Wide Area Network (WAN) Bandwidth / Application Impact Analysis

A BSC Systems, Inc. White Paper

12007 Sunrise Valley Drive
Suite 300
Reston, VA 20191
703.391.1200
www.bscsys.com

Wide Area Network (WAN)
Bandwidth / Application Impact Analysis

Developed for

Department of Education
Federal Student Aid
Union Center Plaza
830 First Street, NE
Washington, DC 20202-5409

May 2006

BSC Systems, Inc. White Paper:
WAN Bandwidth /
Application Impact Analysis

Contents Introduction
At the request of the Department of
Introduction 2 Education, Federal Student Aid, the
Purpose 2 Common Services for Borrowers (CSB)
Independent Verification and Validation
Background 3
(IV&V) contractor, BSC Systems, Inc. has
Network Capacity Planning Concepts 4 developed this white paper to provide
Quality of Service Concepts 4 information concerning the Wide Area
Application Traffic Management 5 Network Bandwidth (WAN) / Application
Impact Analysis. This white paper was
Bandwidth / Application Impact Analysis
discussed with Cathy Power,
Process 6 ISCG/Production Operations for Borrower
Conclusion 11 Services at Federal Student Aid following a
Appendix 12 monthly Data / Telecommunication meeting
between Federal Student Aid, the developer
The 7 Layers of the OSI Model 12
of CSB (ACS-Educational Solutions, LLC),
and the Federal Student Aid WAN / Virtual
Data Center Provider (CSC, Incorporated).
During this meeting, some questions arose
concerning the bandwidth requirements for a
new sub-system of the CSB Application that
was scheduled to be deployed. As a result
of that meeting, it was clear there was some
confusion and misunderstanding as to what
the Federal Student Aid WAN / Virtual Data
Center Provider was looking for to assist in
provision of the WAN. This is the basis for
the development of a high-level and non-
technical white paper to discuss Bandwidth
and Application Impact Analysis and to
explain the core concepts in order to
facilitate a better understanding.

Purpose
The purpose of this white paper is to provide
the reader with an understanding of what
WAN Bandwidth / Application Impact
Analysis is, a generic process for conducting
the analysis, and the potential benefits from
doing the analysis prior to deploying a
system or application that will transverse the
WAN.

May 2006 Page 2 of 16

WAN Bandwidth /

This white paper DOES NOT make recommendations about the various tools and
applications that can assist in performing the types of analysis outlined in this document.
In some cases this document will discuss specific tools, and in those cases the tools being
discussed and referenced by this white paper are an example only.

Background
Bandwidth / Application Impact analysis is one of the techniques used for provisioning
and network capacity planning (either for a Local Area Network (LAN) or for a WAN).
Currently Federal Student Aid has a WAN in place supporting its legacy borrower
services systems, and, at the same time, has the CSB initiative underway that is
redeveloping, re-platforming, and transitioning many of the legacy systems but utilizing
the existing WAN, with periodic modifications to the WAN to support the CSB initiative.

The CSB system, in its legacy state and planned end state, operates in a distributed
environment that requires applications to be accessed, provided for, and run across WAN
connections (or “run across the wire”). During the development of applications and
systems that are designed to “run across the wire,” typically there are three core
components involved that help ensure the successful implementation of the solution:
• Application Analysis – Application Analysis profiles the impact the application
will have concerning software licensing needs, server requirements, support
requirements (such as help desk and system maintenance requirements), and
minimum requirements for end user systems.
• Load Testing – Load Testing (also referred to as Performance Testing) looks at
the performance of servers and server configurations under various conditions to
help identify potential bottlenecks or potential application concurrent use issues,
and to develop a baseline for performance tuning of the application when it enters
production.
• Bandwidth / Application Impact Analysis (capacity planning) – Bandwidth /
Application Impact Analysis is concerned with the number of users that will
utilize the application / system (concurrence), communication requirements for
each connection (bandwidth or traffic requirements), prioritization of application
traffic and end-user response time (quality of service), and impact of the
application traffic on the overall performance of the WAN (provisioning).
This paper is focusing on Bandwidth / Application Impact Analysis, but before
discussing this topic, a basic explanation of some network capacity planning concepts is
provided for the reader.


WAN Bandwidth /

Network Capacity Planning Concepts
Network Capacity Planning, at its core, is the process of modeling, forecasting, and
anticipating the future from a networking perspective. Capacity Planning, at its best, is a
complicated and complex discipline. Depending on the type of analysis and planning
being done, capacity planning:
• Can be accomplished with the use of a few tools along with some manual
analysis, or
• May require specialized expertise from various networking fields and become
unwieldy and overwhelming.
The ultimate goals of network capacity planning are:
• The creation and maintenance of simple, practical, and flexible architecture;
• Cost-Effective use of communication lines, equipment, bandwidth, and budget ;
• An architecture that supports the organization and performance expectations;
• The provision for some immediate and incremental growth in capacity without
having to acquire new circuits or equipment; and
• A design that can support “Just-In-Time” delivery of bandwidth.
WAN provision is expensive and can require between 30 to 90 days for new circuits to be
installed to add bandwidth capacity. WAN provisioning is typically viewed only in terms
of bandwidth. However, with the advent of web services, video streaming, Voice Over IP
(VOIP), and more distributed application processing, WAN provisioning should be
viewed and addressed in terms of quality of service.

