compression (CF-RDC), which use heuristics to determine whether
destination files are similar to the file being replicated. If so, CF-RDC will use blocks
from these files to minimize the amount of data transferred over the network. The risk
of automated replication is that administrator do not typically monitor replication events
in real-time and do not respond quickly when errors occur, unless the automation tool
has an alerting feature. Some tools can be configured to automatically restart the copy
process in the event of a failure. For example, Robocopy supports “resume copying” in
the event that the network connection is interrupted.
For medium and large-sized projects, use a tool to automate vDisk replication. Select a tool
that is capable of resuming from network interruptions, copying file attributes and preserving the
original timestamp.
Note: Load balancing and high availability will not work unless the vDisks have identical
timestamps.
Decision: Microsoft Volume Licensing

Provisioning Services supports Microsoft Key Management Service (KMS) and Multiple
Activation Key (MAK) volume licensing:
 KMS Volume Licensing: KMS volume licensing allows organizations to activate their
Microsoft products on their local network, eliminating the need for individual computers
to connect to Microsoft for product activation. A single KMS host can support an
unlimited number of KMS clients; however Microsoft recommends deploying at least
two KMS hosts for failover.
Additional KMS considerations include:
o Office 2010 deployments only support KMS license activations.
o When deploying Windows 7 or Office 2010, Windows Server 2008 is not
supported as a KMS host. Instead, Windows Server 2008 R2, Windows Server
2003, or Windows 7 VL edition must be used.
o When deploying Office 2013, Windows Server 2003 and Windows Server 2008
is not supported as a KMS host. Instead, use Windows Server 2008 R2, or
Windows 7 VL edition.
For more information, please refer to Citrix eDocs - Configuring Microsoft KMS Volume
Licensing.
 MAK Volume Licensing: MAK volume licensing is used for one-time activation with
Microsoft‟s hosted activation services. There are two ways products can be activated
using MAK. The first method requires each computer to independently connect and be
activated. The second method is MAK proxy activation, where a computer running the
Volume Activation Management Toolkit (VAMT), gathers information from multiple
computers on the network and sends the activation request on their behalf.
Provisioning Services leverages this proxy activation mechanism for Standard Image
mode vDisks that have MAK licensing mode selected. In order for MAK licensing to
work, the Volume Activation Management Toolkit (VAMT) must be installed on all
Provisioning Servers within a farm.
MAK is recommended for computers that rarely or never connect to the corporate
network and for environments that do not meet the KMS activation threshold of five
Windows Server 2008/2008R2/2012 devices or twenty-five Windows Vista/7/8 devices.
Before target devices can be activated, MAK user credentials must be entered for a
site. The user must have administrator rights on all target devices, and on all
Provisioning Servers that will stream the vDisks to the target devices. After all licenses
for a given MAK have been used, a new key is required in order to allow additional
target devices to share the vDisk image.
Note: To update a MAK activated vDisk, the AutoUpdate feature must be used so that
the MAK activation information is maintained.
For more information, please refer to Citrix eDocs - Configuring Microsoft MAK Volume
Licensing.
For virtual desktop deployments of 50 computers or less, MAK licensing is recommended. For
all other deployments, KMS licensing is recommended.
Decision: High Availability
Provisioning Services is a critical component of the virtual desktop infrastructure. The following
recommendations should be followed to eliminate single points of failure:
 Database: Provisioning Services provides two options for supporting a highly available
database configuration:

o Offline Database Support, which is covered in the Farm Design Decisions
section earlier in this document.
o The Provisioning Services SQL database can be part of a mirrored
configuration. In the event that the primary version becomes unavailable,
Provisioning Services can work from the mirrored version. For more
information, please refer to the Citrix Knowledgebase article CTX129993 -
Provisioning Services 6.0 Administrative Guide.
 Provisioning Server: All Provisioning Servers within a site can be configured to
provide the same vDisks to target devices. A minimum of two Provisioning Servers
should always be implemented per site. Sufficient redundancy should be incorporated
into the design so that a single server failure does not reduce the total number of target
devices that can be supported per site.
The Provisioning Services boot file should be configured for high availability. Up to four
Provisioning Servers may be listed in the boot file. Target devices will try to contact the
servers in the order that they are listed. The server that responds may not necessarily
be the server that will provide streaming services to the target device. If Load
Balancing is enabled, the target device may be reassigned to another server in the site
that is less loaded than the others.
 vDisks and Storage: For vDisk Stores hosted on Direct Attached Storage, replication
should be used to synchronize the vDisks. vDisks using Network Attached Storage
should be hosted on a highly available network share.
Note: When using Direct Attached Storage and replicating the vDisks, it is important
that the timestamp is identical on all vDisks, or failover will not occur. Some automatic
replication tools may not preserve the timestamp.
 Networking: The Provisioning Servers network adapters should be teamed, and
Provisioning Services NIC failover enabled by specifying the NICs that the Provisioning
Services drivers bind to.
Note: The target devices will only failover to NICs that are in the same subnet as the
PXE boot NIC.
There are several options available to make the TFTP service highly available. Some of
the more commonly used options are:
o DNS Round Robin: A DNS entry is created for the TFTP service with multiple
A records corresponding to the TFTP services running on the Provisioning
Servers in the farm. This method is not recommended since the state of TFTP
service is not monitored. Clients could potentially be sent to a non-working
server.
o Hardware load balancer: Use a hardware load balancer, such as the Citrix
NetScaler, to create virtual IPs that corresponds to the Provisioning Servers.
The NetScaler can intelligently route traffic between the Provisioning Servers.
In the event one of the servers becomes unavailable, the NetScaler will not
route TFTP requests to that server.
o Proxy DHCP: Use the Provisioning Servers PXE service to provide the
bootstrap information. If one of the servers is down, the next available server in
the farm can provide the bootstrap information.
o Boot Device Manager: Use the Boot Device Manager to create an ISO file and
place it on a shared network location. Configure the target devices to boot from
the CD/DVD-ROM drive and specify the location of the ISO file. An alternate
method is to use the Boot Device Manager to write the bootfile information to a

hard drive partition local to the target device. When either method is utilized,
the TFTP service is not used at all.
For more information, please refer to the Citrix whitepapers - High Availability for TFTP
with Provisioning Services and Provisioning Services Networking Planning Guide.
High availability should always be incorporated into the Provisioning Services design. Although
high availability may require additional resources and increased costs, it will provide a highly
stable environment so that users experience minimal impact due to service outages.
Decision: Bandwidth Requirements
It is imperative that available bandwidth is sized correctly to prevent network bottlenecks
affecting virtual desktop performance. Network utilization is most significant when target
devices are booting. The amount of data required to boot the target device will vary by
operating system. The necessary time required by target devices to boot can be estimated
using the following formula:
For example, 500 Windows 7 target devices requiring 40MB to boot with 1Gbps Ethernet:
1. 1 GB Ethernet = 125MB/s
2. 500 targets x 40 MB = 20GB
3. 20GB/125MBs = 160 seconds
The following table shows the approximate bandwidth and load times when booting 500 target
devices on a 10GB network using various Windows desktop operating systems.
Operating System Avg Boot Data Usage (MB) Bandwidth (Mb) Load Time (s)
Windows 7 x86 166 83,000 66.4
Windows 7 x64 210 105,000 84
Windows Vista x86 190 95,000 76
Windows Vista x64 240 120,000 96
Windows XP x86 90 45,000 36
Windows XP x64 240 120,000 96
Table 49: Sample calculations when booting 500 target devices on a 10GB network
Tip: Firewalls can add latency and create bandwidth bottlenecks in Provisioning Services
environments. Whenever possible, firewalls should be avoided between Provisioning Services
infrastructure components. If the use of firewalls cannot be avoided, the following ports should
be enabled for full functionality:

Source/Destination Service Ports
Target Device  Infrastructure Servers PXE
DHCP
TFTP
Broadcast
UDP 67, 68
UDP 69
Provisioning Servers  Infrastructure
Servers
SQL TCP 1433
Provisioning Server Admin Console 
Provisioning Server / Infrastructure Servers
SOAP Service
Active Directory
TCP 54321, 54322
TCP 389
Provisioning Server  Target Device Target Device
Streaming Service
Boot login
Write Cache
UDP 6901
UDP 6910-6930
UDP 6910
UDP 10802, 10803
Provisioning Server  Provisioning
Server
Server communication
(version 6.0 and later)
(Prior to version 6.0)
UDP 6890-6904
UDP 6905-6909
Table 50: Provisioning Services Port Requirements
For a complete listing of all ports used by Citrix products please refer to the Citrix whitepaper
CTX101810 - Communication Ports Used By Citrix Technologies.
Decision: Network Configuration
Providing Provisioning Servers with a 10Gbps network will provide sufficient bandwidth for most
environments. The use of 1Gbps Ethernet typically provides sufficient bandwidth for target
devices, but can become a bottleneck for Provisioning Servers. Teaming multiple NICs can
provide greater throughput, increasing network performance and help to alleviate potential
bottlenecks. Teaming NICs also prevents them from becoming a single point of failure. The
table below lists Provisioning Server interface options, from most to least capacity:
Interface Bandwidth Justification
Chassis 10Gbps+ 80Gbps+ PVS servers, targets, and infrastructure servers hosted
within the same blade chassis offer the maximum
bandwidth potential across a common backplane.
Bonded/Teamed
10Gbps
Up to 80Gbps Outside of a chassis option. Bonding/Teaming offers
the best capacity and resiliency.
Single 10Gbps 10Gbps PVS servers with a 10Gbps interface are becoming the
most broadly deployed configuration. A single network
interface creates a single point of failure and is not
recommended.
Bonded/Teamed
1Gbps
Up to 8Gbps When 10 Gb is not available Bonding/Teaming is the
ideal option to bolster server bandwidth using multiple
interfaces without the added complexity of multi-
homing.
Single 1Gbps 1Gbps Functions for small environments, but 1Gb targets
quickly overrun 1Gb servers. A single network interface
creates a single point of failure and is not
recommended.
Table 51: Provisioning Server NIC Speeds

For Provisioning Services implementations on low bandwidth networks (1Gbps or slower),
performance may be improved by isolating streaming traffic from other network traffic on the
LAN, except for small implementations.
Note: Microsoft does not support NIC teaming with Hyper-V on Windows Server 2008 or 2008
R2; however, third party solutions are available. All support queries regarding teaming with
Hyper-V should be directed to the NIC OEM. For more information, please refer to Microsoft
Knowledgebase Article KB968703 - Microsoft Support Policy for NIC Teaming with Hyper-V.
The following NIC properties should be modified on the Provisioning Servers and target
devices:
Property Justification
TCP Large Send
Offload
TCP Large Send Offloads allows the AIX TCP layer to build a TCP message up
to 64 KB long and send it in one call down the stack through IP and the
Ethernet device driver. The adapter then re-segments the message into
multiple TCP frames to transmit on the wire. Re-segmenting and storing up
packets to send in large frames causes latency and timeouts to the
Provisioning Server.
Checksum Offloading When data comes in through a network, it is verified using a checksum in the
packet header. If the data and checksum do not match, the packet is
determined to be bad and has to be re-transmitted. To improve performance,
some network cards can “offload” the checksum process to the network card,
freeing up the CPU to do more work on its own. However, in Windows XP and
Windows Server 2003 this has been known to cause the network link to drop
unexpectedly. To prevent this from affecting Provisioning Services, Citrix
recommends disabling checksum offloading.
Auto Negotiation Auto negotiation can cause long starting times and PXE timeouts, especially
when starting multiple target devices. To avoid this, hard-code all Provisioning
Services ports (server and client) on the NIC and on the switch.
Jumbo Frame On 10 GB networks, enabling jumbo frames improves Provisioning Services
performance by increasing throughput while consuming fewer CPU cycles.
Ensure all servers in the site have jumbo frames enabled to ensure consistent
performance.
Table 52: NIC Properties
For high availability, ensure that the NICs on the Provisioning Servers and target devices are
teamed. For optimal performance, use 10Gbps networks and keep the latency between the
Provisioning Servers and the target devices as low as possible. If security policies allow,
disable firewalls on and between the Provisioning Servers and target devices.
Decision: Hyper-V Networking
Hyper-V supports two types of network adapters: Legacy and Synthetic. The synthetic adapter
is a better performing adapter that is installed by default when a VM is created. In Provisioning
Services environments, however, the legacy network adapter must be used in order to PXE
boot the desktops. As such, two options must be considered during the Hyper-V design:
 Only use the legacy network adapter, which does not perform as well as the synthetic
adapter, and adds overhead to the hypervisor since the network card has to be
emulated.
 Configure the Provisioning Services target devices with dual network adapters: one
synthetic and one legacy adapter. Once the virtual desktop PXE boots, the synthetic
adapter has precedence over the legacy adapter, but Provisioning Services traffic is
bound to the legacy adapter and is not able to take advantage of the better performing
synthetic adapter. Additional DHCP addresses are also required with this configuration.

The legacy network connection will be suitable for all but the most network intensive
environments. For simplicity, utilize the legacy network adapter unless an associated network
bottleneck is identified.
Decision: Switch Settings
The network switches used by the Provisioning Servers and target devices should be optimized
for broadcast and streaming traffic:
 Spanning Tree Protocol / PortFast: The Spanning Tree Protocol prevents loops from
being formed when switches or bridges are interconnected via multiple paths. The
Spanning Tree Protocol places ports on a switch into a blocked state while it submits
Bridged Protocol Data Units (BPDUs) and listens to ensure the BPDUs are not in a
loopback configuration. The time it takes to complete this process depends on the size
of the switched network, which may cause the Pre-boot Execution Environment (PXE)
to time out. Citrix recommends disabling the Spanning Tree Protocol or enabling
PortFast. For more information, please refer to the Citrix Knowledgebase article
CTX117374 - Best Practices for Configuring Provisioning Server on a Network.
 Storm Control: The Cisco Catalyst line of switches has a feature called Storm Control
which prevents LAN traffic from being disrupted by a broadcast, multicast, and unicast
storm on one of the physical interfaces. Provisioning Servers communicate with clients
using Unicast UDP. If the storm control threshold on the Cisco switch port interface is
set too low or too strict, target devices may experience performance issues and
intermittent freezing. To avoid this behavior, change the threshold for Unicast traffic to
a greater value than what the target devices require, or disable Unicast filtering on that
switch port. For more information, please refer to the Citrix Knowledgebase article
CTX121618 - Target Device Performance and Freezing Issue.
 Broadcast Helper: Broadcast Helper is a switch feature that directs broadcasts from
server to clients that reside on a different subnet. When using Provisioning Services, it
is necessary to forward PXE boot requests when clients are not on the same subnet as
the Provisioning Servers. When this feature is enabled the switch behaves as a router
by forwarding PXE and TFTP requests from target devices to specific Provisioning
Servers in different subnets.
Decision: Bootstrap Delivery
A target device initiates the boot process by first loading a bootstrap program which initializes
the streaming session between the target device and the Provisioning Server. There are three
methods in which the target device can receive the bootstrap program:
 Using PXE with DHCP Options:
1. When the target device boots, DHCP sends a request to Provisioning Services
for an IP address and Scope Option settings 66 (the name or IP address of the
Provisioning Services TFTP server) and 67 (the name of the bootstrap file).
The Provisioning Server returns the information requested.
2. Using TFTP, a request for the bootstrap file is sent from the target device to the
Provisioning Server. The Provisioning Server downloads the boot file on the
target device.
3. The target device boots the assigned vDisk image.
Requires UDP/DHCP Helper to be configured when targets are not on the same subnet
as the PVS servers in order to receive PXE broadcasts.

 Using PXE without DHCP Options:
1. When a target device boots from the network, DHCP sends a request to the
Provisioning Server for an IP address and Scope Option settings 60 (PXEClient
identifier). The Provisioning Server returns the information requested.
2. The target device sends a request to the Provisioning Server for the bootstrap
file name and location to the PXE servers. The PXE service returns the
information to the target device.
3. Using TFTP, a request for the bootstrap file is sent from the target device to the
Provisioning Server. The Provisioning Server downloads the bootstrap file to
the target device and the target device boots.
Note: Make sure no other PXE services are in use on the same subnet like the Altiris
PXE service or conflicts may occur with Provisioning Services.
Requires UDP/DHCP Helper to be configured when targets are not on the same subnet
as the PVS servers in order to receive PXE broadcasts.
 Using Boot Device Manager: The Boot Device Manager (BDM) creates a bootfile that
target devices obtain through an ISO image mounted from a network share, or the
bootfile information is written to a hard drive partition local to the target devices
(preferably the write cache drive).
In the event the organization is not able to provide PXE or TFTP services, or allow changes to
the DHCP scope options, consider using the Boot Device Management utility to provide the
bootstrap file via an ISO file.
Decision: Auditing
Provisioning Services provides an auditing tool that records configuration actions on
components within the Provisioning Services farm. Auditing provides administrators with the
ability to troubleshoot and monitor recent changes that might impact system performance and
behavior. The Provisioning Services administrator privileges determine the audit information
that can be viewed and the menu options that are visible.
By default the auditing feature is disabled. When enabled, the audit trail information is written to
the Provisioning Services database along with general configuration data. The data written to
the database is not archived automatically and will continue to grow indefinitely until archived
manually.
Objects that can be audited include tasks performed on the farm, site, Provisioning Servers,
collections, target devices, store, and vDisks. Only tasks, which are performed through the
Admin Console, PowerShell, SOAP Server, or command line interface, are audited. Since
auditing is capturing user initiated actions it will have no impact on the performance of
streaming services. Tasks that are not performed by these methods are not audited, such as
booting target devices.
For organizations with high security concerns, enable auditing for tracking purposes. Ensure
that archiving is performed on the database at regular intervals in order to prevent indefinite
growth.

Decision: Antivirus
By default, most antivirus products scan all files and processes, which will have a significant
impact on performance. For optimal performance, antivirus software should be optimized for a
Provisioning Services environment. For more information, please refer to CTX124185 -
Provisioning Services Antivirus Best Practices.
Antivirus software can cause file-locking issues on Provisioning Servers by contending with files
being accessed by Provisioning Services. The vDisk Store and write cache should be excluded
from any antivirus scans in order to prevent file contention issues.
When a virtual disk is running in standard mode and needs to be restarted, it downloads all of
the previously loaded virus definitions. This can cause performance degradation when
restarting several target devices at a time, often causing network congestion while the operation
persists. In extreme cases, the target device and Provisioning Server can become sluggish and
consume more resources than necessary. If the antivirus software supports it, definition files
should be redirected to the write cache drive so that they are preserved between reboots.
Infrastructure Controllers
Active Directory
Active Directory provides a global management infrastructure and resource directory that stores
information about users, computers, services, and other objects of an organization. It is required in
almost every enterprise infrastructure for authentication and authorization and is therefore also a
vital piece of the overall Citrix infrastructure. Active Directory is used for resource assignments and
can be used to assign Citrix specific policies to user groups and desktops. A well-defined
integration of the Citrix components into an existing Active Directory structure ensures a consistent
user experience and can simplify operational procedures and maintenance of the infrastructure.
Decision: Active Directory Membership
Most Citrix components integrate with Active Directory for their operation as outlined earlier in
this chapter. Selecting an Active Directory forest and domain for housing the Citrix components
such as Controllers, virtual desktops or user accounts is the first key decision which has to be
made. The simplest and most widely used approach is to locate all components and user
accounts within the same forest and domain. This ensures a fast and simple authentication and
authorization process between all components. In some enterprise or service provider
environments a more sophisticated Active Directory structure is required for security or
delegated administration purposes. For these scenarios it is important to consider the
requirements of the Citrix components as outlined below:
XenDesktop Considerations
XenDesktop uses Active Directory's built-in Kerberos infrastructure to sign and encrypt network
traffic to guarantee the authenticity and confidentiality of desktop and controller communication.
Therefore a minimum level of trust relationship in between various domains and/or forests is
required.