Quality of Service Concepts
While Quality of Service (QoS) concepts and requirements are commonly associated
only with services such as video streaming and VOIP, QoS has broader applications for
use, that, when employed, can optimize WAN traffic and ensure levels of services
between different classes of network traffic. It is also a misconception that networks
with high-capacity backbones and links do not need QoS management. In reality, all
networks can benefit from QoS Management from the standpoint that all networks have
congestion points and QoS can help level the spikes and drop offs of circuit or segment
utilization and provide a more consistent performance rate, especially from the point of
view of the user.


WAN Bandwidth /

Application Traffic Management
Application Traffic Management is a discipline that helps manage technical requirements
to help ensure that performance guarantees and business objectives are met. Adding
more “bandwidth” is a strategy that delivers diminishing
Web Services / Web margins of return in today’s world of 24x7 operations,
Applications using SOAP /
Web Services and web-based extended enterprise
XML will require more
bandwidth than traditional applications (See Table 1 – Evolution of Distributed
distributed applications models. Processing for a brief history of Distributed Processing).
- SOAP Primer by W3C The old paradigm of adding bandwidth is no longer
considered a good strategy where: a) reserve bandwidth
is more of an oxymoron in the context of web services and web-based extended
enterprise applications; and b) web services can consume as much as 10 times more
bandwidth than older legacy client-server traffic.

Imperative for
Type of Network
Era Network Environment Application Traffic
Communications
Management
1960’s-1 Mainframe Systems Dedicated Circuits Low
980’s
1980’s-1 Centralized Corporate IT– Local / Campus Networks Medium
990’s Client Server Based Separate Voice & Data
Networks
2000 - Distributed/De-centralized Local / Campus Networked High
Present IT Departments – Mission supporting Multimedia, VPN
Critical Web Services, Internet access
Extranets, Intranets, Video / Shared / Combined Voice &
Voice services, Multi-Tier Data networks
Architecture
Table 1 – Evolution of Distributed Processing

In today’s world, bandwidth is now a service, not a product or a provision. This requires
end-to-end visibility by WAN administrators and Network Planners into the nature of
communications required by and being performed by each application that is required to
transverse the WAN (and/or LAN). To accomplish this end-to-end visibility,
organizations typically deploy appliances or other tools that help identify how WAN (and
Internet) resources are consumed. Some of the more commonly used tools and
appliances for doing this are:
• CISCO® IOS NetFlow™
• Packeteer PacketSeeker™
• Compuware Application Expert™
• AdventNet’s NetFlow Analyzer™


WAN Bandwidth /

Bandwidth / Application Impact Analysis Process
Bandwidth / Application Impact Analysis can be performed at either a high-level or a
detailed level with the goal of gaining visibility into the nature and type of
communications in use or required. However, the high-level and detailed analysis do not
seek to answer the same questions. The high-level Bandwidth / Application Impact
Analysis looks at all the traffic on the WAN and shows bandwidth utilization by protocol
as well as by segment. Along with this, the high-level analysis is used to develop trend
and performance information and is used to help the organization identify rogue traffic
present on the WAN.

The detailed Bandwidth / Application Impact Analysis Process is where each application
or service, individually, is analyzed to understand the traffic the application or service
generates. This analysis results in the creation of a profile of the application or service
for the traffic generated by it, a model of how the application or service makes use of
bandwidth. The profile that is developed for each application or service can then be used
for multiple purposes; from WAN capacity planning or as a baseline for potential future
tuning of the application, and it has the potential to be used in developing load testing
scenarios, performance testing scenarios, and metrics for the application itself.

Performing both the high-level and detailed Bandwidth / Application Impact Analysis
provides the organization visibility into the traffic flows and identifies applications and
services generating traffic that transverse the WAN. When Bandwidth / Application
Impact Analysis is incorporated as part of the System Development Life Cycle for both
the Network (WAN or LAN) AND for application development, this process can become
simple and efficient. More importantly, it can reduce costs and allow capacity to be
modeled or forecasted to allow bandwidth to be optimized and/or added “Just-In-Time”
as the traffic patterns and volume change.