Domain which
contains
Trust required Domain which contains Without the trust…
Controllers
(Bi-directional trust is
recommended)
(Optional: Controllers
domain trusts the VDA
domain using a selective
one-way trust
1
)
VDAs
Controller  VDA
communication is not
possible
(Controllers domain must
trust the users domain)
Users
Virtual desktop access
cannot be granted to users
from un-trusted domains
(DB domain must trust
the Controllers domain)
Database Server
Controllers cannot access
the site database
VDA
(VDA domain must trust
the users domain)
Users
User cannot access
desktops in un-trusted
domains
XenDesktop
Studio
(License Server domain
must trust the Studio
domain)
License Server
License management from
within Studio is not
possible
(Controllers domain must
trust the Studio domain)
Controllers
Management of the
XenDesktop site is not
possible
(DB domain should trust
the Studio domain)
Database Server
Studio user will be
prompted to enter
administrative credentials
when performing “non-
runtime” tasks such as
updating the DB schema or
removing a controller from
the site.
Table 53: Active Directory Domain Trusts for XenDesktop
Note: For scenarios where the XenDesktop components are located in different forests,
external trusts or forest trusts must be implemented accordingly. Further information can be
found within CTX134971 - Successfully Deploying XenDesktop 5.x in a Complex Active
Directory Environment.
Note: Please refer to CTX127998 - Database Access and Permission Model for XenDesktop 5
for further information in regards to database access of XenDesktop Controllers and/or Studio.
XenApp Considerations
XenApp can be integrated into Active Directory for the authentication and authorization of users
and administrators. The XenApp server(s) which provide resources to the users must be
members of a domain or a forest which trusts the user‟s domain or forest.
Note: XenApp supports trust-based routing; XenApp servers in domains that do not trust each
other can still be members of the same farm. When a server needs to perform an operation on

an untrusted domain, the requesting server determines from the data store which servers can
perform the operation and routes the request to the appropriate server.
Storefront Considerations
Storefront must be integrated into the same Active Directory forest as the XenDesktop site(s) or
XenApp farm(s) which provide resources to the users. Two-way trusts are required between
the domain containing the Storefront server(s), the domain containing the database server(s)
and the domain containing the XenDesktop and/or XenApp resources.
Decision: Organizational Unit Structure
All Active Directory objects, such as computers, users and user groups are structured by means
of Organizational Units (OUs). A well-designed OU structure will help to simplify operational
tasks such as policy assignments, worker group management and the delegation of
administrative user rights. This section provides sample OU structures for XenDesktop and
XenApp environments.
XenDesktop
In the following sample OU structure the Active Directory objects required for the XenDesktop
infrastructure are split into two groups:
Central Infrastructure
The “Citrix” OU contains all central
components, such as XenDesktop
Controllers or Storefront Servers.
In addition pooled (non-persistent)
desktops are located underneath a
dedicated “Desktops” OU. In this
sample the virtual desktops are
organized based on their
respective XenDesktop Machine
Catalogs. In scenarios with
multiple Delivery Groups per
Machine Catalog a structure based
on Delivery Groups is preferable.
Local Infrastructure
Virtual desktops assigned to
individual users are located
underneath a location specific OU,
in order to allow delegated
administration of the virtual
desktops. This OU structure
simplifies the assignment of
location specific policies and/or
logon scripts to the virtual
desktops.

Figure 24: Sample XenDesktop OU structure
In scenarios without branch specific admin teams, assigned desktops should be located within
the central “Citrix” OU as shown below:
Figure 2: Sample XenDesktop OU structure for centralized administration
XenApp
In the following sample OU structure the Active Directory objects required for the XenApp
infrastructure are located within a single “Citrix” organizational unit. The XenApp Controllers are
located within the “Controllers” sub-OU, while the infrastructure systems and XenApp related
user groups are located within separate sub-OUs.

Figure 25: Sample XenApp OU structure
Locating the XenApp workers in a separate organizational unit helps to simplify delegated
administration and policy assignment. This structure also enables the Citrix administrators to
configure OU based Worker Groups, which can help to simplify the management of XenApp
workers.
Decision: Resource and User Right Assignments
Most Citrix products rely on Active Directory for authentication and authorization services.
Centralized resources, such as published applications or virtual desktops, can be assigned to
individual user accounts or Active Directory user groups. Citrix recommends that resources are
assigned to user groups rather than individual user accounts to simplify the management of the
environment. For efficient management of the resource and user right assignments, especially
in multi-domain environments, a well-designed user group strategy is vital. Citrix recommends
following the Microsoft AGDLP user group design principle for this purpose, which is outlined in
the diagram below:
Figure 26: AGDLP design principle
The AGDLP principle states that user Accounts should be member of a domain Global user
group which is member of a Domain Local user group to which Permissions or, in Citrix
environments, resources are assigned.
This user group nesting approach enables organizations to delegate tasks from central
administrators to local administrators very easily. The following diagram outlines the AGDLP
approach in a Citrix environment:

Figure 27: AGDLP design principle applied to XenDesktop OU structure
In this example, only the central admin team has full admin rights on the “Citrix” OU and all
descendant objects, including the domain local user groups. All XenApp and/or XenDesktop
resources are published to these domain local groups, which contain nothing but the domain
global groups from each of the branch offices. The local admin teams can add/modify/delete
Active Directory objects, including the users and domain global groups, within their respective
branch OU only. This approach ensures that the central admin team is not required for adding

or removing users to/from resources and that local admins cannot administer objects outside of
their respective branch OU.
Decision: Service Accounts
Citrix Windows services must run under the context of a specific user account. In most cases, a
local user account is created during installation or an existing account (e.g. Network Service) is
used for this purpose. In some scenarios, an Active Directory user account is required. For
example, in environments in which the Provisioning Services vDisks are hosted on a file server
rather than the local hard disks, a user account with sufficient access permissions needs to be
used. While it is technically possible to leverage existing user accounts, Citrix strongly
recommends that dedicated service accounts are created for these scenarios. This enables
administrators to create Active Directory accounts with the following attributes:
 Password Never Expires: If the password expires all dependent services are unable
to start. In case the password is changed all dependent services have to be
reconfigured.
 Strong Password with High Complexity: This ensures the service accounts cannot
be hacked using simple dictionary attacks and brute force based attacks are very time
consuming. This offsets the security implication of a non-expiring password.
 Elevated User Rights: Service accounts often require access rights beyond normal
user permissions for certain parts of the infrastructure. In general service accounts
should follow the principle of least privilege, allowing that each account in a system be
granted the most restrictive set of privileges needed for the performance of authorized
tasks. This will limit the ability to of the account to impact other components in the
environment should the service account be exposed. Therefore it is recommended to
create one dedicated service account for all services of an infrastructure component.
For example one service account for PVS should be created, which is used to run the
“SOAP” and “Streaming” services.
 No User Policies: In order to ensure a resilient service and to simplify troubleshooting
it is strongly recommended that service accounts are created within an organizational
unit which does not have any Group Policy Objects applied.
Citrix infrastructure components, which require a service account:
 XenApp: Database access
 Provisioning Services: Database and vDisk access and server management tasks
(e.g. start / stop services)
 XenServer: ISO store access
 NetScaler: User authentication in Active Directory
Decision: Server Naming Scheme
A standard naming convention should be defined for all key infrastructure components,
including servers, databases and service accounts which identifies the role or function, location
and affiliation of the component. A standard naming convention will help to simplify operations
and troubleshooting. A sample naming scheme is provided:
Naming: LLLFFF####
Explanation: LLL = Location
FFF = Function

#### = Unique number
Samples: FLLXDC0001 (XenDesktop controller in Fort Lauderdale with the number 0001)
BERWEB1234 (Web server in Berlin with the number 1234
In order to ensure compatibility with old operating systems the computer names should not
exceed 15 characters. Computer names should be restricted to alphabetical characters (A-Z),
numeric characters (0-9), the minus sign (-), and the period (.) to ensure compliance with
RFC952.
Further guidance for server name schemes can be found within KB909264 - Naming
conventions in Active Directory for computers, domains, sites, and OUs.
Best Practices
Don‟t forget to follow these top recommendations from Citrix Consulting:
1. Time Synchronization: Due to the VDA registration process relying on Kerberos
authentication, the VDAs and Controllers should have their time synchronized. The
Network Time Protocol (NTP) should be used to synchronize the time across all
components.
2. Active Directory Sites: Ensure the subnets in which the XenDesktop / XenApp
components are located are created in Active Directory and assigned to the correct
site. Otherwise an arbitrary domain controller is selected for user and computer
authentication and traffic may be sent across the WAN instead of the nearest
domain controller.
3. Functional Level: To ensure all features of XenDesktop / XenApp are available a
domain functional level of Windows 2003 native or higher needs to be selected.
4. Global Catalog: To ensure fast user lookups and authentication the Active Directory
site in which the XenDesktop / XenApp components are located should contain at
least one domain controller configured as a Global Catalog.

Database
The majority of Citrix products discussed within this document make use of a database for various
reasons. The following table outlines the usage on a per product basis:
Product Configuration Data Runtime Data Audit/Change Log Data
XenApp  
XenDesktop  
Provisioning Services  
XenClient   
“”: Optional
Table 54: Database usage
A shared database helps to simplify administration by allowing configuration settings to be applied
across multiple objects of the same type. In addition, XenDesktop leverages a shared database to
store dynamic runtime information, for example connected users and available desktops. In this
scenario the database effectively becomes a message bus between controllers within the same
site. Some Citrix products can utilize a database to store a configuration or audit log, which
outlines the administrative changes performed within the infrastructure. Because of these
dependencies the database design is a vital part of every Citrix Architecture Design.
Decision: Vendor and Version
There are a variety of vendors that provide database software including Microsoft, Oracle,
Sybase and IBM. In order to streamline the development and testing of Citrix products, support
has been limited to a subset of vendors and versions: Citrix XenApp supports database
software from Microsoft, Oracle and IBM, while support for all other Citrix products has been
limited to Microsoft SQL Server. Therefore, it is recommended that all Citrix related databases
are hosted on Microsoft SQL Server.
Consequently this guide will focus on Microsoft SQL Server only.
For more information in regards to supported database vendors and versions, please refer to
CTX114501 - Supported Databases for Citrix Products.
Decision: Edition
Microsoft SQL Server is shipped in six versions: Express, Workgroup, Web, Standard,
Enterprise and Datacenter. The following table outlines the features of each version, which are
most applicable to Citrix databases:
Feature Datacenter Enterprise Standard Web Workgroup Express
Scalability and Performance
Number of
CPUs
OS
maximum
8 4 4 2 1
Maximum
memory
utilized
OS
maximum
2 TB 64 GB 64 GB 4 GB 1 GB
Maximum
database
size
524 PB 524 PB 524 PB 524 PB 524 PB 10 GB

High Availability
Log
Shipping
    
Database
Mirroring
  Yes
(Single
ReDo
Thread,
synchronou
s only)
1
Witness
Only
Witness
Only
Witness
Only
Number of
Failover
Cluster
Nodes
16 16 2 - - -
Virtualization Support
Hyper-V
Live
Migration
   - - -
Table 55: SQL 2008 R2 Edition Feature Comparison
1
In environments with a large number of mirrored databases a single ReDo thread can increase
the time required to successfully failover from principal to mirror server. For further information,
please review MSDN - Estimating the Interruption of Service During Role Switching.
Based on the capabilities of the various SQL Server versions available, Citrix recommends
using the Standard edition for hosting Citrix databases in production environments, if the SQL
server is dedicated to Citrix databases.
Decision: Database and Transaction Log Sizing
When sizing a SQL database, two aspects are important:
 Database File: Contains the data and objects such as tables, indexes, stored
procedures, and views stored in the database.
 Transaction Log File: Contains a record of all transactions and database modifications
made by each transaction. The transaction log is a critical component of the database
and, if there is a system failure, the transaction log might be required to bring the
database back to a consistent state. The usage of the transaction log varies depending
on which database recovery model is used. Following the three available models and
their impact of the transaction log are outlined:
o Simple Recovery: No log backups required. Log space is automatically
reclaimed, to keep space requirements small, essentially eliminating the need
to manage the transaction log space. Changes to the database since the most
recent backup are unprotected. In the event of a disaster, those changes must
be redone.
o Full Recovery: Requires log backups. No work is lost due to a lost or
damaged database data file. Data of any arbitrary point in time can be
recovered (for example, prior to application or user error). Full recovery is
required for database mirroring.
o Bulk-Logged: This model is an adjunct of the full recovery model that permits
high-performance bulk copy operations. It Is typically not used for Citrix
databases.

For further information, please refer to MSDN – SQL Server Recovery Models.
In order to estimate the storage requirements, it is important to understand the disk space
consumption for common database entries. This section will outline the storage requirements
on a per product basis and provide sizing calculations.
XenDesktop
Since the database of a XenDesktop site contains static configuration data and dynamic runtime
data, the size of the database file depends not only on the physical size of the environment but
also the user pattern. The following factors all impact the size of the database file:
 The number of configured and registered workstations
 The number of connected sessions
 The connection rate of users
 The number of MCS-based and PVS-based desktops
Each desktop (regardless of it‟s type) requires:
 ~2.9KB for registration and worker state information
 ~5.1KB for session state
 ~0.04 KB for each connection log record. By default, these records are purged after 48
hours.
Each Machine Creation Service (MCS) provisioned desktop also requires:
 ~1.8KB for AD computer account information
 ~1.94KB for MCS machine information
In addition, approximately 10MB is used for the database structure and general configuration
data. A further 10% overhead should be allowed for policies, catalogs, published applications
and other infrastructure configuration information. In order to illustrate the database growth, the
following table outlines three sample configurations:
# of desktops 5,000 10,000 20,000
For every virtual desktop
Per Worker (KB) 14,500 29,000 58,000
Per Session (KB) 25,500 51,000 102,000
Per Connection (KB) 840 1,680 3,360
Sub-Total (KB) 40,840 81,680 163,360
Additional space for MCS based virtual desktops
Per AD Account (KB) 9,000 18,000 36,000
Per MCS machine
(KB)
9,700 19,400 38,800
Sub-Total (KB) 59,540 119,080 238,160
General infrastructure requirements
Initial Database (KB) 10,000 10,000 10,000
Sub-Total (KB) 69,540 129,080 248,160

10% allowance for
additional data
~7,000 ~13,000 ~25,000
Total (KB) 76,540 142,080 273,016
Total (MB) 76 142 273
Table 56: XenDesktop site DB sample size calculations
Note: This sizing information is a guide only. Actual database usage might be slightly larger
because of SQL server having spare pages in tables.
The transaction log is more difficult to size because it depends on a number of factors,
including:
 The SQL Database recovery model in use
 Launch rate at peak times
 The number of desktops
Testing with the full recovery model shows that logging on 20,000 users to 20,000 virtual
desktops (at a launch rate of 40 users per second) consumed ~1.3MB per second of transaction
log space. This would consume ~670MB of transaction log in total for the logon period.
Idle desktops consume ~62KB of transaction log space every hour. This means that one
desktop will consume 1.45MB and 20,000 desktops will consume ~29GB in a 24-hour period.
In a large environment, this level of transaction log usage requires careful management.
By default, Desktop Studio configures any new SQL databases to use the simple recovery
model. However, Citrix recommends that if full recovery model is in use, that the transaction log
be sized so that it rarely has to grow. Testing showed that when SQL Server auto-grew the
transaction log, all transactions stalled causing an increase in controller response times.
Instead, Citrix recommends that a fixed-size transaction log is used and a SQL Alert is setup so
that when the transaction log is 50% full, the transaction log is backed up, thus freeing up
space. This maintains the transaction log at a reasonable size, without impacting SQL Server
performance.
Please refer to CTX127939 - XenDesktop 5 Database Sizing and Mirroring Best Practices for
further information.
XenApp – Data Store
In contrast to XenDesktop the XenApp data store only contains static configuration data. This
means that the size of the infrastructure, rather than the number of users (e.g. number of
XenApp servers, published applications or Load Evaluators), will impact the size of the
database. Initially the XenApp data store will be less then 5MB in size and will grow as
additional XenApp servers are added or configuration is performed. In order to illustrate the
database growth, the following table outlines a sample configuration in which only the number of
servers and the number of server folders is increased.

Configuration Item
Static configuration items
Number of applications published
to all servers
50
Number of user policies 25
Printers per server 5
Printer drivers installed per server 25
Network print servers with printers 5
Number of Load Manager load
evaluators
10
Number of application folders in
management console
10
Number of Application Isolation
Environments
10
Number of Citrix administrators 10
Variable configuration items
Number of server folders in
management console
8 16 25 50 50
Number of servers in farm 50 100 250 500 1000
Size of data store in MB 32 51 76 125 211
Table 57: XenApp Farm DB sample size calculations
Due to the nature of the data in the XenApp data store the transaction log is not expected to
grow quickly unless a large number of configuration changes take place. Nevertheless, over
time the transaction log will grow and it can reach multiple gigabytes in size. In order to
minimize the space required, backup the transaction log on a regular schedule, which will shrink
the size of the log file. Alternatively, the recovery model of the database can be set to “Simple”.
XenApp – Configuration Log
The Configuration Logging database will only store records of the administrative changes
performed in the XenApp farm. Therefore, the database as well as the respective transaction
log is expected to be very small. When an administrative change is made the following entries
will be made in the database:
 1 entry in CtxLog_AdminTask_LogEntry (row size = 3096)
 1 entry in CtxLog_AdminTask_Object (row size = 1068)
 3 entries in CtxLog_AdminTask_ReferenceList (row size = 108)
In order to calculate the size of the Configuration Logging database, the following formula can
be used, which assumes a page size / packet size of 8192 bytes and a fill factor of 80%:
or

Based on this formula, one change with the aforementioned assumptions, will add 5.6KB to the
Configuration Logging database which has an initial size of 1,5MB. Therefore, a Configuration
Logging database in a XenApp environment with 10,000 administrative changes will have a size
of approximately 57.5MB.
Due to the nature of the data in the XenApp configuration logging database the transaction log
is not expected to grow very quickly, unless a large amount of configuration changes take place.
Please refer to the previous section for further guidance.
The Provisioning Services farm database contains static configuration and configuration logging
(audit trail) data. Therefore, the database as well as the respective transaction log is expected
to be relatively small. In order to calculate the size of the database, the record size
requirements outlined below can be used (important configuration items are marked bold):
Configuration Item DB space
required
(KB)
Sample
Number of items Total (KB)
Base Farm configuration 112 - 112
User group w/ farm access 50 10 250
Site 4 5 20
Device Collection 10 50 500
Farm View 4 10 40
Farm View to Device relationship 5 1 5,000
Site View 4 5 20
Site View to Device relationship 5 1 5,000
Device 2 5,000 10,000
Device Bootstrap 10 - -
Device to Disk relationship 35 1 175,000
Device Printer relationship 1 - -
Device Personality data 1 - -
Device status (when booted) 1 5,000 5,000
Device custom property 2 - -
vDisk 1 20 20
vDisk version 3 5 300
Disk Locator 10 1 200
Disk Locator custom property 2 - -
Server 5 10 50
Server IP 2 1 20
Server status (when booted) 1 20 20
Server custom property 2 - -

vDisk Store 8 5 40
vDisk Store to server relationship 4 1 40
Connection to XenServer (VirtualHostingPool) 4 - -
vDisk update task 10 10 100
Administrative change (Auditing enabled) 1 10,000 10,000
Total Overall 211,732KB
Table 58: Provisioning Services Farm DB sample size calculations
During the PVS farm setup a database with an initial file size of 20MB is created. Therefore the
database is not expected to grow automatically initially. Due to the nature of the data in the
PVS farm database the transaction log is not expected to grow very quickly unless a large
amount of configuration is performed. Please refer to the XenApp data store section for further
guidance.
In contrast to XenApp, which also offers the ability to track administrative changes, the related
information is not written to a dedicated database but directly into the Provisioning Server farm
database. In order to limit the size of the Provisioning Services database it is recommended to
archive the Audit Train data on a regular schedule. Please refer to the Operations chapter for
further details.
XenClient
The XenClient database contains both static configuration and active computer data. The static
data includes the endpoint computers, policies, virtual machine images and their assignments.
The active computer data which includes the last check-in time and disk usage is only captured
when the endpoint computer communicates back to the Synchronizer at its set internal. This
data is only refreshed and does not progressively build up. The XenClient event log data does
however progressively build up as the environment grows. This data includes all events that
occur such as image creation, image updates, remote help desk logons, and scheduled tasks.
Initially, the XenClient database will be less then 10MB in size including the transaction logs
(full) although it will grow as additional computers, images, and policies are added.
Decision: High-Availability
The following table highlights the impact of a database outage:
Component Impact of a database outage
XenDesktop site Database Users will be unable to connect or reconnect to a virtual desktop.
Administrators are unable to use XenDesktop Studio or XenDesktop
Director. Any Users with existing connections will be unaffected
XenApp Farm Data Store Each XenApp server stores a subset of the data store in the Local Host
Cache (LHC). If the data store is unreachable, the LHC contains
sufficient information about the farm to allow normal operations for an
indefinite period of time. However, administrators will be unable to
manage the farm until the farm data store becomes operational again.
Neither users with existing sessions nor new users will be impacted.
XenApp Configuration
Logging Database
If “Allow changes to the farm when logging database is disconnected”
has been enabled in the XenApp farm properties, then an outage of the
configuration logging database has no impact (other than configuration
changes not being logged). Otherwise administrators cannot perform
any changes in the XenApp farm configuration. Users are not
impacted.

Farm Database
When offline database support is enabled and the database becomes
unavailable, the Stream Process uses a local copy of the database to
retrieve information about the Provisioning Server and the target
devices available to the server. This allows Provisioning Servers and
target devices to remain operational. However, when the database is
offline, the Console and the management functions listed below
become unavailable:
 AutoAdd target devices
 vDisk creation and updates
 Active Directory password changes
 Stream Process startup
 Image Update service
 PowerShell and MCLI based management
If offline database support has not been enabled, all management
functions become unavailable and boot as well as failover procedures
of target devices will fail.
XenClient Database Administrators will be unable to access the Synchronizer webpage to
manage the environment, new computers will be unable to register and
existing computers will be unable to receive image updates or policies.
Existing users will still be able to function normally with access to their
virtual machines being uninterrupted.
Table 59: Impact of a database outage
In addition to the built-in Citrix database redundancy options, Microsoft SQL Server as well as
the underlying hypervisor (in virtual environments) offers a number of high availability features.
These enable administrators to ensure single server outages will have a minimal impact (if any)
on the Citrix infrastructure. Following the SQL / hypervisor high availability features are
outlined:
 VM-Level HA: This high availability option is available for virtual SQL servers only,
which need to be marked for High Availability at the hypervisor layer. In case of an
unexpected shutdown of the virtual machine or the underlying hypervisor host, the
hypervisor will try to restart the VM immediately on a different host. While “VM-level
HA” can minimize downtimes in power-outage scenarios, it cannot protect from
operating system level corruption. This solution is less expensive than mirroring or
clustering because it uses a built-in hypervisor feature. However, the automatic failover
process is slower, as it can take time detect an outage and start the virtual SQL server
on another host. This may interrupt the service to users.
 Mirroring: Database mirroring increases database availability with almost
instantaneous failover. Database mirroring can be used to maintain a single standby or
mirror database, for a corresponding principal or production database. Database
mirroring runs with either synchronous operation in high-safety mode, or asynchronous
operation in high- performance mode. In high-safety mode with automatic failover
(recommended for XenDesktop) a third server instance, known as a witness, is
required, which enables the mirror server to act as a hot standby server. Failover from
the principal database to the mirror database happens automatically and is typically
completed within a few seconds. It is a good practice to enable VM- level HA (or a
similar automatic restart functionality) for at least the witness to ensure SQL service
availability in case of a multi-server outage.

 Clustering: Failover clustering provides high-availability support for an entire instance
of SQL Server. A failover cluster is a combination of two or more nodes, or servers,
with two or more shared disks. A SQL Server failover cluster instance appears on the
network as a single computer, but has functionality that provides failover from one node
to another if the current node becomes unavailable. The transition from one note to the
other node is seamless for the clients connected to the cluster.
The following table outlines the recommended high availability features for Citrix databases:
Component VM-Level HA Mirroring Clustering
XenDesktop site Database   
XenApp Farm Datastore  
1

XenApp Configuration Logging Database   
Provisioning Services Farm Database   
XenClient Database   
“”: Recommended. “”: Viable. ““: Not Supported.
“”: Recommended for test environments only.
1
Please refer to CTX111311 - Using SQL Database Mirroring to Improve Citrix XenApp Server Farm
Disaster Recovery Capabilities for further information
Table 60: Recommended SQL high availability options
Please note that in case a component supports database mirroring and clustering, database
mirroring is preferred due to its less complex setup.
Decision: SQL Server Sizing
The SQL server must be sized correctly to ensure the performance and stability of an
environment. Since every Citrix product uses SQL server in a different way, no generic all-
encompassing sizing recommendations can be provided. Instead, pre-product SQL server
sizing recommendations are provided below.
XenDesktop
XenDesktop Brokers use the database for storing configuration data as well as a message bus
for broker communications. The database is constantly in use and the performance impact on
the SQL server can be considered as high.
Based on results from Citrix internal scalability testing the following SQL server specification is
recommended:
 2 Cores / 4 GB RAM for environments up to 15,000 users
 4 Cores / 8 GB RAM for environments with 15,000+ users
The database files and transaction logs should be hosted on separate hard disk subsystems in
order to cope with the high number of transactions. For example, registering 20,000 virtual
desktops during a 15 minute boot storm causes ~500 transactions / second and 20,000 users
logging on during a 30 minute logon storm causes ~800 transactions / second.
XenApp
XenApp uses the database for storing static configuration data. In addition, every XenApp
server polls the database on a regular schedule (30 minutes) in order to ensure that all
configuration changes have been replicated to the Local Host Cache (LHC), which is used for
configuration data lookups during normal operations. Performance intense database operations

are joining servers to the farm, applying hotfixes and recreating the LHC. Since these
operations occur rarely the performance impact of this database can be considered as low.
Based on results from Citrix internal scalability testing the following SQL server specification is
recommended:
 2 Cores / 2 GB RAM for environments up to 250 servers
 2 Cores / 4 GB RAM for environments up to 500 servers
 4 Cores / 8 GB RAM for environments up to 1,000 servers
In addition to static configuration data Provisoning Servers store runtime and auditing
information in the database. Depending on the boot and management pattern, the performance
impact of the database can be considered as low to medium.
Based on this categorization, a SQL server specification of 4 Cores and 4 GB RAM is
recommended as starting point. During testing and pilot phase a thorough monitoring of the
SQL server performance should be performed, in order to determine the optimal configuration of
the SQL server.
XenClient
Similar to Provisioning Servers, XenClient Synchronizers store static configuration data in
addition to runtime and auditing / event information in the database. Depending on the client
synchronization and management pattern the performance impact of the database can be
considered as low to medium.
Based on this categorization, a SQL server specification of 4 Cores and 4 GB RAM is
recommended as starting point. During testing and pilot phase a thorough monitoring of the
SQL server performance should be performed, in order to determine the optimal configuration of
the SQL server.
Decision: Authentication and Authorization
Authentication and authorization is required in order to access a database hosted on a Microsoft
SQL Server. There are two authentication options available: Windows Authentication mode and
mixed mode.
 Windows Authentication: Enables a user or service to leverage Active Directory
credentials to access a database. When a user connects through a Windows user
account, SQL Server validates the account name and password using the Windows
principal token in the operating system. This means that the user identity is confirmed
by Windows. SQL Server does not ask for the password, and does not perform the
identity validation. Windows Authentication is the default authentication mode, and is
much more secure than SQL Server Authentication. Windows Authentication uses the
Kerberos security protocol, provides password policy enforcement with regard to
complexity validation for strong passwords, provides support for account lockout, and
supports password expiration. A connection made using Windows Authentication is
sometimes called a trusted connection, because SQL Server trusts the credentials
provided by Windows.
 Mixed mode: Enables Windows Authentication and SQL Server Authentication. When
using SQL Server Authentication, logins are created in SQL Server that are not based
on Windows user accounts. Both the user name and the password are created using
SQL Server and stored in SQL Server. Users connecting via SQL Server
Authentication must provide their credentials (login and password) every time that they
connect.

For security reasons, Citrix recommends using Windows Authentication whenever possible.
Service accounts should be created for every Citrix product, as outlined within the Active
Directory chapter.
For reference purposes, following the minimum SQL server and database rights required for
initial setup, maintenance and runtime are provided below:
XenDesktop
 Setup and Maintenance: For initial database creation, adding controllers, removing
controllers or applying database schema updates the administrator who performs these
tasks requires dbcreator and securityadmin server rights as well as the db_owner
database right.
 Runtime: During initial database setup XenDesktop configures specific database
security roles, which limit the XenDesktop services to read, write and execute and
should not be modified. Please refer to CTX127998 - Database Access and Permission
Model for XenDesktop 5 for further information.
XenApp
 Setup and Maintenance: For initial database creation the administrator who is
performing this task requires dbcreator and securityadmin server rights. The service
account which is used by XenApp to connect to the database, requires db_owner rights
for farm setup, upgrading or applying hotfixes.
 Runtime: During normal runtime the XenApp service account requires read
(db_datareader), write (db_datawriter) and execute permissions only. Further
information can be found in eDocs – Datastore Database Reference.
 Setup and Maintenance: For initial database creation the administrator who is
performing this task requires dbcreator and securityadmin server rights.
 Runtime: During normal runtime the Provisioning Services service account requires
read (db_datareader), write (db_datawriter) and execute permissions only. Further
information can be found in eDocs – Installing and Configuring Provisioning Services.
XenClient
 Setup, maintenance and runtime: The service account used for connecting to the
XenClient database requires db_owner rights for all phases. Further information can be
found in CTX136006 - Citrix XenClient Enterprise 4.5 Synchronizer Installation Guide.

Hardware Layer
The hardware layer is responsible for the physical devices required to support the entire solution. It
includes servers, processors, memory, storage devices. This layer is broken into three groups focused on
providing the necessary resources for specific parts of the entire solution. One group of servers will
support the XenApp (shared) components (if applicable). A second group of servers will support the
XenDesktop (VDI) components (if applicable). A final group of servers will support the underlying
infrastructure for the entire environment, which is identified as the Control Layer.
Hypervisor
Hyper-V
This chapter provides design guidance for Citrix XenDesktop 5.x deployments that leverage
Microsoft Hyper-V 2008 R2. It will give the reader a high-level understanding of the key
implementation decisions when building a Hyper-V 2008 R2 environment.
Decision: Host Hardware
The hardware selected for the Hyper-V hosts will have a direct impact on the performance,
scalability, stability and resilience of the XenDesktop solution. As such, the following key design
decisions must be considered during the hardware selection process:
 Hardware Compatibility: For reasons of stability, it is imperative that the hardware
selected is compatible with Hyper-V. The Microsoft Windows Server Catalog website
maintains a list of servers and hardware components capable of supporting Hyper-V.
Servers and hardware components with the “Certified for Windows Server 2008 R2”
and “Certified for Windows Server 2008” logo pass the Microsoft standards for
supporting Windows Server 2008 R2.
 Processor Specification: The Hyper-V design must ensure that the processor
specification selected is capable of supporting Hyper-V:
o 64-bit Intel VT or AMD-V processors with a clock speed of at least 1.4GHz,
however 2GHz or faster is recommended.
o Hardware Data Execution Protection (DEP), specifically Intel XD bit (execution
disable bit) or AMD NX bit (no execute bit), must be enabled and available. For
more information, please refer to Microsoft TechNet Article 912923 - How to
determine that hardware DEP is available and configured on your computer.
o For improved performance and scalability, ensure that Second Level Address
Translation (SLAT) capable processors are selected. Intel calls this technology
Extended Page Tables (EPT) and AMD calls it Nested Page Tables (NPT).
The technology provides hardware assistance for memory virtualization,
memory allocation, and address translation between the physical and virtual
memory. The reduction in processor and memory overhead improves
scalability allowing a single Hyper-V host to run more virtual machines.
Note: Live Migration allows administrators to move virtual machines between physical
hosts. In order to use this feature, all physical host servers in a cluster must have
processors from the same family. Live Migration will not work from a host with an Intel
based processor to a host with an AMD based processor. There is a circumstance in
which Live Migration may fail even when processors are from the same family. For
example, an application may be aware of CPU features it discovered on the host server
that are no longer available after it has been migrated to a new host. In these
situations, Hyper-V Processor Compatibility Mode should be used to hide the processor
features from the application. For more information please refer to Microsoft TechNet
article When to use Processor Compatibility Mode to Migrate Virtual Machines.

 Memory Capacity: The hardware selected for the Hyper-V hosts must have sufficient
memory to support the parent partition as well as the guest virtual machines that they
support:
o At a minimum, the parent partition requires 512MB of RAM, however 2GB or
greater is recommended for VDI environments.
o Virtual desktops typically require between 768MB and 4GB of RAM depending
on workload and operating system used. For more information, please refer to
the Resource Requirements chapter.
o Each Citrix XenApp user typically requires between 200MB and 1GB of RAM
depending on their workload. For more information, please refer to the
Resource Requirements chapter.
Note: With the exception of Microsoft Windows Server 2008 R2 Standard edition, which
is limited to 32GB, Hyper-V can support up to a total of 1TB of physical memory.
Decision: Host Scalability
During the design, it‟s necessary to estimate the number of physical host servers that will be
required to support the XenDesktop implementation. Ideally, scalability testing should be
performed prior to the hardware being ordered, however this is not always possible. In these
situations, consider the following guidelines when determining single host scalability:
 The parent partition typically utilizes between 8 and 12% of the processor resources
available leaving between 88 and 92% available for the child partitions.
 At a minimum, the parent partition requires 512MB of RAM, however 2GB or greater is
recommended for VDI environments.
 Microsoft supports up to 12 virtual processors per logical processor for Hyper-V 2008
R2 hosts supporting Windows 7. A maximum of 8 virtual processors per logical
processor is supported for all other operating systems.
Physical
Processors
Cores per
Processor
Threads per
Core (Hyper-
Threading)
Max Virtual
Processors –
Win7
Max Virtual
Processors – All
Other OS
2 2 2 96 64
2 4 2 192 128
2 6 2 288 192
2 8 2 384 256
2 10 2 480 320
4 2 2 192 128
4 4 2 384 256
4 6 2 512 (Maximum) 384
4 8 2 512 (Maximum) 512 (Maximum)
4 10 2 512 (Maximum) 512 (Maximum)
Table 61: Maximum virtual processors for Windows
It is important to realize that these are maximum values. Testing from Cisco shows that a Cisco
UCS B230 M2 Blade Server (2 processor / 10 core / 2 threads) supports 145 Windows 7
desktops running the Login VSI medium workload (rather than the 480 maximum supported).

Processor was identified as the primary bottleneck. For more information, please refer to the
Cisco Whitepaper – Citrix XenDesktop on FlexPod with Microsoft Private Cloud.
Decision: Scale Up or Out
There are a number of environmental factors that need to be considered when determining
whether the Hyper-V host hardware specification should be “scaled up” (reduced number of
high-performance hosts) or “scaled out” (increased number of less powerful hosts), including:
 Data Center Capacity: The data center may have limited space, power and/or cooling
available. In this situation, consider scaling up.
 Infrastructure and Maintenance Costs: When determining the overall cost of the
Hyper-V hosts, careful consideration should be taken when planning to include costs
such as rack space, support contracts and the network infrastructure required.
 Hosting Costs: There may be hosting and/or maintenance costs based on the number
of physical servers used. If so, consider “scaling up” to reduce the long-term costs of
these overheads.
 Redundancy: Spreading user load across additional less-powerful servers helps
reduce the number of users affected from hardware or software failure on a single host.
If the business is unable to accept the loss of a single high-specification server,
consider “scaling out”.
Decision: Hyper-V Edition
The design must determine which variation of Hyper-V will be used:
 Hyper-V Server 2008 R2: Standalone version of Hyper-V that does not include a
Graphical User Interface (GUI). Administration is either performed remotely, or via the
command line. Hyper-V Server does not require a Windows 2008 R2 license, however
licenses are required for each guest virtual machine. For more information, please refer
to the Microsoft Hyper-V Server 2008 R2 website.
 Windows Server 2008 R2: Hyper-V is also available as an installable role for Windows
Server 2008 R2 Standard, Enterprise and Datacenter editions, each of which includes
different levels of “virtual server image use rights” which range from one Microsoft
server operating system license to unlimited. For more information, please refer to the
Microsoft Windows Server 2008 R2 website.
The following table highlights the key differences between Hyper-V Server 2008 R2 and the
different editions of Windows Server 2008 R2 when it comes to supporting XenDesktop:
Capability Hyper-V Server
2008 R2
Windows Server 2008 R2
Standard Enterprise Datacenter
Virtual Server Image Use Rights 0 1 4 Unlimited
Sockets 8 4 8 64
Memory 1TB 32GB 1TB 1TB
Cluster Shared Volumes Yes No Yes Yes
Table 62: Hyper-V Server 2008 R2 / Windows 2008 R2 Comparison
Hyper-V Server 2008 R2 is typically preferred for XenDesktop 5.x deployments because it has a
decreased attack service, improved stability, reduced maintenance and does not require a
Windows Server 2008 R2 license. Windows Server 2008 R2 Enterprise and Datacenter is
recommended when familiarity with the command-line is limited or there is third-party software
to be installed on the server that is not supported on the Hyper-V Server 2008 R2 version.

The virtual server image rights included with Windows Server 2008 R2 Datacenter make it an
excellent choice for hosting XenApp servers.
Note: The Standard edition of Windows Server 2008 R2 is rarely used due to the limited
memory support. In addition, the Standard edition lacks Cluster Shared Volumes support which
is a recommended feature for hosting infrastructure servers and dedicated virtual desktops.
Decision: Network Connectivity
If unfamiliar with the concepts in this section, please refer to the Microsoft Whitepaper -
Understanding Networking with Hyper-V.
When integrating Hyper-V with XenDesktop it is important to consider the following key
networking topics:
 Physical Network Connections: If sufficient infrastructure exists, performance may be
improved by separating different types of network traffic across multiple physical
Network Interface Cards (NICs), for example management, cluster, virtual machine,
storage, provisioning and backup traffic can all be isolated from each other. A Hyper-V
host used to support a XenDesktop environment typically utilizes between four and
eight physical network connections:
o 2 x NICs (teamed): Live Migration Network, Cluster Shared Volume Network,
Virtual Machine Network
o 1 x NIC: Management Network
o 1 x NIC: Cluster Private Network
If standard network adapters, rather than HBAs, are used to access shared storage, it
may be beneficial to separate out the storage traffic onto a dedicated network:
o 2 x NICs (teamed): Storage Network
In addition, if the servers are network-constrained and Provisioning Services is utilized it
may be beneficial to separate out the provisioning traffic:
o 2 x NICs (teamed): Provisioning Network
For more information, please refer to CTX120955 - How to Setup a Multi-Homed VM
Separating Large Scale Traffic from ICA Traffic for XenDesktop.
Note: All hosts within a cluster must have the same number of network interfaces and
the interfaces must be connected to the same networks.
 NIC Teaming: The implementation of NIC teaming provides increased resiliency by
allowing up to two network cards to function as a single entity. In the event of one NIC
failing, traffic is automatically routed to the second NIC. The Hyper-V design should
determine which network connections should be teamed for resiliency. Citrix
recommends teaming all network connections apart from the Management and Cluster
Networks because they can be failed over to alternative networks in the event of a
failure.
Many servers today are supplied with NICs offering two or more ports. As such, it is
important that any teams created consist of connections from two separate physical
NICs so that a single card failure does not bring down the team.
Redundancy should also encompass the external network. Teamed NICs should be
diversely connected to external switches to help reduce the risk from a single switch
failure.

Note: Microsoft does not support NIC teaming with Hyper-V 2008 R2; however, third
party solutions are available. All support queries regarding teaming with Hyper-V 2008
R2 should be directed to the NIC OEM. In addition, some OEMs may not support TCP
Chimney Offload and VMQ functionality with NIC teaming enabled. Please refer to the
OEM‟s teaming documentation for more information. For more information, please refer
to Microsoft Knowledgebase Article KB968703 - Microsoft Support Policy for NIC
Teaming with Hyper-V.
 IP Addressing: IP addresses need to be assigned to the Hyper-V network interfaces
and individual virtual machines. As such, the design must consider the IP addressing
requirements for these components. If DHCP is used to provide the IP configuration for
the Hyper-V hosts, ensure that DHCP reservations are created for the appropriate MAC
addresses to prevent DNS resolution issues when IP addresses change.
The Hyper-V network design should ensure that the Hyper-V traffic is routed via the
appropriate virtual and physical networks. For example, shared storage traffic is routed
via the parent partition and not directly from the virtual machine. Depending on the
networking architecture used, static routes may need to be added to the parent
partition‟s routing table to ensure that the correct interface is used.
 VLANs: Many network environments utilize VLAN technologies to reduce broadcast
traffic, enhance security and to enable complex virtual network configurations that
would otherwise not be possible. The Hyper-V network design should determine which
VLANs will be required. Typical VLANs used in a XenDesktop environment include:
o Storage VLAN for storage traffic
o Desktop VLAN for virtual desktops
o Server management VLAN
o Provisioning VLAN for streaming services
Separate physical networks are typically created for the Hyper-V Live Migration,
storage, cluster and management networks.
Note: Hyper-V supports the configuration and use of 802.1Q tagged VLANs. Each
Hyper-V host will need to have its physical NICs connected to specific VLAN trunk ports
to allow for the correct routing of VLAN tagged traffic. For more information, please
refer to the TechNet Article – Hyper-V: Configure VLANs and VLAN Tagging.
Decision: Virtual Network Adapter
Hyper-V supports two types of network adapters:
 Synthetic Adapter: Virtual network adapter that uses a network driver installed via
Hyper-V Integration Services to communicate with the hardware abstraction layer.
Since the driver has to be loaded, the synthetic network adapter isn‟t available until the
operating system boots.
 Legacy Adapter: Emulated network adapter based on the Intel 21140 PCI Fast
Ethernet adapter. The legacy network adapter supports guest VMs that are not
compatible with Hyper-V Integration Services and VMs that are configured to network
boot.

The synthetic adapter is better performing, requires reduced host CPU, and is installed by
default when a VM is created. In XenDesktop and XenApp environments using Provisioning
Services the legacy network adapter must be used in order to PXE boot the desktops and
provisioned XenApp servers. As such, two options must be considered during the Hyper-V
design:
1. Only use the legacy network adapter, which does not perform as well as the synthetic
adapter.
2. Configure the Provisioning Services targets with dual network adapters: one synthetic
and one legacy adapter. Once the virtual machine PXE boots, the synthetic adapter
has precedence over the legacy adapter, but PVS traffic is bound to the legacy adapter
and is not able to take advantage of the better performing synthetic adapter.
Note: If the dual NIC configuration is used, each virtual machine will require two IP addresses
from the DHCP server.
The legacy network connection will be suitable for all but the most network intensive
environments. Therefore, for simplicity, utilize the legacy network adapter unless an associated
network bottleneck is identified.
Decision: Storage Connectivity
For an introduction to storage, please refer to Citrix Knowledgebase Article CTX118397 –
Introduction to Storage Technologies. For detailed XenDesktop storage recommendations,
please refer to Citrix Knowledgebase Article CTX130632 – Storage Best Practices.
Storage has a major impact on the performance, scalability and availability of the XenDesktop
implementation. As such, the storage design should focus on the following key areas:
 Direct-attached Storage/Storage-attached network: The Hyper-V 2008 R2 storage
design must determine whether the virtual machines should be hosted on a Storage-
attached Network (SAN), Direct-attached Storage (DAS) or a combination of the two.
There are a number of factors that need to be considered, including:
o Unlike DAS, SAN supports Live Migration, and Failover Clustering. Although
these features are less critical when supporting pooled virtual desktops and
XenApp servers, they are still very important for dedicated desktops and
infrastructure servers. A SAN must be used if these features are required.
o Another disadvantage of local storage is a reduced level of scalability due to
the hardware limit on the number of disks supported, particularly for blade
systems. As the number of virtual desktops per host increases, additional disks
are required to accommodate the number of IOPS generated.
o Although using local storage may require additional disks and array controllers
to be purchased per Hyper-V host, the overall cost is likely to be significantly
less than that of an enterprise storage solution.
o With a SAN there is a potential for a single point of failure. If the SAN
experiences hardware failure or other type of interruption, all services running
off the SAN may be impacted. If the SAN performance is degraded by a
faulting service, all SAN dependent services may suffer as well.
In many cases, the optimal solution is to use SAN storage for the infrastructure servers
and dedicated desktops so that they can benefit from Failover Clustering and Live
Migration while using less expensive DAS storage for pooled desktops and XenApp
servers.

Note: Network-attached storage (NAS) is not supported with Hyper-V 2008 R2. Please
refer to the Microsoft TechNet article KB2698995 regarding Microsoft support policy for
Hyper-V environments utilizing network-attached storage.
Decision: Dynamic Memory
Citrix has conducted many scalability tests with Dynamic Memory enabled and found it to be a
very useful and recommended tool. However, there are a few areas to consider during
planning:
 Dynamic Memory does not oversubscribe the memory pool. When all of the physical
memory has been assigned, desktops will not be able to request additional memory.
 The CPU costs to manage Dynamic Memory are real, but very reasonable (<1%
additional CPU overhead).
 An increase in IOPS is normal and should be planned for. When a virtual machine
requests additional memory, it will also start to utilize its page file in preparation for its
request being denied by Hyper-V and the operating system running out of physical
memory. Therefore, it should be expected that a virtual machine will increase its I/O
characteristics during this time, which can result in an increase in IOPS of up to 10%.
This is a short term increase, and one noticed more in mass scalability testing than in
real life situations. However, on systems where RAM is more plentiful than IOPS,
Dynamic Memory settings should be conservative. For example if a Windows 7 virtual
machine, with a subset of applications, has been tested and determined to require a
minimum of 700MB to run then the Dynamic Memory default setting of 512MB should
be increased to 700MB. By accepting the default setting there will be an increase in
IOPS consumption in relation to paging occurring for the guest.
For more information, including recommended default values for Startup RAM by operating
system, please refer to the Microsoft TechNet Article – Hyper-V Dynamic Memory Configuration
Guide.
Decision: SCVMM Architecture
When designing a VMM 2012 environment there are two design choices that should be
considered.
 Design 1: A single VMM virtual machine exists within a host cluster, or a series of VMM
servers each supporting no more than 2,000 Virtual Desktops each. These VMM virtual
machines have no VMM application level HA, however they do have VM based high
availability enabled with the Hyper-V cluster. The VM operating system and the VMM
application components can easily migrate between hosts within the cluster in the event
of a hardware failure. This design is recommended for deployments that can afford
minor disruption to service as some downtime is required for failover to occur. This
design offers a simpler setup and is the most often used.
 Design 2: Two VMM 2012 servers are used in a guest failover cluster
configuration. This design offers VMM application layer HA, however it can also
increase complexity during the initial setup of the environment. This design is more
restrictive than host clustering, and also requires iSCSI shared storage, since it is the
only storage type supported by guest clusters. This design is recommended for critical
deployments that can‟t afford any downtime due to a disruption of service.
Both of these designs are verified in the Microsoft Fast Track Reference Architecture Guide,
which also provides additional VMM 2012 architectural guidance.

Although Hyper-V virtual machines can function without VMM, XenDesktop relies on its
availability to manage virtual desktops. Therefore, it is imperative that the following VMM
components are protected from failure:
 VMM server: If the VMM server is unavailable, XenDesktop will not be able to manage
the power state of the virtual desktops that it manages or create additional virtual
desktops. Therefore, Microsoft Failover Clustering should be included in the design to
ensure that the VMM server is highly available. For more information, please refer to
the Microsoft Blog Post – Creating a Highly Available VMM server.
 VMM Library: The VMM Library is typically required for the creation of new desktops,
XenApp servers and infrastructure servers. Therefore, the loss of the VMM Library is
unlikely to impact the day-to-day activities of a standard XenDesktop deployment.
However, if required VMM supports the use of highly available library shares on a
Failover Cluster.
 SQL Server: The VMM database contains all VMM configuration information. If the
VMM database is unavailable, virtual machines will still be available but the VMM
Console will be unavailable. Therefore, Hyper-V Failover Clustering should be used to
ensure that the SQL Server is highly available.
Decision: SCVMM Sizing
The following recommendations should be taken into consideration when determining the
number of VMM servers required:
 A VMM 2012 server can manage up to 8,000 virtual machines. When a VMM server is
used to manage both virtual desktops and servers, Citrix has found that best
performance is achieved when each VMM server is limited to managing 2,000 virtual
desktops. There is no limitation placed on the number XenApp servers, or other control
VMs also managed by each VMM server.
 While it is possible to run other applications on the VMM server, it is not recommended,
especially other System Center 2012 applications because they tend to have heavy
resource demands and could significantly impact VMM performance.
 The following table provides recommended specifications for the key VMM
infrastructure components
Requirement Recommended Recommended (>50 Hosts)
VMM Management Server
Processor Dual processor, dual core,
2.8GHz (x64) or greater
Dual processor, dual core,
Memory 4GB 8GB
Disk Space
(no local DB)
40GB 50GB
Disk Space
(local DB)
150GB Use a dedicated SQL Server
Dedicated SQL Server
2GHz (x64) or greater
Memory 4GB 8GB
Disk Space 150GB 200GB

VMM Library Server
Memory 2GB 2GB
Disk Space
(no local DB)
Size will depend on what will
be stored
Size will depend on what will
be stored
VMM Admin Console
Processor Single processor, 1GHz or
greater
Single processor, 2GHz or
greater
Memory 2GB 4GB
Disk Space 2GB 4GB
Table 63: SCVMM recommendations
For more information on VMM 2012 system requirements see the following Microsoft
TechNet articles:
o System Requirements: VMM Management Server
o System Requirements: VMM Console
o System Requirements: VMM Database
o System Requirements: VMM Library
Decision: Failover Clustering
A Hyper-V failover cluster is a group of Hyper-V hosts (cluster nodes) that work together to
increase the availability of hosted virtual machines. Virtual machines on a failed host are
automatically restarted on another host within the same cluster, provided that sufficient
resources are available.
The following should be considered when designing the Hyper-V Failover Cluster:
 Virtual Machines to Cluster: The following virtual machines should be hosted on a
failover cluster because they are critical to the successful operation of XenDesktop on
Hyper-V:
o VMM servers
o Web Interface servers
o License servers
o SQL servers
o Desktop Controllers
o Provisioning Services
o XML Brokers
o XenApp Data Collectors
o Dedicated Desktops
o Pooled Desktops (with personal vDisk)
Pooled desktops and XenApp servers do not typically need to be hosted on a cluster
because they shouldn‟t have user-specific data or applications.

Note: Cluster Shared Volume services depend on Active Directory for authentication
and will not start unless they can successfully contact a Domain Controller. Therefore,
always ensure that there is at least one Domain Controller that is not hosted on a
Cluster Shared Volume. Please refer to the Microsoft Support article Things to
Consider When You Host Active Directory Domain Controllers In Virtual Hosting
Environments.
 Number of Failover Clusters: The number of Hyper-V clusters required to support an
implementation of XenDesktop varies depending on the following key factors:
o Infrastructure Nodes: When designing Hyper-V clusters, consider separating
the host server components such as AD, SQL and XenDesktop Controllers
from the dedicated virtual desktops by placing them on different Hyper-V
clusters. This will ensure that dedicated virtual desktop resources are isolated
from infrastructure resources for optimal performance and availability.
o Performance: Many businesses will have dedicated desktops that require
guaranteed levels of performance. As such, it is sometimes necessary to
create dedicated clusters to meet the service level agreements associated with
these desktops.
o Application Set: It may be beneficial to dedicate clusters for large dedicated
desktop groups as they share a common, predictable resource footprint and
application behavior. Alternatively, grouping dedicated desktop groups
together based on differing resource footprints could help to improve desktop
density per host. For example, splitting processor intensive dedicated desktops
across several clusters will help to distribute the impact from processor
saturation. If XenApp is used to provide virtualized application access, it is
advisable to separate the XenApp servers onto a separate Hyper-V cluster to
balance resource utilization.
o Physical Network: Some environments may have complex network
configurations that require multiple clusters to be deployed. For example, some
dedicated desktop groups may need to be isolated onto specific subnets for
reasons of security, whilst others may have requirements for network
connectivity that can only be satisfied within specific data centers or network
locations.
o Virtual Network: Depending on the environment, it may be overly complex to
trunk every VLAN to every Hyper-V cluster. As such, it may be necessary to
define clusters based on the VLANs to which they are connected.
o Processor Architecture: Hyper-V requires that nodes within the same cluster
have processors from the same family. Therefore, the Hyper-V design will
need to consider the processor architecture available when determining the
number of clusters required.
 Nodes per Failover Cluster: A Failover Cluster supports up to 16 nodes, however
consider reserving at least one node in the cluster for failover and maintenance.
A single cluster can support up to a maximum of 1,000 virtual machines. It is unlikely
that a cluster hosting dedicated virtual desktops would consist of 16 nodes as this
would limit each node to 63 dedicated virtual machines (66 if one host is dedicated to
HA). It is far more likely that a Hyper-V cluster supporting XenDesktop dedicated virtual
desktops will consist of between 7 and 13 nodes. The following formula can be used to
calculate the number of VMs per node:
o 7 nodes per cluster - 167 VMs per node with one dedicated to HA

o 14 nodes per cluster - 76 VMs per node with one dedicated to HA
For more information, please see the Microsoft TechNet article Requirements and
Limits for Virtual Machines and Hyper-V in Windows Server 2008 R2.
Note: When multiple virtual machines exist for each server role, ensure that they are
not all hosted on the same physical host. This will help to ensure that the failure of a
single virtualization host does not result in a service outage. In addition, the physical
hosts supporting the core infrastructure components should ideally be located in
different chassis/racks.
 Cluster Shared Volumes: A Cluster Shared Volume (CSV) allows virtual machines
that are distributed across multiple cluster nodes to access their Virtual Hard Disk
(VHD) files at the same time. The clustered virtual machines can all fail over
independently of one another. CSVs are required for Failover Clustering and Live
Migration functionality. Therefore infrastructure servers and dedicated desktops are
typically hosted on CSVs.
The following recommendations should be considered during the Cluster Shared
Volume design:
o Microsoft recommends that the CSV communications take place over a
different network to the virtual machine and management traffic.
o The network between cluster nodes needs to be low latency to avoid any lag in
disk operations but doesn‟t need to be high bandwidth due to the minimal size
of metadata traffic generated under normal circumstances.
Note: Since Clustered Shared Volume communication occurs over the Server Message
Block (SMB) protocol, the Client for Microsoft Networks and File and Printer Sharing for
Microsoft Networks services must be enabled on the network adapters used for the
cluster network. Disabling NetBIOS over TCP/IP is recommended.
For more information, please see the Microsoft TechNet article Requirements for Using
Cluster Shared Volumes in a Failover Cluster in Windows Server 2008 R2.

Overview
Deploying a desktop virtualization solution follows a similar process to the design phase, except the entire
process moves in reverse. Hardware must be in place before the appropriate controllers and desktops
can be deployed. Just like design, without the standardized and proven process, administrators tend to
randomly jump from component to component. The right approach focuses on building components in
order, and validating along the way.
The recommended approach focuses on working through five distinct layers, in the correct order:
Figure 28: Five-Layer Deploy Model
The supporting foundation must be deployed first. The Hardware and respective controllers must be
deployed before desktops or access configurations are built. The specifics for each of the layers, what
must be built, how they are configured and in what order are determined within the Assess and Design
phases.
User Layer
Hardware Layer
Desktop Layer
Control Layer
Access Layer
User Layer
Desktop Layer
Access Layer
User Layer
Desktop Layer
Access Layer
Deploy

Overview
With any virtual desktop solution, support and maintenance aspects for new environments often get
overlooked. Effectively maintaining a virtual desktop environment requires reliable systems are in place to
ensure smooth day to day operations. The Monitor phase of the methodology covers the responsibilities
involved in keeping a virtual desktop infrastructure in peak operating efficiency. These activities include:
 Support: When problems arise, technical support often is the first point of contact for issue
resolution. This section addresses the proper staffing, organization, training, and tools utilized in
effective support organizations.
 Operations: Maintenance, issue prevention and resolution are core responsibilities in running a
Citrix infrastructure. When the responsibilities and assignments are structured properly, friction is
kept to a minimum, reducing issues and their resolution times. This section discusses routine
operations that Citrix environments require for optimal performance.
 Testing: Regular upgrades are required to ensure a farm environment is up to date. Change
management processes are critical to ensure improvements are properly approved, tested, and
validated by appropriate parties. This section covers the proper processes that ensure changes in
production environments are deliberate, proven, and accountable.
 Monitoring: By having an in-depth understanding of current and expected behavior of the
various components, administrators are better equipped to discover an issue before it impacts the
user community. Furthermore the data tracked during normal operations can be used for trending
and capacity planning.
Support
Well trained support members with a clear organizational structure are critical for an effective support
team. Since technical support is the first place many users go, thought needs to be given to its structure,
training, and tools utilized by support members.
Structure
Multi-tier support levels have been found to be the most effective means in addressing user support
issues. Low criticality, low complexity or frequently occurring issues should be managed and resolved
at the lower tiers. High criticality and complex issues are escalated to more experienced architects or
infrastructure owners. The diagram below outlines a common multi-tier support structure:

Figure 29: Support Structure
The first level should resolve 75% of all issues encountered. These issues include routine problems
requiring limited knowledge of the Citrix environment. Issues are quickly resolved and may also be
automated, like password resets.
When non-routine problems exceeding level one‟s abilities are encountered, they are then escalated
to level two. Information on the end user‟s problem and attempted troubleshooting steps are
documented at the first level allowing level two technicians to immediately begin addressing the
problem. Level two technicians should handle only 20% of the support tickets and are highly
knowledgeable on the virtual desktop and application environment and components.
Should a complex issue arise that exceeds level two‟s abilities, it is escalated to the third active
support tier… level three. Level three issues are complicated and often mission critical requiring
expert knowledge of the virtual desktop and application environment. Tier three support tickets should
be uncommon, amount to no more than 5% of all support issues, and be resolved as soon as possible.
The final tier, level four, is not involved in active support of a production environment. It is focused
solely on the strategic improvements for the solution, testing new technologies, planning migrations,
and other high level changes.
Should support discover an issue is related to an application or underlying infrastructure, the ticket is
handed to the appropriate team for troubleshooting. If a program bug is discovered, the issue is then
re-escalated and a ticket is established with the program‟s vendor.
In parallel with the active support structure, self-service portals are available for end users to assist
themselves with basic troubleshooting tasks. This assists in resolving frequently encountered issues
and helps prevent the creation of new support tickets.

Support Level Description Responsibilities Skill Set
Level 1
(Call Entry)
Provide first-line support
of reported issues.
Initially, servicing support
messages and phone
calls. This level needs to
perform initial issue
analysis, problem
definition, ticket routing,
and simple issue
resolution. Additionally
can handle requests for
application access or
support with configuring
plugins.
 Perform issue definition,
initial analysis and basic
issue resolution
 Perform initial
troubleshooting to
determine the nature of the
issue
 Collect basic user and
session information
 Create ticket and log all
troubleshooting steps
performed
 Resolve basic Citrix related
issues, connectivity
problems and application
related issues using
existing Wiki articles
 Escalate issue to Tier-2 if it
requires advanced skills or
elevated permissions
 If the issue is related to
particular applications or
other technologies and not
the Citrix infrastructure,
escalate the issue to the
appropriate application or
technology support units
 If the issue is deemed
important enough, escalate
directly to Tier-3
 Generate requests for
additional issue resolution
guides as necessary
Tier-1 support personnel should
be provided with basic training
on Citrix XenApp, Citrix
XenDesktop and supporting
technologies. This can include
internal training from subject
matter experts or formal training.
The training provided should
focus on the following topics:
 High-Level overview of the
XenApp and XenDesktop
implementation
 Using Citrix Desktop
Director to manage user
sessions
 Troubleshooting Citrix
XenApp and XenDesktop
sessions
 Managing AD group
policies user profiles
 Troubleshooting
methodology
In addition, regular training
should be provided to the Tier-1
team members on the latest
troubleshooting
recommendations from the Tier-
2 and Tier-3 teams as well as
details on any relevant changes
to the environment. This will
help ensure a base knowledge
level amongst the team and
consistent customer service.

Level 2
(Production
Support
Engineer)
Primarily supporting day-
to-day operations of
virtual desktops
environment, may
include proactive
monitoring and
management. In
addition, this role would
also perform advanced
troubleshooting and
utilize available
monitoring / advanced
troubleshooting tools.
Assist with resolving
issues escalated by
Level One Support.
 Perform intermediate issue
analysis and resolution
 Identify root cause of
issues
 Respond to server alerts
and system outages
 Review vendor knowledge
base articles
 Respond to out-of-hours
helpdesk calls
 Respond to critical
monitoring alerts
 Generate internal Wiki
articles and issue
resolution scripts and
maintain Tier-1
troubleshooting workflows
 Perform basic server
maintenance and
operational procedures
 Manage user profiles and
data
 Escalate ticket to Tier-3 or
appropriate technology
owner if advanced skills or
elevated permissions are
required
 Generate requests for
additional issue resolution
scripts and Wiki articles as
necessary
All members of the Tier-2 team
should achieve the Citrix
Certified Administrator (CCA)
certification for both XenApp
and XenDesktop.
In addition, basic and advanced
training on Windows concepts
will be essential for team
members who do not have
desktop or server support
experience.
Finally, on-the-job training along
with close integration with Tier-3
administrators is essential as
the Tier-2 roles are formalized
and responsibilities are handed
over from Tier-3 to Tier-2.
Regular training should be
delivered to the Tier-2 team to
maintain competencies on
technology and architectural
updates. This will ensure a
baseline knowledge level across
the team.

Level 3
(Build
Engineer)
Central point for
implementing,
administering and
maintaining Citrix
desktop and application
virtualization
infrastructure. This role
focuses on deploying
new use cases and
leading lifecycle
management initiatives.
Generally, one Build
Engineer could focus on
one use-case at a time.
For example, three new
concurrent use cases
would require three Build
Engineers. Escalates
issues to software
vendor specific Technical
Support and notifies
Level 4 about this issue.
 Perform advanced issue
analysis and resolution
 Perform maintenance and
environment upgrades
 Addressing high severity
issues and service outages
 Manage the Citrix
environment
 Oversee and lead
administrative tasks at the
Tier-2 level
 Manage network and
storage infrastructure as it
relates to the Citrix
environment
 Review periodic reports of
server health, resource
usage, user experience,
and overall environment
performance
 Review vendor knowledge
base articles and newly
released updates
 Perform policy-level
changes and make Active
Directory updates
 Review change control
requests that impact the
Citrix environment
 Perform advanced server
and infrastructure
maintenance
 Review Wiki articles and
issue resolution scripts for
accuracy, compliance, and
feasibility
 Create Wiki articles and
issue resolution scripts to
address Tier-2 requests
Tier-3 support team members
with a minimum of three years of
experience implementing and
supporting XenApp,
XenDesktop, Provisioning
Services and Windows
operating systems.
The Tier-3 team should have
enterprise experience with
industry recognized
certifications.
The Tier-3 administrators should
complete the Citrix Certified
Enterprise Engineer certification
track as this will prepare them to
proactively manage the user
community and the Citrix
solution. Citrix Certified
Integration Architect (CCIA)
level certifications would be
optimal at this level.

Level 4
(Architect)
The Level 4 team has
minimal exposure to
administrative tasks but
focuses on translating
business requirements
into technical
architectures, designing
the infrastructure or
planning migrations. Not
involved in day-to-day
support.
 Provide technical
leadership for upcoming
projects
 Lead design updates and
architecture revisions
 Address high severity
issues and service outages
 Oversee technology
integration workflows
 Review periodic reports of
server health, resource
usage, user experience,
and overall environment
performance to determine
next steps and upgrade
paths
 Review change control
requests which impact the
Citrix environment
 Review frequently recurring
helpdesk issues
 Ensure technical
specifications continue to
meet business needs
 Update design
documentation
A qualified Tier-4 resource
should have a minimum of five
to six years of experience
implementing, supporting, and
serving in a technology architect
role for a XenApp and/or
XenDesktop environment as
well as additional administrative
experience with integrated
technologies such as application
and profile management
solutions. The ideal candidate
will have served in such a
capacity at two or more
environments for purposes of
product exposure and in at least
one environment of over 1,200
concurrent users.
The Tier-4 resource should have
enterprise experience with
industry recognized
certifications. A Citrix Certified
Integration Architect (CCIA)
certification or comparable
training and experience should
be a prerequisite of the role.
Training or practical knowledge
on XenCenter, AD Group
Policies and the utilized
hypervisor should be considered
essential.
Vendor
Support
Vendor assistance may be necessary should defects in a program be discovered. At this stage
level 3 engineers need to establish a support tickets with the appropriate vendor to assist with
finding a solution.
Self Service A self-service portal is utilized for noncritical tasks such as application access, permissions, etc.
The portal can range from a simple FAQ page to a fully automated process requiring no human
interaction. The purpose of the self-service portal is to add an additional touch point for end
users to address basic issues, preventing the creation of new support tickets.
Table 64: Helpdesk Support Levels
Tools
Providing the appropriate tools for the aforementioned support levels is a key aspect in order to
achieve a high First-Time-Fix-Rate and a low Time-to-Resolution. The table below highlights several
tools available for supporting Citrix environments.

Tools Details
Ticket Management System Used to document customer information and issues. Help Desk should also
document date and time of occurrence, the application and module, steps to
repeat error, obtains screenshots as needed.
Call Scripts The first contact help desk personnel should have documented scripts or
triage guides to ensure that all relevant data is captured while the user is on
the phone. This practice also assists in proper triage and allows the next
support tier to perform research off-line and review logs prior to customer
contact.
Shadowing Tool Shadowing users is a useful way for troubleshooting new issues. This allows
for support technicians and administrators to remotely observe a user‟s
actions.
Knowledge Base Documentation should be created and maintained in a knowledge base or
library of known issues. Articles should be searchable for quick recovery.
Knowledge bases help quickly resolve issues upon reoccurrences and reduce
the need to perform future research.
Citrix Administrative
Consoles
The incident management team will have access to the Citrix Administrative
Consoles for focused troubleshooting. Citrix recommends that the Help Desk
also have same level of access. The initial support tier will primarily focus on
resetting connections and obtaining data for escalation.
Table 65: Helpdesk Tools
The appropriate tools should be available to their respective support levels as follows:
Tool Description
Products Support Level
XD XA PVS XS L1 L2 L3
Desktop
Director
Desktop Director provides an overview of
XenDesktop hosted desktops and XenApp
sessions. It enables support teams to perform basic
maintenance tasks and to monitor and troubleshoot
system issues.
X X X X
Desktop
Studio
Desktop Studio enables administrators to perform
configuration as well as maintenance tasks for a
XenDesktop site and associated virtual desktops.
X X X
AppCenter AppCenter provides support teams the ability to
monitor and troubleshoot user sessions as well as
server systems within a XenApp farm. In addition
AppCenter enables administrators to perform
configuration and maintenance tasks for a XenApp
farm.
X X X
HDX
Monitor
HDX Monitor is a tool to validate the operation of
the Citrix ICA/HDX stack of a user session. Hereby
HDX Monitor provides information about client
capabilities, network performance / activity, session
settings and many more items.
X X X X
Provisioning
Services
Console
The Provisioning Services Console enables
administrators to perform configuration and
maintenance tasks for a Provisioning Services farm.
X X X
XenCenter XenCenter enables administrators to perform
configuration and maintenance tasks for a
XenServer Resource Pool.
X X

Table 66: Helpdesk Tool Assignment
Delegated Administration
In addition to providing the appropriate tools for the appropriate support levels, it is also important to
provide each support level with the proper rights so they can be effective in their role. The following
provides recommendations on delegated administration rights within the different management
consoles.
XenApp Delegated Rights
Console Permission Tier-1 Tier-2 Tier-3 Tier-4
Farm Node
Farm Management √
Edit All Other Farm Settings √
Edit Configuration Logging
Settings
√
Edit Zone Settings √
View Farm Management √ √
Administrators Node
Administrators √
Edit centrally configured Web
Interface sites
√
Log on to Management
Console
√ √ √ √
View Administrators √ √
Applications Node
Published Applications √ √
Publish Applications and Edit
Properties
√ √
View Published Applications
and Content
√ √ √
Servers √ √ √
Terminate Processes √ √ √
Sessions √ √ √
Connect Sessions √ √ √
Disconnect Users √ √ √
Log Off Users √ √ √
Reset Sessions √ √ √
Send Messages √ √ √
View Session Management √ √ √ √
Load Evaluators Node
Load Manager √

Assign Load Evaluators √
Edit Load Evaluators √
View Load Evaluators √ √
Load Balancing Policies Node
Load Balancing Policies √
Edit Load Balancing Policies √
View Load Balancing policies √ √
Policies Node
Policies √
Edit Policies √
View Policies √ √
Printer Management Node
Printers √
Edit Printer Drivers √
Replicate Printer Drivers √
View Printers and Printer
Drivers
√ √
Servers Node
Published Applications √
Assign Applications to
Servers
√ √
Servers √
Edit Other Server Settings √
Move and Remove Servers √
Terminate Processes √
View Server Information √ √
Sessions √
Connect Sessions √
Disconnect Users √
Log Off Users √
Reset Sessions √
Send Messages √
View Session Management √ √
Worker Groups
Assign Applications to Worker
Groups
√
View Worker Groups √ √

Table 67: XenApp Delegated Rights
For further information about delegated rights within a XenApp Farm, please refer to
eDocs - Delegating Tasks to Custom Administrators.
XenDesktop Delegated Rights
Admin Role Support Tier
Help Desk Administrator Tier-1
Assignment Administrator Tier-2
Full Administrator Tier-3
Read-Only Administrator Tier-4
Table 68: XenDesktop Delegated Rights
For further information about delegated rights within a XenDesktop Site, please refer to eDocs –
XenDesktop Delegated Administration.
Provisioning Services Delegated Rights
- Tier-1
Device Operators Tier-2
Farm Administrator Tier-3
Farm Administrator Tier-4
Table 69: Provisioning Services Delegated Rights
For further information about delegated rights within a Provisioning Services Farm, please refer to
eDocs – Provisioning Services Managing Administrative Roles.
XenServer Delegated Rights
- Tier-1
Virtual Machine Operator Tier-2
Pool Administrator Tier-3
Read-Only Tier-4
Table 70: XenServer Delegated Rights
For further information about delegated rights within a Provisioning Services Farm, please refer to
CTX130420 – XenServer 6.0 Administrators Guide (see chapter Role Based Access Control).

Testing
Regular updates and maintenance is an everyday part of Citrix environments. Standard processes must
be closely followed to ensure modifications do not negatively impact a production environment. This
includes maintaining dedicated infrastructure where modifications are tested prior to production and
processes that ensure changes are deliberate and accountable.
The process for testing should resemble the following:
Figure 30: Testing Process
Change Control
Standardized processes that manage application lifecycles through installation, updating, and end of
life are necessary to ensure consistent and accountable performance of a farm. The change control
process should be closely followed starting with a change request. For any changes to be made to a
production environment a change request form must first be filled out detailing changes requested,
reasons for the change, and intended timeframes for the action. This is then reviewed and edited if
required by a change manager and then advisory board. When the change request has gone through
the entire change approval process it is then given to a change implementer who then stages the
change for testing, and finally conducts the implementation in production. The table and steps below
further detail the process:
Change
Request
Development Testing
Pre-
production
Production

Figure 31: Change Control Process

The process is as follows:
1. The Request for Change (RfC) form is completed by any person requesting a change
2. After appropriate manager approvals have been acquired, the RfC is forwarded to the
appropriate Change Manager(s).
3. The Change Manager validates the RfC for completeness and logs the RfC information into
the Change Control Log for tracking. Incomplete CR‟s are returned to the requestor for update
and re-submission.
4. The Change Manager assesses the impact of the change in conjunction with subject matter
experts and/or managers of the teams associated/affected by this change.
5. The Change Manager works with the associated/affected teams as well as the change
requestor in order to confirm the priority, category and type of the change as well as the
proposed back-out plan.
6. If the process succeeds the RfC is forwarded to the Change Advisory Board for approval. If
the process does not succeed the change control log is updated with the current status as well
as the reason of the rejection and the RfC is send back to the requestor.
7. The Change Advisory Board reviews and validates the change in detail, and discusses and
evaluates purpose, reasons, impact, cost and benefits. Each Board member represents their
department and provides guidance on the change requests. The Change Advisory Board also
reviews multiple requests to coordinate implementations and “package” requests into a single
release schedule.
8. Upon approval the change is send back to the Change Manager to schedule the change for
implementation into the staging environment.
9. The change is implemented and test are conducted. The results are sent back to the Change
Manager.
10. If the staging implementation and testing was successful the change is scheduled for
production implementation. In case the staging phase was not successful another staging
iteration will be conducted.
11. The change is rolled out to production.
12. The Change Manager reviews the implementation and finally updates the Change Control
Log.
13. On a periodic basis, the Change Manager reviews the Change Control Log to analyze for
trends on type, frequency and size of changes and forwards the results to the Change
Advisory Board for review.
In the instance of an emergency, the processes may be expedited. Should an issue be declared an
emergency, a change request form is still filled out and delivered to the appropriate change
management representative. When approved, the requested change is immediately implemented and
the advisory board notified. Due to the criticality of time in an emergency, timeframes are not required.

Testing Phases
Since changes to Citrix infrastructure can impact thousands of virtual desktop and application users,
multi-phase testing is critical for the reliability of the environment. As such, it is a best practice to
implement the following testing strategy:
 Development: The development infrastructure exists outside of the production network.
Typically, it consists of short-lived virtual machines whose configuration should match the
configuration in production as close as possible. The purpose of the development phase is to
provide change requestors a non-production environment to perform proof of concepts,
determine integration requirements and perform iterative testing as part of a discovery phase
to making changes. Proposed changes should be documented to they can be applied in the
test phase.
 Test: The test environment is a standalone 1:1 copy of the production infrastructure and is
used to confirm that the proposed changes can be easily repeated prior to the pre-production
staging environment. The changes made should follow exactly from the documentation
provided from discoveries made in the development stages. If testing fails within the testing
stage, the architect must determine the severity of failure and determine whether minor
updates to documentation is sufficient or a full development cycle is needed.
 Pre-production: The staging environment can be co-located with the production environment
as the changes should be tightly managed. The goal of staging is to implement the proposed
changes with little risk or uncertainty. It is expected that any changes made to the staging
infrastructure have been tested and documented for repeatability. There should not be any
iterations or adjustments required within this phase. For updates, the starting point for all
systems in the staging environment should be the current production status. However, for
new deployments, the architect will determine the appropriate start point; whether an existing
system or newly created. Suring this phase and within this environment User Acceptance
Testing (UAT) should be performed.
 Production: The production environment is the fully redundant and scalable solution for
normal usage of the XenDesktop and XenApp environments. There should be minimal
changes to the environment.
Operations
It is best practice to keep operating systems, applications, and Citrix components all up to date with
applicable hotfixes and updates to ensure stability, performance, and security for the environment. In
general, Citrix recommends that
 Hotfix Rollup Packs and Service Packs always be installed
 General hotfixes be installed on an “as needed” basis only
 All changes follow standardized change management processes for any changes
It is recommended that support staff and administrators involved in implementation review the product
documents, articles, and guides to support, manage and administer the environment. All of these
documents are available from the Citrix Product Documentation Library
(http://support.citrix.com/proddocs/index.jsp) or on the Citrix Knowledge Base (http://support.citrix.com).
Each operation is categorized by the associated component of the solution as well as the frequency of the
operation (daily, weekly, and yearly).

Daily
The following table outlines the necessary tasks that should be undertaken per component by the
appropriate administrators on a daily basis:
Component Task Description
XenDesktop Event log checking Checking of the event logs on all Citrix XenDesktop (Desktop Delivery
Controller) servers including Windows System, Application and Error
logs.
XenDesktop Citrix XenDesktop
Alerts
Checking of the Citrix XenDesktop alerts in Desktop Studio needs to be
performed at least once per day to ensure no issues have occurred with
the Citrix XenDesktop environment.
XenDesktop Check Citrix
XenDesktop logs
Checking of Citrix XenDesktop specific application logs including the
Pool Management Logs. To assist with troubleshooting efforts, refer to
Comprehensive List of Event Log Messages for XenDesktop on
http://support.citrix.com/article/CTX119978
XenDesktop Test Virtual Desktop
Access
Simulate a connection to ensure access to Desktop Groups (Pooled
and Assigned) is available to users.
XenApp Event log checking Checking of the event logs on all Citrix XenApp servers including
Windows System, Application and Error logs.
XenApp Reboot schedule
checking
All Citrix XenApp servers are rebooted on a regular basis as part of an
automated process. Ensuring the reboot of the servers has been
successful is essential to maintain server health and performance.
XenApp Citrix XenApp Alerts Checking of the Citrix XenApp server alerts in the Access Management
Console needs to be performed at least once per day to ensure no
issues have occurred with the XenApp environment.
XenApp Citrix XenApp
commands
Qfarm /load - to check that the load of the Citrix XenApp servers are
reporting correctly
Qfarm /zone - to confirm that the correct Zone Data collector is
assigned
Qfarm /online - to confirm all Citrix XenApp servers are online
XenApp Test Application
Access
Simulate an ICA connection to ensure key applications are available to
users.
Provisioning
Services
Event log checking Checking of the event logs on all Citrix Provisioning Servers including
Windows System, Application and Error logs.
Provisioning
Services
Check Citrix
Provisioning Server
logs
Checking of Citrix Provisioning Server specific application logs including
the following:
Citrix Provisioning Server logs - %APPDATA%CitrixProvisioning
Serviceslogs
Reference of PXE Error Codes -
http://support.citrix.com/article/CTX120956
Provisioning
Services
Check Citrix
Provisioning Server
utilization
Check the number of target devices connected to the Citrix Provisioning
Servers and balance the load across the servers, if required.
Table 71: Daily Operations

Weekly
appropriate administrators on a weekly basis:
XenDesktop Review Access
Management Console
Check the status of the Citrix XenDesktop Desktop Groups, Idle Pool
count and connectivity to backend hosting infrastructure.
XenDesktop Hotfixes and Patches Review the latest Citrix hotfixes and ascertain whether the
XenDesktop Controller and virtual desktops require them.
XenDesktop Printer Drivers Review the local installation of printer drivers on the virtual desktops,
if required for auto-creation.
XenDesktop Review Citrix Knowledge
Base for latest incidents
Review incidents logged for any common issues related to Citrix
XenDesktop on http://support.citrix.com/
XenDesktop Backup Citrix
XenDesktop Database
Perform full-data backup of the Citrix XenDesktop Datastore (could
be a daily task due to size of the transaction log)
XenDesktop Backup User Data Backup user-related file data.
XenApp Administration
Configuration and
Logging Reports
Confirm that Citrix XenApp farm changes taken place during the
previous week have been approved.
XenApp Citrix XenApp
Commands to be run
DSCheck.exe - verify that the contents of the datastore are still valid.
XenApp Hotfixes and Patches Review the latest Citrix hotfixes and ascertain whether the Citrix
XenApp farm requires them.
XenApp Review Citrix Knowledge
XenApp on http://support.citrix.com/
XenApp Printer Drivers Review the local installation of printer drivers on each Citrix XenApp
server to ensure no rogue drivers have been installed
XenApp Backup Citrix XenApp
related databases
Perform full-data backups of the following Citrix XenApp databases:
Citrix XenApp Datastore
Citrix SmartAuditor Datastore
Citrix XenApp Configuration Logging Database
XenApp Backup User Data Backup user-related file data.
XenApp Backup Application
Profiles
Perform full-data backup of the Citrix XenApp application profiles
used for streaming.
Provisioning
Services
Backup vDisk Image
Files
Backup of Citrix Provisioning Server vDisk files (.VHD) on local and/or
shared storage.
Provisioning
Services
Backup Citrix PVS
Datastore
Backup of Citrix Provisioning Server database hosted on SQL Server
infrastructure.
Provisioning
Services
Hotfixes and Patches Review the latest Citrix hotfixes and ascertain whether the Citrix
Provisioning Server farm requires them.
Provisioning
Services
Review Citrix Knowledge
Provisioning Server on http://support.citrix.com/

Provisioning
Services
Perform vDisk Updates Update the „golden‟ vDisk image files and apply the following:
Windows software updates and patches
Operating system and application changes
Anti-virus pattern and definitions updates
Deploy updated vDisk version to target devices post update process
Provisioning
Services
Check Storage Capacity Review the storage utilization – used and free storage space for
vDisks and write cache files.
Provisioning
Services
Check Auditing Reports Review the Citrix Provisioning Server Auditing Logs.
Provisioning
Services
Check Network and
Streaming Services
Confirm DHCP, TFTP (or PXE, if applicable) and Streaming services
are functional.
Table 72: Weekly Operations
Yearly
appropriate administrators on a yearly basis:
XenDesktop Perform restore test Conduct backup tests to validate the actual restore process is
functioning correctly.
XenDesktop Citrix Policy Assessment Review Citrix Policies and determine whether new policies are
required and existing policies need to be updated.
XenDesktop Software Upgrade Review and asses the requirement for new Citrix XenDesktop
software releases/versions.
XenDesktop Capacity Assessment Perform capacity assessment of the Citrix XenDesktop environment.
XenApp Perform restore test Conduct backup tests to validate the actual restore process is
XenApp Citrix Policy Assessment Review Citrix Policies and determine whether new policies are
required or existing policies need to be updated.
XenApp Perform BCP/DR test Conduct functional Business Continuity Plan (BCP)/Disaster
Recovery (DR) test to confirm DR readiness.
XenApp Software Upgrade Assess the validity of applications. Review and asses the requirement
for new Citrix Web Interface software releases/versions.
XenApp Capacity Assessment Perform capacity assessment of the Citrix XenApp environment to
determine environment utilization and any scalability requirements.
Provisioning
Services
Archive Audit Reports Perform an archive of the Citrix Provisioning Server Audit Trail
Information for compliance requirements.
Provisioning
Services
Update vDisk Workloads Update the vDisk desktop and/or server workloads (.VHD files) to
reflect the current standard operating environment, if required.
Provisioning
Services
Perform restore test Conduct backup tests to validate the actual restore process is
Provisioning
Services
Perform BCP/DR test Conduct functional Business Continuity Plan (BCP)/Disaster
Recovery (DR) test to confirm DR readiness.
Provisioning
Services
Software Upgrade Review and asses the requirement for new Citrix Provisioning Server
software releases/versions.

Provisioning
Services
Capacity Assessment Perform capacity assessment of the Citrix Provisioning Server
environment to determine environment utilization and scalability
requirements including server, storage and network components.
Table 73: Yearly Operations

Monitoring
Monitoring of a new environment enables administrators to address issues proactively. By having an in
depth understanding of current and expected behavior of the various components, administrators are
better equipped to discover an issue before it impacts the user community.
Make sure to define the monitoring activities that administrators will perform on a periodic basis, as well
as the tools that will be used. In addition, determine which reports will be made available to
management.
A variety of built-in and customizable reports exist within the Citrix products so that administrators can
track trends, effectively plan capacity, and address auditing requirements. In addition, troubleshooting is
typically minimized when administrators fully utilize the tools integrated into the Platinum edition, as well
as other third party tools.
Citrix recommends implementing a monitoring solution, which covers all aspects of an IT environment and
allows aggregating all captured data within a single tool or console. This simplifies correlating events or
counters of multiple components and enables administrators to easily get an end-to-end overview. This
can decrease the time required to determine the root cause of an issue significantly.
Monitoring a virtual desktop and application environment is broken down into the following aspects:
 Performance: This section outlines the key performance counters, which should be tracked
within every environment.
 Services: This section outlines the Citrix and Windows services which play a critical role within
desktop and application virtualization environment. In addition information about the criticality of
every service is provided.
 Events: Within this section, monitoring of the Windows event log is discussed and links to event
log reference documentations are provided.
 Growth: This section provides guidelines on how to monitor and plan for growth.

Performance
Monitoring the performance of the overall environment is crucial towards making sure users have an acceptable experience. Different components
within the overall solution require monitoring of unique metrics with appropriately set thresholds.
The metrics and thresholds presented are based on real world experience but may not apply to all environments. Organizations will need to
perform their own baselining, validity testing and validation before implementing within a production environment.
Note: Some hypervisors, such as VMware vSphere, provide specific performance counters for tracking CPU and Memory utilization within virtual
machines (i.e. “VM Processor % Processor Time”). These performance counters should be used in addition to the general counters listed below.
Component Metric Description Warning
(Yellow)
Critical
(Red)
Troubleshooting / Remediation
General
All servers
and virtual
machines
Processor -
% Processor
Time
% Processor Time is the percentage of elapsed time
that the processor spends to execute a non-Idle thread.
It is calculated by measuring the duration of the idle
thread is active in the sample interval, and subtracting
that time from interval duration. (Each processor has an
idle thread that consumes cycles when no other threads
are ready to run). This counter is the primary indicator of
processor activity, and displays the average percentage
of busy time observed during the sample interval. It is
calculated by monitoring the time that the service is
inactive and subtracting that value from 100%.
80% for 15
minutes
90% for 15
minutes
Identify the processes/services
consuming processor time using Task
Manager or Resource Monitor.
If all processes/services work within
normal parameters and the level of CPU
consumption is an expected behavior it
should be considered to add additional
CPU resources to this system in the
future.
If a process/service can be identified
which works outside normal parameters,
the process should be killed. Please note
that killing a process can cause unsaved
data to be lost.
General
All servers
and virtual
machines
System -
Processor
Queue Length
Processor Queue Length is the number of threads in the
processor queue. Unlike the disk counters, this counter
shows ready threads only, not threads that are running.
There is a single queue for processor time even on
computers with multiple processors. Therefore, if a
computer has multiple processors, you need to divide
this value by the number of processors servicing the
workload. A sustained processor queue of less than 10
threads per processor is normally acceptable,
dependent of the workload.
5 (per
Core) for 5
minutes
10 (per
Core) for 10
minutes
A long CPU queue is a clear symptom of
a CPU bottleneck. Please follow the
steps outlined for counter “Processor -
% Processor Time”

General
All servers
and virtual
machines
Memory –
Available
Bytes
Available memory indicates the amount of memory that
is left after nonpaged pool allocations, paged pool
allocations, process‟ working sets, and the file system
cache have all taken their piece.
<30% of
total RAM
<15% of
total RAM
Identify the processes/services
consuming memory using Task Manager
or Resource Monitor.
normal parameters and the level of
memory consumption is an expected
behavior it should be considered to add
additional memory to this system in the
future.
data to be lost.
General
All servers
and virtual
machines
Memory –
Pages/sec
Pages/sec is the rate at which pages are read from or
written to disk to resolve hard page faults.
>10 >20 A high value reported for this counter
typically indicates a memory bottleneck,
except if “Memory – Available Bytes”
reports a high value at the same time. In
this case most likely an application is
sequentially reading a file from memory.
Please refer to KB139609 for further
information.
General
All servers
and virtual
machines
Paging File -
%Usage
This is the percentage amount of the Page File instance
in use.
>40% >70% Review this value in conjunction with
“Memory - Available Bytes” and
“Memory - Pages/sec” to understand
paging activity on the affected system.
General
All servers
and virtual
machines
LogicalDisk/P
hysicalDisk -
% Free Space
% Free Space is the percentage of total usable space
on the selected logical disk drive that is free.
<10% of
physical
disk
<5% of
physical disk
Identify which files or folders consume
disk space and delete obsolete files if
possible. In case no files can be deleted,
consider increasing the size of the
affected partition or add additional disks.

General
All servers
and virtual
machines
LogicalDisk/P
hysicalDisk -
% Disk Time
% Disk Time marks how busy the disk is. >70%
consistently
>90%
consistently
Identify the processes / services
consuming disk time using Task
Manager or Resource Monitor.
normal parameters and the level of disk
consumption is an expected behavior it
should be considered to move the
affected partition to a more capable disk
subsystem in the future.
data to be lost.
General
All servers
and virtual
machines
LogicalDisk/P
hysicalDisk –
Current Disk
Queue Length
Current Disk Queue Length provides a primary measure
of disk congestion. It is an indication of the number of
transactions that are waiting to be processed.
>=1 (per
spindle)
consistently
>=2 (per
spindle)
consistently
A long disk queue length typically
indicated a disk performance bottleneck.
This can be caused by either
processes/services causing a high
number of I/Os or a shortage of physical
memory. Please follow the steps outlined
for counter “LogicalDisk/PhysicalDisk -
% Disk Time” and counter “Memory –
Available Bytes”
General
All servers
and virtual
machines
LogicalDisk/P
hysicalDisk –
Avg. Disk
Sec/Read
– Avg. Disk
Sec/Write
– Avg. Disk
Sec/Transfer
The Average Disk Second counters show the average
time in seconds of a read/write/transfer from or to a disk.
>=15ms
consistently
>=20ms
consistently
High disk read or write latency indicates
a disk performance bottleneck. Systems
affected will become slow, unresponsive
and application or services may fail.
Please follow the steps outlined for
counter “LogicalDisk/PhysicalDisk - %
Disk Time”

General
All servers
and virtual
machines
Network
Interface –
Bytes
Total/sec
Bytes Total/sec shows the rate at which the network
adaptor is processing data bytes. This counter includes
all application and file data, in addition to protocol
information, such as packet headers.
< 8 MB/s
for 100
Mbit/s
adaptor
<80 MB/s
for 1000
Mbit/s
adaptor
None Identify the processes / services
consuming network using Task Manager
or Resource Monitor.
normal parameters and the level of
bandwidth consumption is an expected
behavior it should be considered to move
the respective process/service to a
dedicated NIC (or team of NICs).
data to be lost.
XenApp
All XML
Brokers
Application
Resolution
Time (ms)
The time in milliseconds that a resolution took to
complete. This is also the time required to determine
the “least-loaded” server during an application
launch. A baseline would need to be established in
the environment in order to accurately establish
threshold values.
Based on
baseline
values
Based on
baseline
values
XenApp
All servers
DataStore
Connection
Failure
The number of minutes that the XenApp server has
been disconnected from the data store. Threshold
should take into account events such as reboots and
scheduled maintenance.
>=1 None
XenApp
All XML
Brokers
Number of
busy XML
threads
The number of XML threads currently being
processed. There are 16 worker threads in the Citrix
XML Service. A count of 16 or more shows that
XML requests are not being processed in a timely
manner.
>=10
consistentl
y
>=16
consistently
XenApp
All XML
Brokers and
Controllers
Resolution
WorkItem
Queue Ready
Count
The number of resolution work items (related to
application launches) that are ready to be executed.
A value above 0 indicates that requests are being
queued while IMA handles other requests.
None >=1

XenApp
All XML
Brokers and
Controllers
WorkItem
Queue Ready
Count
The number of work items that are ready to be
executed. A value above 0 indicates that requests are
being queued while IMA handles other requests.
This counter should not be over 1 for an extended
period of time.
None >=1
XenApp
All servers
Citrix
Licensing –
License
Server
Connection
Failure
Displays the number of minutes that XenApp has
been disconnected from the License Server.
>1 minute >1440
minutes
XenApp
All servers
Citrix
Licensing –
Last
Recorded
License
Check-Out
Response
Time
Displays the last recorded license check-out
response time in milliseconds. A value of more than
5000 milliseconds indicates a performance issue at
the license server.
>2000ms >5000 ms
XenDesktop
All controllers
Database
Avg.
Transaction
Time
The time on average, in seconds, taken to execute a
database transaction. A baseline needs to be
established in the environment in order to accurately
establish threshold values.
Based on
baseline
values
Based on
baseline
values
In case the reported values exceed the
baseline response time constantly, a
potential performance issue needs to be
investigated at the SQL server level.
XenDesktop
All controllers
Database
Connected
Indicates whether this service is in contact with its
database. (1 is connected; 0 is not connected)
None None Both values report connectivity issues of
the XenDesktop Broker service with the
XenDesktop database. In case issues
are reported, SQL server and network
availability needs to be verified.
XenDesktop
All controllers
Database
Transaction
Errors/sec
The rate at which database transactions are failing. None >0 Both values report connectivity issues of
the XenDesktop Broker service with the
XenDesktop database. In case issues
are reported, SQL server and network
availability needs to be verified.

Web Interface
All servers
ASP.NET –
Request
Queued
The number of requests waiting to be processed by
ASP. A baseline needs to be established in the
environment in order to accurately establish
threshold values.
Based on
baseline
values
Based on
baseline
values
In case the queue length exceeds the
critical limit requests may be rejected.
In this case it should be considered to
add additional Web Interface servers
to the load balancing team in order to
distribute the load across more nodes.
Web Interface
All servers
ASP.NET –
Requests
Rejected
The number of requests rejected because the request
queue was full.
None >=1 When this limit is exceeded, requests
will be rejected with a 503 status code
and the message "Server is too busy."
Please follow the steps outlined for
counter “ASP.NET – Request
Queued”
Table 74: Recommended Metrics to Monitor

Services
Windows services that are critical to basic server functionality should be automatically monitored to ensure that they are running properly. The
following table provides a list of the common Windows services that should be monitored. The recommended recovery actions for the services
listed below are as follows:
 First failure: Restart the Service
 Second Failure: Restart the Service
 Subsequent Failures: Put the server in maintenance mode and investigate the root cause
Component Service Functionality Risk
XenApp Citrix
Independent
Management
Architecture
Provides management services and server-to-server
communication for Citrix products.
Dependencies:
 IPSEC Services
 RPC Service
 Server
 WMI Driver Extensions
 Workstation
Failure of this service results in the server not allowing any
additional users. The server will also not respond to IMA
communications from the management console or from any other
Citrix component.
XenApp Citrix XML
Service
Provides an HTTP interface to the ICA browser. It uses TCP
packets instead of UDP, which allows connections to work
across most firewalls. The default port for the Citrix XML
Service is 80.
Frequently, the Citrix XML Service is plugged into IIS.
Therefore the Citrix XML Service may not appear. This does
not present a problem.
Dependencies:
 Independent Management Architecture Service
The failure of the Citrix XML Service is critical on the zone data
collectors. These servers are utilized when the Web Interface
tries to determine the least loaded server. If this server fails, the
Web Interface will look at the backup server. This can slow down
the application enumeration and launching process.
XenApp Citrix XML
Service
The Secure Ticket Authority (STA) service is embedded into
the Citrix XML service.
Is used by the Access Gateway and Web Interface to create and
validate secure tickets. Generally more than one STA is specified
for fault tolerance however it should be monitored.
XenApp Citrix XTE
Server Service
Provides necessary Session Reliability.
Dependencies:
 None
This service needs to be monitored if the Session Reliability is
enabled.

XenApp Citrix CPU
Utilization
Management/Re
source
Management
Manages resource consumption to enforce entitlement
policies
Dependencies:
 RPC Service
These services need to be monitored if CPU optimization is
enabled. (Automatically started if CPU Optimization is enabled in
the XenApp Console.)
XenApp Citrix CPU
Utilization
Management/Us
er Session sync
Synchronizes the CPU Utilization Management user ID of a
process with the user ID of the session owner of the process
Dependencies:
 Citrix CPU Utilization Management/Resource
Management
XenApp Citrix CPU
Utilization
Management/C
PU Rebalancer
Installed on Windows Server 2003 multi-processor systems;
it is not installed on Windows 2000 Server or servers with
only one processor. The CPU Rebalancer service is used to
alleviate a Microsoft issue that appears in environments
where a lot of short-lived processes are started and
stopped. Due to the performance impact the CPU
Rebalancer service can have, by default it is set to “Manual.”
If your environment is running many short-lived applications
that all appear to be running on the same CPU, setting the
service to “Automatic” is recommended. The CPU
Rebalancer service corrects this by balancing the load
equally across processors.
Monitored if service is set to „automatic‟. (Set to „manual‟ by
default and must be manually configured to start automatically on
Windows Server 2003 with multiple processors.)
XenApp Citrix Print
Manager
Service
Supports the Citrix Advanced Universal printing architecture
Dependencies:
 Print Spooler
 RPC Service
This service needs to be monitored if the Citrix Advanced
Universal Printer Driver is used
XenApp Citrix Virtual
Memory
Optimization
Dynamically optimizes applications running on the Citrix
XenApp
Dependencies:
 None
This service needs to be monitored if memory optimization is used
(Automatically started if Memory Optimization is enabled in the
XenApp Console.)
XenApp Citrix MFCOM
Service
Provides access to necessary XenApp objects.
This service will be present if the license server exists on a
server running Citrix XenApp.
Dependencies:
 RPC Service

XenApp Remote Desktop
Services/Termin
al Services
Allows multiple users to be connected interactively to a
machine as well as the display of desktops and applications
to remote computers.
Dependencies:
 RPC Service
If the Remote Desktop Services service fails, the server will
become inoperable requiring a reboot.
XenApp Citrix WMI
Service
Citrix WMI Service is responsible for sending alerts to the
WMI aware monitoring tools.
This service only runs when needed. The failure of the Citrix WMI
Service will only result in not receiving updates to WMI aware
monitoring tools from the WMI Provider. It will not put the server
at severe risk, but it will hinder the ability to monitor the system.
XenApp Citrix Client
Network
Maps client drives and peripherals for access in sessions.
Applicable to XenApp 5 and older.
Dependencies:
 Client Drive Mapping (CDM)
 Workstation
On failure client drives and peripherals are not accessible.
XenApp Citrix Streaming
Service/ Citrix
Streaming
Helper Service
Manages the Citrix Offline Plug-In when streaming
applications.
Dependencies:
 RPC Service
On failure streamed applications cannot be accessed.
XenApp Server Service Provides file, print and named-pipe sharing over the network
for the server.
On failure the server will not accept new user sessions
XenApp Net Logon
Service
Provides pass-through authentication of account logon
events for systems in a domain.
On failure users of new sessions are unable to authenticate and
users of existing sessions may not be able to access network
resources.
XenApp Remote
Procedure Call
Service (RPC)
Provides the endpoint mapper and other miscellaneous
RPC services.
On failure certain network services cannot be accessed.
XenApp Workstation
Service
Creates and maintains client network connections to remote
servers.
On failure certain network services cannot be accessed.
XenApp Print Spooler Manages all local and network print queues and controls all
print jobs.
On failure printing is not possible for users connected to this
server.
XenDesktop Citrix AD Identity
Service
Manages Active Directory computer accounts.
Dependencies:
 WMI Service
Machine Creation Service relies on this service to create virtual
machines. The machine account becomes tainted at the time of
creation if this service is not able to sync with Active Directory.

XenDesktop Citrix Broker
Service
Manages connections to virtual machines and applications. If this service is stopped existing connections are not affected. No
new connections can be established.
XenDesktop Citrix
Configuration
Service
Stores configuration information on the Citrix Services.
Dependencies:
 WMI Service
If this service is stopped, it will not be able to collect information
from the Broker Services.
XenDesktop Citrix Diagnostic
Facility COM
Server Service
Manages and controls Citrix diagnostic trace sessions.
Dependencies:
 RPC Service
This service has no impact on the production environment. It is
used to generate CDF trace files which aid in troubleshooting
issues.
XenDesktop Citrix Host
Services
Manages host and hypervisor connections.
Dependencies:
 WMI Service
It is not possible to connect to the virtual machines when this
service is not available.
XenDesktop Citrix Machine
Creation Service
Creates new virtual machines.
Dependencies:
 WMI Service
The ability to create virtual machines using Machine Creation
Service will not be available if this service is stopped. Provisioning
Services is not affected.
XenDesktop Citrix Machine
Identity Service
Manages storage of virtual machines.
Dependencies:
 WMI Service
If this service is stopped XenDesktop is not able to manage the
storage of the virtual desktops.
Provisioning
Services
Citrix PVS PXE
Service
Provides the PVS PXE Boot Server functionality On failure of this service target devices may not be able to boot
successfully if PXE booting is leveraged.
Provisioning
Services
Citrix PVS
Stream Service
Streams contents of the vDisk to the target device on
demand.
If this service stopped it will not be possible to stream vDisk
images.
Provisioning
Services
Citrix PVS
SOAP Service
Provides framework for external or existing solutions to
interface with Provisioning services.
If this service fails PVS Server to PVS Server communication as
well as PVS Console to PVS Server communication is not
possible.
Provisioning
Services
Citrix PVS TFTP
Service
Provides the TFTP Server functionality On failure of this service target devices may not be able to boot if
this server is used as TFTP server for the bootstrap.
Provisioning
Services
Citrix PVS Two-
Stage Boot
Service
Provides the bootstrap functionality for devices booting by
means of a BDM ISO file
On failure of this service target devices may not be able to boot if
a BDM ISO file is used.
Citrix
Licensing
Citrix License
Service
Provides licensing service for Citrix products. Licensing mode changes to grace period when service is stopped
or License Server can‟t be contacted. If not monitored,
functionality of Citrix products will cease after grace period
expires.

Citrix
Licensing
Citrix Licensing
WMI
The Citrix License Management Console collects license
data information using the WMI service.
Web
Interface
World Wide
Web Publishing
Service
Provides web connectivity and administration through the
Internet Information Services Manager.
Dependencies:
 HTTP
 RPC Service
Access to XenApp published applications or published desktops
will not be available through Web Interface if WWW services is not
available.
Table 75: Recommended Services to Monitor

Events
Monitoring the Windows Event Log for unknown or critical events can help discover issues proactively
and allows administrators to understand event pattern. Citrix recommends monitoring the Windows
event logs of all Windows servers closely and investigate for errors pertaining to:
 Licensing: Errors in the event log dealing with Terminal Services licensing should be
investigated. This might be a result of the installed Citrix product not being able to contact the
Terminal Services Licensing Server or the Citrix Licensing Server. If errors in the event log are
not reviewed, users might eventually be denied access because they cannot acquire a valid
license.
 Hardware Failure: Any event notification that relates to a hardware failure should be looked
at immediately. Any device that has failed will have an impact on the performance of the
system. At a minimum, a hardware failure will remove the redundancy of the component.
 Security Warnings: Customers should investigate security warnings or audit failure events
regarding failed logons in the security log. This could be an indication that someone is
attempting to compromise the servers.
 Disk Capacity: As the drives of a Windows system reach 90% of capacity, an event error
message will be generated. To ensure continuous service, customers should poll these event
errors. As the system runs out of hard disk space, the system is put at severe risk. The
server might not have enough space left to service the requests of users for temporary file
storage. As the space becomes used, customers should analyze and determine where space
is being used and clean up the drives.
 Application / Service errors: Any event notification that relates to application or services
errors should be investigated. All Citrix software components will leverage the Windows Event
Log for error logging. A list of the known Event Log warnings and errors issued by Citrix
components can be found at the following links:
o Event Codes Generated by PVS
o XenDesktop Broker Event Log Messages
o XenDesktop VDA Event Log Messages
o XenDesktop MCS Event Log Messages
o XenApp Critical Event Viewer Messages
o Web Interface - Logged Messages and Event IDs
It is important to periodically check the Event Viewer for Citrix related warnings or errors. Warnings or
errors that repeatedly appear in the logs should be investigated immediately, because it may indicate
a problem that could severely impact the Citrix environment if not properly resolved.
In multi-server environments it becomes easier to administer the servers when logs can be collected
and reviewed from a central location. Most enterprise grade monitoring solutions provide this
functionality. More sophisticated monitoring solutions enable an administrator to correlate event
information with other data points such as performance metrics or availability statistics. In case the
selected monitoring solution does not provide this functionality the Windows 2008 R2 EventLog
subscription feature can be used. This feature allows administrators to receive events from multiple
servers and view them from a designated collector computer. Please see Microsoft TechNet article on
how to set this up.
XenServer is also capable of sending its logs to a central syslog server. The administrator sets the IP
address of the syslog daemon server in the properties of each XenServer in the pool. This
configuration allows administrators to capture real-time activity across multiple XenServer hosts.
Further information can be found within the XenServer Admin Guide.

Growth
In addition to the day-to-day monitoring of system-level metrics, the following metrics should be
tracked from a historical perspective. Each of the metrics will help plan for future growth as more
users access the environment.
For example, the metric “Application Resolutions/sec” might increase as more users access the
environment and as more applications and desktops are available. If the metrics increases too much,
users will have to wait longer for delivery of their available resources. As the trends for the metrics
increases, additional system-level resources (CPU, RAM) can be allocated to keep the system running
at optimal efficiency.
Component Metric Description
System Counters Process – Private Bytes – ImaSrv The current size, in bytes of memory that this
process (IMA) has allocated and cannot be
shared with other processes.
XenApp Counters Application Resolutions/sec The number of resolutions (applicable launch
requests) per second.
XenApp Counters Application Enumerations/sec The number of non-XML-based enumerations
(requests for application lists) per second.
XenApp Counters Filtered Application
Enumerations/sec
The number of XML-based enumerations
(requests for application lists) per second.
XenApp Counters Terminal Services – Total Sessions Total number of Terminal services sessions.
XenApp Counters Citrix Licensing – Average License
Check-Out Response Time (ms)
Displays the average license check-out
response time in milliseconds. Please note
that network congestion can influence the
values reported.
XenDesktop Counters Citrix Broker Service – Brokered
Sessions
The number of virtual desktop sessions
brokered by the Citrix Broker Service.
XenDesktop Counters Citrix XML Service – Avg.
Transaction Time (Instances:
“Enumerated Resources”, and
“Launch *”
The time on average, in seconds, taken to
process an XML transaction in Citrix Broker
Service.
Web Interface Counters ASP.NET – Requests Current The current number of requests, both
executing and queued.
Web Interface Counters ASP.NET – Request Execution
Time
The amount of time (in milliseconds) that it
took to execute the most recent request.
Web Interface Counters Web Service – Concurrent
Connections
Current Connections is the current number of
connections established with the Web
service.
Table 76: Historical Metrics to Monitor

199
Glossary and Abbreviations
Bring your own Device (BYOD): A diverse set of initiatives aiming to enhance employee satisfaction
and morale by enabling the use of personal devices to supplement corporate endpoints. BYOD can
replace a corporate-owned device entirely. Whatever approach an organization chooses to take, a
complete, well thought-out policy is essential for embracing BYOD without increasing risk. The lack of a
structured BYOD policy leaves many organizations exposed to problems such as security gaps,
compliance issues and rising IT complexity.
Consumerization: A trend impacting businesses globally where a new generation of younger workers
that have seen technology as a natural part of daily life and are demanding its usage at work. Unlike prior
generations of employees who learned about technology primarily through the workplace, these users
enter the workforce already primed by consumer technology and commonly own and use consumer
products at least as sophisticated if not more so than the tools they‟re provided at work. These workers
are demanding rather than requesting the use of consumer technologies including hardware and
applications. Organizations are being forced to embrace this trend or risk alienating this new generation
of workers.
Desktop Virtualization: The concept of isolating then separating the operating system from the
underlying client hardware used to access it. Several models exist, ranging from hosted virtual desktop
infrastructure to client side hypervisor virtualization.

200
Baseline Policy Reference
The following sample outlines the initial policy configurations recommended for various scenarios,
including baseline configuration, network, security, mobile device and profile related policy
considerations. Each policy configuration may contain the following policy settings:
Policy Settings
Enabled - Enables the setting. Where applicable, specific settings are
detailed.
Disabled – Disables the setting
Note: Disabling the policy overrides lower priority policies settings.
Allow – Allows the action controlled by the setting. Where applicable,
specific settings are detailed.
Prohibit – Prohibits the action controlled by the setting
Note: Prohibiting a feature or functionality overrides lower priority policies
settings.
Not Configured – Unless specifically set, un-configured policies use default
settings.
Table 77: Policy Configuration Settings

201
Citrix User Policy Settings
User Policy Setting XA XD Baseline Low Security High Security LAN
Speed
WAN
Speed
Tablet Profile
ICA
Client clipboard redirection X X Allow Prohibit
Desktop launches X Disable
Launching of non-published programs
during client connection
X Disable
ICAAdobe Flash DeliveryFlash
Redirection
Flash acceleration X X Enabled Disable
Flash default behavior X X Enable Flash
Acceleration
Disable Flash
Acceleration
Flash event logging X X Enabled
Flash intelligent fallback X X Enabled
Flash latency threshold X X 30 milliseconds
ICAAdobe Flash DeliveryLegacy
Server Side Optimization
Flash quality adjustment X Allow
ICAAudio
Audio over UDP Real-time Transport X Enabled Enabled
Audio Plug N Play X Allow
Audio quality X X Medium
Client audio redirection X X Allow
Client microphone redirection X X Prohibit
ICA Client Sensors Location

202
Allow applications to use the physical
locations of the client device
X Enable if
secure
connection
ICADesktop UI
Aero Redirection X Allow Prohibit
Aero Redirection Graphics Quality X High Disable
Desktop wallpaper X X Enable Disable
Menu animation X X Allow Prohibit
View window contents while dragging X X Allow Prohibit
ICAFile Redirection
Auto connect client drives X X Allow Prohibit
Client fixed drives X X Enable Disable
Client floppy drives X X Allow Prohibit
Client network drives X X Allow Prohibit
Client optical drives X X Allow Prohibit
Client removable drives X X Allow Prohibit
Host to client redirection X Disable
Preserve client drive letters X Disable
Read-only client drive access X X Disable
Use asynchronous writes X X Disable Enable
ICA Mobile Experience
Automatic Keyboard Display X Enable
Launch touch-optimized desktop X Enable
Remote the combo box X Enable
ICA Multi Stream Connections
Multi-Stream X X Enable
with QoS
Enable
with QoS
ICAPort Redirection

203
Auto connect client COM ports X X Disable
Auto connect client LPT ports X X Disable
Client COM port redirection X X Disable
Client LPT port redirection X X Disable
ICAPrinting
Client printer redirection X X Allow Prohibit
Default printer X X Set to client‟s
main printer
Direct connections to print servers X X Enable Disable
Printer auto creation log preference X X Errors
Wait for printers to be created (desktop) X X Disabled
ICAPrintingClient Printers
Auto-create client printers X X Default printer
only
Auto-generate generic universal driver X X Disabled
Client printer names X X Standard names
Printer properties retention X X Retained in
profile only
Retained and restored client printers X X Allowed
ICAPrintingDrivers
Automatic installation of in-box printer
drivers
X X Disabled
Universal driver usage X X Use Universal
Printing only if
requested driver
is unavailable
ICAPrintingUniversal Printing
Universal printing EMF processing
mode
X X Spool to printer

204
Universal printing image compression
limit
X X Best Quality
Universal printing optimization defaults X X Best
Quality
Standard
Quality
Universal printing preview preference X X Use for auto-
generated and
generic
ICASecureICA
SecureICA minimum encryption level X RCS 128 bit
Logon only
RCS 128 bit
ICASession Limits
Disconnected session timer X Disabled Enabled
Disconnected session timer interval X 30 Minutes
Linger Disconnect Timer Interval X 5 Minutes
Linger Terminate Timer Interval X 10 Minutes
Pre-Launch Disconnect Timer Interval X 15 Minutes
Pre-Launch Terminate Timer Interval X 30 Minutes
Session connection timer X Disabled
Session idle timer X Disabled Enabled
Session idle timer interval X 2 hours
ICAShadowing
Input from shadow connections X Allow Prohibit
Log shadow attempts X Allow
Notify user of pending shadow
connections
X Allow
Users who can shadow other users X Defined by
security
ICATime Zone Control

205
Estimate local time for legacy clients X Enable
Use local time of client X X Use Client time
zone
ICATWAIN devices
Client TWAIN device redirection X X Allow Prohibit
TWAIN compression level X X Low High
ICAUSB devices
Client USB device redirection X X Enable Disable
Client USB device redirection rules X X Allow Prohibit
Client USB Plug and Play device
redirection
X Allow Prohibit
ICAVirtual Desktop Agent
SettingsICA Latency Monitoring
Enable Monitoring X Disabled
ICA Virtual Desktop Agent Settings
Profile Load Time Monitoring
ICAVisual Display
Max Frames per Second X X 30 15
ICAVisual DisplayMoving Images
Moving Image Compression X X Enabled
Minimum Image Quality X Very
High
Low
Target Minimum Frame Rate X 24 10
ICAVisual DisplayStill Images
Extra Color Compression X X Disable Enabled
Extra Color Compression Threshold X X 8192
kbps
8192
kbps
Lossy compression level X X Low High

206
Lossy compression level threshold
value
X X Unlimited Unlimited
Profile Management
Enable Profile Management X X Enabled
Process Groups X X Configure
groups
Path to User Store X X UNC Path
Active Write Back X X Enabled
(Persistent
desktops)
Process logons of local administrators X X Enabled
Profile Management Advanced
Settings
Delete Redirected Folders X X Enabled
Directory of MFT Cache Files X X Local or
persistent
location
Process Internet cookie files on logoff X X Enabled
Profile Management File System
Exclusion list – directories X X Exclude
redirected
folders
Synchronization
Directories to Synchronize X X Exclude
directories
Files to Synchronize X X Selected
files
Folders to Mirror X X Selected
folders
Profile Management Profile handling

207
Table 78: Citrix User Policy Settings
Local profile conflict handling X X Delete local
profile
Migration of existing profiles X X None
Profile Management Profile
Streamed user profiles
Profile Streaming X X Enable if
large profile
Server Session Settings
Session importance X Normal
Single Sign-on X X Disabled

208
List of excluded files for Profile Management
 AppDataLocal
 AppDataLocalLow
 AppDataRoamingCitrixPNAgentAppCache
 AppDataRoamingCitrixPNAgentIcon Cache
 AppDataRoamingCitrixPNAgentResourceCache
 AppDataRoamingICAClientCache
 AppDataRoamingMicrosoftWindowsStart Menu
 AppDataRoamingSunJavaDeploymentcache
 AppDataRoamingSunJavaDeploymentlog
 AppDataRoamingSunJavaDeploymenttmp
 Application Data
 Citrix
 Contacts
 Desktop
 Documents
 Favorites
 Java
 Links
 Local Settings
 Music
 My Documents
 My Pictures
 My Videos
 Pictures
 UserData
 Videos
 AppDataRoamingMacromediaFlash
Playermacromedia.comsupportflashplayersys
 AppDataRoamingMacromediaFlash Player#SharedObject
 AppDataRoaming
 Downloads
 Saved Games
 Searches
Synchronized Directories
 AppDataRoamingMicrosoftCredentials
 AppDataRoamingMicrosoftCrypto
 AppDataRoamingMicrosoftProtect
 AppDataRoamingMicrosoftSystemCertificates
 AppDataLocalMicrosoftCredential

209
Synchronized Files
Example Synchronized Files for Microsoft Outlook and Google Earth
 AppDataLocalMicrosoftOffice*.qat
 AppDataLocalMicrosoftOffice*.officeUI
 AppDataLocalLowGoogleGoogleEarth*.kml
Mirrored Folders
 AppDataRoamingMicrosoftWindowsCookies

210
Citrix Computer Policy Settings
Computer Policy Setting X
A
X
D
Baseline Low
Security
High
Security
LAN
Speed
WAN
Speed
Tablet Profile
ICA
ICA listener connection timeout X X 120000 ms
ICA listener port number X X 1494
ICA Auto Client Reconnect
Auto client reconnect X X Allow Prohibit
Auto client reconnect authentication X Not required Required
Auto client reconnect logging X X Disabled Enabled
ICA End User Monitoring
ICA round trip calculation X X Enable
ICA round trip calculations for idle
connections
X X Disable
ICA Graphics
Display memory limit X X 32768 KB
Display mode degrade preference X X Degrade Color
Depth First
Dynamic Windows preview X X Enabled
Image caching X X Enabled
Maximum allowed color depth X 32 bit
Notify user when display mode is degraded X X Disabled
Queuing and tossing X X Enabled
ICAGraphicsCaching
Persistent Cache Threshold X X 3000000 Kbps
ICA Keep Alive
ICA keep alive timeout X X 60 seconds

211
Table 79: Citrix Computer Policy Settings
ICA keep alives X X Enabled
ICA Multimedia
Windows Media Redirection X X Allowed
Windows Media Redirection Buffer Size X X 10 seconds
Windows Media Redirection Buffer Size Use X X Enabled
ICA Multistream Connections
Multistream X X Enabled
(QoS)
Enable
d
(QoS)
ICA Session Reliability
Session reliability connections X X Allow
ICA Virtual Desktop Agent Settings CPU
Usage Monitoring
ICA Shadowing
Shadowing X Allow
Licensing
License server host name X License Server
Name
License server port X 27000

212
Microsoft Windows Policy
User Policy
Policy Path Setting Description Applies to
Control Panel Prohibit Access to the Control Panel Enable Disables all control panel programs XenApp, XenDesktop
Control Panel Personalization Enable screen
saver
Enable Enables the use of a Screen Saver XenApp, XenDesktop
Control Panel Personalization Force specific
screen saver
Enable
scrnsave.scr
Forces the use of the “blank” screen saver in Windows XenApp, XenDesktop
Control Panel Personalization Password protect
the screen saver
Enabled Forces password protection on the screen saver XenApp, XenDesktop
Control Panel Personalization Screen saver
timeout
Enabled
X Minutes (default
15)
Sets the amount of time in minutes that elapse before the
screen saver is activated
XenApp (Published
Desktop), XenDesktop
Desktop Don‟t save settings on exit Enabled Prevents users from changing some desktop configurations
such as the size of the taskbar or the position of open
windows on exit.
XenApp
Desktop Hide Network Locations icon on desktop Enabled Removes the Network Locations icon from the desktop. XenApp
Desktop Prohibit user from manually redirecting
Profile Folders
Enabled Prevents users from manually changing the path to their
profile folders.
XenApp, XenDesktop
Desktop Remove Recycle Bin icon from desktop Enabled Removes most occurrences of the Recycle Bin icon. XenApp, XenDesktop
Start Menu and Taskbar Change Start Menu
power button
Enabled
Log Off
Set Start Menu power button functionality to Log Off user. XenApp, XenDesktop
Start Menu and Taskbar Prevent changes to
Taskbar and Start Menu settings
Enabled Removes the Taskbar and Start Menu settings from Settings
on the Start Menu.
XenApp
Start Menu and Taskbar Remove and prevent
access to the Shut Down, Restart, Sleep and
Hibernate commands
Enabled Prevents user from performing these commands from the
Start Menu or the Windows Security screen.
XenApp
Start Menu and Taskbar Remove links and access
to Windows Update
Enabled Prevents users from connecting to the Windows Update
website.
XenApp, XenDesktop
Start Menu and Taskbar Remove network icon
from the Start Menu
Enabled Removes the network icon from the Start Menu XenApp, XenDesktop

213
Start Menu and Taskbar Remove Run menu from
the Start Menu
Enabled Removes the Run command from the Start Menu, Internet
Explorer, and Task Manager
XenApp
System Prevent access to registry editing tools Enabled Disables the Windows Registry Editor XenApp, XenDesktop
System Prevent access to the Command Prompt Enabled Prevents users from running the interactive command
prompt “cmd.exe”
XenApp
System Ctrl+Alt+Del Options Remove Task
Manager
Enabled Prevents users from starting Task Manager XenApp
System Folder Redirection Do not automatically
make redirected folders available offline
Enabled Prohibits redirected shell folders Contacts, Documents,
Desktop, Favorites, Music, Pictures, Videos, Start Menu and
AppDataRoaming from being available offline
XenApp, XenDesktop
System User Profiles Exclude Directories in
Roaming Profile
Citrix, Contacts,
Desktop,
Downloads,
Favorites, Links,
Documents,
Pictures, Videos,
Music, Saved
Games, Searches
Excludes the specified directories from the Roaming Profile XenApp, XenDesktop
Windows Components Windows Update Remove
access to use all Windows Update features
Enabled Removes all Windows Update functions XenApp, XenDesktop
Windows Explorer Do not move deleted files to the
Recycle Bin
Enabled Prohibits deleted files from being placed in the Recycle Bin.
All files are permanently deleted.
XenApp, XenDesktop
Windows Explorer Hide these specified drives in
My Computer
Enabled
Local hard drives
Hides local hard drives from My Computer XenApp
Windows Explorer Prevent access to drives from
My Computer
Enabled
Local hard drives
Prevents access to local hard drives from My Computer XenApp
Table 80: Microsoft User Policy Settings

214
Machine Policy
Policy Path Setting Description Applies to
Internet Communication settings Turn off Windows
Customer Improvement Program
Enabled Turns off the Windows Customer Improvement Program for
all users
XenApp, XenDesktop
System Group Policy User Group Policy loopback
processing mode
Merge or Replace Applies alternate user settings when a user logs on to a
computer affected by this setting
XenApp, XenDesktop
System Power Management Select an active
power plan
High Performance Specifies a power plan from a list of available plans. XenApp, XenDesktop
System System Restore Turn off System Restore Enabled Turns off Windows System Restore features XenApp, XenDesktop
System User Profiles Do not check for user
ownership of Roaming Profile folders
Enabled Disables security check for roaming profile folders XenApp, XenDesktop
Windows Components AutoPlay Policies Turn off
AutoPlay
Enabled Turns off AutoPlay for removable devices. XenApp
Windows Components Internet Explorer Turn off
reopen last browsing session
Enabled Disables ability to reopen the user‟s last browsing session XenApp
Windows Components Remote Desktop Services
RD Licensing License server security group
XenApp server
security groups
Specifies the servers to which RDS will provide licenses XenApp
Remote Desktop Session Host Licensing Set the
Remote Desktop licensing mode
Per User or Per
Device
Specifies the licensing mode used by Remote Desktop
Server
XenApp
Remote Desktop Session Host Licensing Use the
specified Remote Desktop license servers
Specified servers Specifies the preferred license servers for Remote Desktop
Services
XenApp
Windows Components Windows Update
Configure Automatic Updates
Disabled Specifies whether the computer system will receive
automatic updates through the Windows Update process.
XenApp, XenDesktop
Table 81: Microsoft Machine Policy Settings

215
Folder Redirection Policy
User PolicyWindows SettingsSecurity SettingsFolder Redirection
Folder Setting Options
AppData (Roaming) Basic Grant User Exclusive Rights: Disabled
Move Contents to new location: Enabled
Apply Policy to Windows 2000, Windows XP, Windows 2003: Disabled
Policy Removal Behavior: Leave Contents
Contacts Basic Grant User Exclusive Rights: Disabled
Desktop Basic Grant User Exclusive Rights: Disabled
Documents Basic Grant User Exclusive Rights: Disabled
Move Contents to new location: Disabled
Downloads Basic Grant User Exclusive Rights: Disabled
Favorites Basic Grant User Exclusive Rights: Disabled
Links Basic Grant User Exclusive Rights: Disabled
Music Follow the
Documents Folder
Pictures Follow the
Documents Folder

216
Saved Games Basic Grant User Exclusive Rights: Disabled
Searches Basic Grant User Exclusive Rights: Disabled
Start Menu Basic Grant User Exclusive Rights: Disabled
Videos Follow the
Documents Folder
Table 82: Folder Redirection Settings

Desktop Virtualization Design
217
Policy Quick Reference
The following table provides a description for all Citrix policy settings contained in this document. For complete and up-to-date policy settings, consult
the policy settings references sections for the various technologies in Citrix eDocs.
User Policy Reference
User Policy
Policy Group Policy Description Applies to
ICA
Client clipboard redirection Allow or prevent the clipboard on the client device to be mapped to the clipboard on the
server.
XA 6, XD 5
Desktop launches When allowed, non-administrative users can connect. XA 6 RDS only
Launching of non-published programs
during client connection
Specifies whether to launch initial applications or published applications on the server. XA 6
ICAAdobe Flash DeliveryFlash Redirection
Flash acceleration Enables or disables, in Legacy mode only, Flash content rendering on client devices
instead of the server.
XA 6, XD 5
Flash backwards compatibility Enabling Flash backwards compatibility allows earlier versions of Citrix Receiver to
work with legacy Flash Redirection features
XA 6.5, XD 5.5
Flash default behavior Establishes the default behavior of second generation Flash acceleration. XA 6.5, XD 5.5
Flash event logging Allows Flash events to be recorded in the Windows application event log. XA 6, XD 5
Flash intelligent fallback If enabled, the system attempts to automatically revert to server-side rendering for
Flash Player instances for which client-side rendering is unnecessary or would provide
a poor experience
XA 6.5, XD 5.5
Flash latency threshold Maximum latency threshold for Flash redirection. Only applies to Legacy mode
features. Flash backwards compatibility must be enabled.
XA 6, XD 5
ICAAdobe Flash DeliveryLegacy Server Side Optimizations
Flash quality adjustment Adjusts quality of Flash content rendered on session hosts to improve performance. XA 6
ICA Audio

218
Audio over UDP Real-time Transport Allows transmission of audio between host and client over Real-time Transport
Protocol (RTP) using the user datagram protocol (UDP).
XD 5.5
Audio Plug N Play Allows the use of multiple audio devices. XA 6
Audio quality Specify the sound quality as low, medium, or high.
Select "Medium - optimized for speech" for delivering Voice over IP applications. Audio
sent to the client is compressed up to 64Kbps.
XA 6, XD 5
Client audio redirection Allows or prevents applications hosted on the server to play sounds through a sound
device installed on the client computer. Also allows or prevents users to record audio
input.
XA 6, XD 5
Client microphone redirection Enables or disables client microphone redirection. XA 6, XD 5
ICA Client Sensors Location
Allow applications to use the physical
locations of the client device
Enables or disables the ability for applications to use the physical location of the client
device.
XA 6.5 FP1
ICA Desktop UI
Aero Redirection Allow the redirection of Aero commands from VDA to client to enrich user experience. XD 5.5
Aero Redirection Graphics Quality Determine the quality of graphics for Aero Redirection. XD 5.5
Desktop wallpaper Enables or disables the desktop wallpaper in user sessions. XA 6, XD 5
Menu animation Allows or prevents menu animation. XA 6, XD 5
View window contents while dragging Controls the display of window content when dragging a window across the screen. XA 6, XD 5
ICA File Redirection
Auto connect client drives Allows or prevents automatic connection of client drives when users log on. XA 6, XD 5
Client drive redirection Enables or disables file (drive) redirection to and from the client. XA 6, XD 5
Client floppy drives Allows or prevents users from accessing or saving files to floppy drives on the client
device.
XA 6, XD 5
Client fixed drives Allows or prevents users from accessing or saving files to fixed drives on the user
device.
XA 6, XD 5
Client network drives Allows or prevents users from accessing and saving files to client network (remote)
drives.
XA 6, XD 5
Client optical drives Allows or prevents users from accessing or saving files to CD-ROM, DVD-ROM, and
BD-ROM drives on the client device.
XA 6, XD 5

219
Client removable drives Allows or prevents users from accessing or saving files to removable drives on the user
device.
XA 6, XD 5
Host to client redirection Enables or disables file type associations for URLs and some media content to be
opened on the client device.
XA 6
Preserve client drive letters Enables or disables preservation of client drive letters. XD 5
Read-only client drive access When enabled, files/folders on mapped client drives can only be accessed in read-only
mode.
When disabled, files/folders on mapped client drives can be accessed in regular
read/write mode.
XA 6.5, XD 5.5
SecureICA minimum encryption level Specifies the minimum level at which to encrypt session data sent between the server
and a client device.
XA 6
Use asynchronous writes Enables or disables asynchronous disk writes. XA 6, XD 5
ICA Mobile Experience
Automatic Keyboard Display Enables or disables the automatic display of the soft keyboard on mobile devices. XA 6.5 FP1, XD 5.6 FP1
Launch touch-optimized desktop Enables or disables the launching of a touch-optimized desktop for mobile clients. XA 6.5 FP1, XD 5.6 FP1
Remote the combo box Enables or disables the remoting of the combo box on mobile devices. XA 6.5 FP1, XD 5.6 FP1
Multi-Stream Enables or disables the Multi-Stream feature for specified users. XA 6.5, XD 5.5
ICA Port Redirection
Auto connect client COM ports When enabled, COM ports from the client are automatically connected. XA 6, XD 5
Auto connect client LPT ports When enabled, LPT ports from the client are automatically connected. XA 6, XD 5
Client COM port redirection When enabled, COM port redirection to and from the client is allowed. XA 6, XD 5
Client LPT port redirection When enabled, LPT port redirection to the client is allowed. XA 6, XD 5
ICA Printing
Client printer redirection Allows or prevents client printers to be mapped to a server when a user logs on to a
session.
XA 6, XD 5
Default printer Specifies how the client‟s default printer is established in an ICA session. XA 6, XD 5
Printer auto-creation event log
preference
Specifies which events are logged during the printer auto-creation process. You can
choose to log no errors or warnings, only errors, or errors and warnings.
XA 6, XD 5

220
Wait for printers to be created (desktop) Allows or prevents a delay in connecting to a session so that desktop printers can be
auto-created.
XA 6, XD 5
ICA Printing Client Printers
Auto-create client printers Specifies which client printers are auto-created. XA 6, XD 5
Auto-create generic universal printer Enables or disables auto-creation of the Citrix UNIVERSAL Printer generic printing
object for sessions with a UPD capable client.
XA 6, XD 5
Client printer names Selects the naming convention for auto-created client printers. XA 6, XD 5
Direct connections to print servers Enables or disables direct connections from the host to a print server for client printers
hosted on an accessible network share.
XA 6, XD 5
Printer properties retention Specifies whether and where to store printer properties. XA 6, XD 5
Retained and restored client printers Enables or disables the retention and re-creation of client printers. XA 6, XD 5
ICA Printing Drivers
Automatic installation of in-box printer
drivers
Enables or disables the automatic installation of printer drivers from the Windows in-
box driver set or from driver packages which have been staged onto the host using
"pnputil.exe /a".
XA 6, XD 5
Universal driver usage Specifies when to use universal printing. Universal printing employs generic printer
drivers instead of standard model-specific drivers potentially simplifying burden of
driver management on host machines.
XA 6, XD 5
ICA Printing Universal Printing
Universal printing EMF processing mode Controls the method of processing the EMF spool file on the Windows client machine. XA 6, XD 5
Universal printing image compression
limit
Defines the maximum quality and the minimum compression level available for images
printed with the Universal Printer driver.
XA 6, XD 5
Universal printing optimization defaults Specifies the default settings for the Universal Printer when it is created for a session. XA 6, XD 5
Universal printing preview preference Specifies whether to use the print preview function for auto-created or generic
universal printers.
XA 6, XD 5
ICA Security

221
SecureICA minimum encryption level Specifies the minimum level at which to encrypt session data sent between the server
and a client device.
XA 6
ICA Session Limits
Disconnected session timer Enables or disables a timer to determine how long a disconnected, locked workstation
can remain locked before the session is logged off.
XD 5
Disconnected session timer interval Determines how long, in minutes, a disconnected, locked workstation can remain
locked before the session is logged off.
XD 5
Linger Disconnect Timer Interval Disconnects an existing session the specified number of minutes after the last
application exits.
XA 6.5
Linger Terminate Timer Interval Terminates an existing session the specified number of minutes after the last
application exits.
XA 6.5
Pre-launch Disconnect Timer Interval Disconnects an existing Pre-launch session after the specified number of minutes. XA 6.5
Pre-launch Terminate Timer Interval Terminates an existing Pre-launch session after the specified number of minutes. XA 6.5
Session connection timer Enables or disables a timer to determine the maximum duration of an uninterrupted
connection between a user device and a workstation.
XD 5
Session idle timer Enables or disables a timer to determine how long an uninterrupted user device
connection to a workstation will be maintained if there is no input from the user.
XD 5
Session idle timer interval Determines, in minutes, how long an uninterrupted user device connection to a
workstation will be maintained if there is no input from the user.
XD 5
ICA Shadowing
Input from shadow connections Allows or prevents shadowing users to take control of the keyboard and mouse of the
user being shadowed during a shadowing session.
XA 6
Log shadow attempts Allows or prevents recording of attempted shadowing sessions in the Windows event
log.
XA 6
Notify user of pending shadow
connections
Allows or prevents shadowed users to receive notification of shadowing requests from
other users.
XA 6
Users who can shadow other users Specifies the users who can shadow other users. XA 6
ICA Time Zone Control
Estimate local time for legacy clients Enables or disables estimating the local time zone of client devices that send
inaccurate time zone information to the server.
XA 6

222
Use local time of client Determines the time zone setting of the user session. XA 6, XD 5
ICA TWAIN devices
Client TWAIN device redirection Allows or prevents users to access TWAIN devices, such as digital cameras or
scanners, on the client device from published image processing applications.
XA 6, XD 5.5
TWAIN compression level Specifies the level of compression of image transfers from client to server. XA 6, XD 5.5
ICA USB devices
Client USB device redirection Enables or disables redirection of USB devices to and from the client (workstation
hosts only).
XA 6 VM Hosted Apps, XD 5
Client USB device redirection rules Lists redirection rules for USB devices. XA 6 VM Hosted Apps, XD 5
Client USB Plug and Play device
redirection
Allows or prevents plug-n-play devices such as cameras or point-of-sale (POS) devices
to be used in a client session.
XA 6 Terminal Server
ICA Visual Display
Max Frames per Second Sets the maximum number of frames per second that the virtual desktop will send to
the client.
XA 6, XD 5
ICA Visual Display Moving Images
Minimum Image Quality Adaptive Display JPEG Quality Floor. XD 5.5
Moving Image Compression Enables Adaptive Display. XA 6.5 (with hotfix
XA650W2K8R2X64011), XD 5.5
Target Minimum Frame Rate The system will try its best to maintain this many frames per second when bandwidth is
low.
XD 5.5
ICA Visual Display Still Images
Extra Color Compression Extra color compression improves responsiveness over low bandwidth connections at
the expense of image quality.
XA 6.5, XD 5
Extra Color Compression Threshold Threshold at which Extra Color Compression is applied. XA 6.5, XD 5
Lossy compression level Degree of lossy compression used on images. XA 6, XD 5
Lossy compression threshold value The maximum bandwidth in kilobits per second for a connection to which lossy
compression is applied.
XA 6, XD 5
Server Session Settings
Session importance Specifies the importance level at which a session is run. XA 6

223
Single Sign-On Enables or disables the use of Single Sign-On when users connect to servers or
published applications in a XenApp farm.
XA 6.5, XD 5.5
ICA Virtual Desktop Agent Settings ICA Latency Monitoring
Enable Monitoring Enable or disable ICA Latency monitoring. XD 5.5
Monitoring Period Period of time, in seconds, during which the moving average for ICA Latency is
calculated.
XD 5.5
Threshold Threshold, in milliseconds, that triggers a High Latency condition, displayed in Desktop
Studio and Desktop Director.
XD 5.5
ICA Virtual Desktop Agent Settings Profile Load Time Monitoring
Enable Monitoring Enable or disable Profile load time monitoring. XD 5.5
Threshold Threshold, in seconds, that triggers a High Profile Load Time condition, displayed in
Desktop Studio and Desktop Director.
XD 5.5
Table 83: Citrix User Policy Reference Guide
Computer Policy Reference
Computer Policy
Policy Group Policy Description Applies to
ICA
ICA listener connection timeout Maximum wait time for a connection using the ICA protocol to be completed. XA 6 (VM Hosted Apps), XD 5
ICA listener port number The TCP/IP port number used by the ICA protocol on the server. XA 6 (VM Hosted Apps), XD 5
ICA Auto Client Reconnect
Auto client reconnect Allows or prevents automatic reconnection by the same client after a connection has
been interrupted.
XA 6, XD 5
Auto client reconnect authentication Requires authentication for automatic client reconnections. XD 5
Auto client reconnect logging Records or prevents recording auto client reconnections in the event log. XA 6, XD 5
ICA End User Monitoring
ICA round trip calculation Enables or disables the calculation of ICA round trip measurements. XA 6, XD 5

224
ICA round trip calculations for idle
connections
Determines whether ICA round trip calculations are performed for idle connections. XA 6, XD 5
ICA Graphics
Display memory limit Specifies the maximum video buffer size in kilobytes for the session. XA 6, XD 5
Display mode degrade preference Degrades either color depth or resolution first when the session display memory limit
is reached.
XA 6, XD 5
Dynamic Windows preview Dynamic Windows preview enables the state of seamless windows to be seen on the
various windows previews (Flip, Flip 3D, Taskbar Preview, and Peek).
XA 6.5, XD 5.5
Image caching Cache image to make scrolling smoother XA 6, XD 5
Maximum allowed color depth Specifies the maximum color depth allowed for a session. XA 6
Notify user when display mode is
degraded
Displays a popup with an explanation to the user when the color depth or resolution is
degraded.
XA 6, XD 5
Queuing and tossing Discards queued images that are replaced by another image. XA 6, XD 5
ICAGraphicsCaching
Persistent Cache Threshold Caches bitmaps on the client disk. XA 6, XD 5
ICA Keep Alive
ICA keep alive timeout Seconds between successive ICA keep-alive messages. XA 6, XD 5
ICA keep alives Sends or prevents sending ICA keep-alive messages periodically. XA 6, XD 5
ICA Multimedia
Windows Media Redirection Controls and optimizes the way XenApp servers deliver streaming audio and video to
users.
XA 6, XD 5
Windows Media Redirection Buffer Size Specify a buffer size from 1 to 10 seconds for Windows Media Redirection. XA 6, XD 5
Windows Media Redirection Buffer Size
Use
If this setting is enabled, the system will use the buffer size specified in the "Windows
Media Redirection Buffer Size" setting.
XA 6, XD 5
Multi-Stream Enables or disables the Multi-Stream feature on the server. By default, Multi-Stream is
disabled. This policy need not be enabled when using branch repeater that supports
Multi-Stream. Enable this policy when using 3rd party routers or legacy branch
repeaters to achieve desired QoS. Restart the server for the changes to take effect.
XA 6.5, XD 5.5

225
ICA Session Reliability
Session reliability connections Allow or prevent session reliability connections. XA 6, XD 5
ICA Virtual Desktop Agent Settings CPU Usage Monitoring
Enable Monitoring Enable or disable CPU usage monitoring. XD 5.5
Monitoring Period Period of time, in seconds, during which the moving average for CPU usage is
calculated.
XD 5.5
Threshold Threshold, as a percentage, that triggers a High CPU condition, displayed in Desktop
Studio and Desktop Director.
XD 5.5
ICA Shadowing
Shadowing Allow shadowing of ICA sessions XA 6
Licensing
License server host name The name of the server hosting XenApp licenses. XA 6
License server port The port number of the server hosting XenApp licenses. XA 6
Profile Management
Enable Profile Management Turns on Citrix Profile Management UPM 2.0
Process Groups Active Directory groups that will use Citrix Profile Management UPM 2.0
Path to User Store Network location of end-user profile store UPM 2.0
Active Write Back Files and folders (but not registry keys) will be synchronized as they are modified. UPM 3.0
Process logons of local administrators Process the profile of a user who is a local administrator on a system. UPM 2.0
Profile Management Advanced Settings
Delete Redirected Folders Folder is deleted from the local profile when the user next logs on. UPM 3.2
Directory of MFT Cache Files Identifies the location for the MFT Cache file. The MFT cache file should be saved in
a persistent, easily accessible location for best performance
UPM 2.0
Process Internet cookie files on logoff Stale Internet cookie files are removed on user logoff UPM
Exclusion list – directories Identifies what directories to exclude from the user profile UPM 2.0
Profile Management File System Synchronization

226
Directories to Synchronize Identifies which directories should be synchronized from the system to the profile on
logoff.
UPM 2.0
Files to Synchronize Identifies specific files, which should be synchronized from the system to the profile on
logoff.
UPM 2.0
Folders to Mirror Mirroring folders allows Profile management to process a transactional folder and its
contents as a single entity, thereby avoiding profile bloat.
UPM 3.1
Profile Management Profile handling
Local profile conflict handling Identifies how UPM handles conflicts between Windows local profiles and Citrix
profiles.
UPM 2.0
Migration of existing profiles Determines which types of existing user profiles to migrate. UPM 2.0
Profile Management Profile Streamed user profiles
Profile Streaming Enables streaming of profiles as files are requested. UPM 3.0
Table 84: Citrix Machine Policy Reference Guide

Revision History
Revision Change Description Updated By Date
1.0 Assess Released Andy Baker (Architect)
Daniel Feller (Lead Architect)
February 5, 2013
1.1 Design (Part 1) Released Andy Baker (Architect)
March 31, 2013
1.2 Maintain phase released Daniel Feller (Lead Architect) May 13, 2013
1.3 Printing and User Profiles added Thomas Berger (Architect)
June 3, 2013
1.4 User Policies, Computer Policies
and Personal vDisk added
Rich Meesters (Architect)
Adeel Arshed (Principal Consultant)
June 7, 2013
1.5 XenDesktop Sites, XenApp
Farms, Active Directory, Images
Thomas Berger (Architect)
Rich Meesters (Architect)
Andy Baker (Architect)
June 26, 2013
1.6 StoreFront 2.0, Hyper-V,
Database
Thomas Berger (Architect)
Ed Duncan (Senior Consultant)
August 30
th
, 2013
Authors
The creation of the handbook is a time consuming process and requires real deployment experience
across many scenarios. Citrix Worldwide Consulting Solutions would like to thank all of the individuals
that contributed to the Citrix Virtual Desktop Handbook:
 Daniel Feller
 Andy Baker
 Thomas Berger
 Rich Meesters
 Matthew Brooks
 Adeel Arshed
 Roger LaMarca
 Ed Duncan
About Citrix
Citrix Systems, Inc. (NASDAQ:CTXS) is a leading provider of virtual computing solutions that help
companies deliver IT as an on-demand service. Founded in 1989, Citrix combines virtualization,
networking, and cloud computing technologies into a full portfolio of products that enable virtual
workstyles for users and virtual datacenters for IT. More than 230,000 organizations worldwide rely on
Citrix to help them build simpler and more cost-effective IT environments. Citrix partners with over 10,000
companies in more than 100 countries. Annual revenue in 2011 was $2.20 billion.
©2012 Citrix Systems, Inc. All rights reserved. Citrix®, Access Gateway™, Branch Repeater™, Citrix
Repeater™, HDX™, XenServer™, XenApp™, XenDesktop™ and Citrix Delivery Center™ are
trademarks of Citrix Systems, Inc. and/or one or more of its subsidiaries, and may be registered in the
United States Patent and Trademark Office and in other countries. All other trademarks and registered
trademarks are property of their respective owners.

Citrix virtual desktop handbook (5 x)

More Related Content

What's hot (19)

Similar to Citrix virtual desktop handbook (5 x) (20)

More from Nuno Alves (20)

Recently uploaded (20)

Citrix virtual desktop handbook (5 x)