Bandwidth / Application Impact Analysis is centered around the following processes:
• Identification and Categorization of Traffic
• Impact Analysis of adding new Application Traffic
• Prioritization of Services / Applications
• Baselining / Trending Traffic Patterns
• Monitoring and Tuning
• Reporting

Identification and Categorization of Traffic
Identification and Categorization of Traffic is the profiling of traffic, identifying what
traffic is associated with each application / service, and categorizing the traffic into
typically one of three types:
• Legitimate, Business/Mission supportive Traffic – This category includes all
traffic that is related to doing business or supporting the mission and can be
identified and traced to the enterprise applications.


WAN Bandwidth /

• Inappropriate Traffic – This category includes traffic related to recreational
Internet Traffic, viruses, or rogue applications.
• Unwise Traffic – This category is traffic that results from business-related
applications / services that are performed at inappropriate times or where
business-related applications are “stealing” bandwidth. This could be traffic
scheduled to occur at different times (such as system backup or database
synchronization) other than peak periods to avoid the need to upgrade bandwidth
and improve performance.

During the Identification and Categorization of Traffic, it is important to identify the
overhead required for the traffic, especially since overhead typically adds about 30% to
the volume of traffic. Overhead includes items such as packet acknowledgements,
routing information, as well as the various protocol headers.

Impact Analysis of Application Traffic (new or existing)
Impact Analysis of Application Traffic creates a profile for an application that:

• Defines the traffic characteristics of key transactions that make up the application
(i.e., latency requirements1, protocols used and/or identifying the traffic generated
across the 7 Layers of the OSI Model2, and measured bandwidth requirements.)
• Defines how users will interact with the application and includes information such
as the number of concurrent users, number of locations, and peak use by time of
day as well as by location.
• Defines how remote sites / remote users will access the application, routing
information, existing bandwidth loads and latencies.

To create this profile and define the information discussed above requires a combination
of interviews (to define how users will interact with the application and the locations
where the application will be used) and the use of tools to examine and capture actual
information in a test environment. Tools can include network sniffers or traffic analyzers.
Development of this profile will also require a mix of personnel from Network Engineers
to Application Owners and, in some cases Application Developers.

1
Latency is the time delay between the moment something is initiated and the moment one of its effects
begins. Latency in network terms is typically based on the time from the source sending a packet to the
source receiving a response that the packet has been received (round-trip latency).
2
A brief high level overview of the 7 Layers of the OSI Model is included as an appendix to this white
paper.


WAN Bandwidth /

For example, a very basic Impact Analysis Profile Report might look like this:
Application Impact Analysis Profile

Application Name: _______________________________
Estimated Date for Deployment: _______________________________
Application Owner: _______________________________

Key Transactions (please attach any supporting information created from the use of network sniffers
or traffic analyzer tools):
Transaction / Protocol / Bandwidth / Latency Duration Avg. #
Service Name Port Packet Size Requirements Peak / Non
Peak
Transactions
1) TX1-
RX2-
2) TX-
RX-
1 – TX = Transmit 2 – RX - Receive

User Information:
Number of total users: __________
Total Expected Concurrent Users (peak usage): __________
Total Expected Concurrent Users (non-peak usage): __________
Number of Users via Remote Access (i.e. VPN): __________

Locations Where the Application will be used:
Location Total # of Non Peak Peak Avg. # Peak Avg. # Non
Users Concurrent Concurrent Transactions Peak
Users Users Transactions

Existing Location Infrastructure:
Location Type of Size of Circuit Current Available Current
Circuit Utilization Bandwidth Circuit
Latency

Prioritization of Services / Applications
Once all the services and/or applications that are running across the WAN are identified,
an organization will need to review and prioritize each of the services and/or applications
based on mission need. This prioritization ideally is done in conjunction or coordinated
with Business Continuity Planning and Disaster Recovery Planning. After the services /
applications are ranked based on the mission requirements of the organization,
information concerning the nature of the traffic is added to develop the WAN
prioritization in order to develop policies for traffic control and/or quality of service


WAN Bandwidth /

control. For this discussion, WAN priority is concerned with a priority policy. Some
typical prioritization policies are:

Priority Policy Description
Priority Policy Establishes a priority for traffic without specifying a set transmission rate.
Typically this priority is a number between 0 and 5 or 7, where 0 is the lowest
priority.
Rate Policy Delivers a specified guarantee rate for class of service defined as a rate.
Discard Policy All traffic associated with this policy is dropped theoretically blocking the service
/ application.

A sample prioritization table might look like this:

Network Service / Application Prioritization

Nature of Traffic
Mission
Application / Time Packet Size (Large/ Prone to WAN
Supporting
Service Sensitive Small , Bursty/Non- Jitter (drop Priority
Rank
(Latency) Bursty) packets)
Accounting 1 Yes Varies No Priority 5
VoIP 1 Yes Varies Yes Rate
CRM 2 Yes Varies No Priority 5
Oracle 2 Yes Varies No Priority 5
E-Mail 3 No Large / Bursty No Priority 3
DNS 3 No Small / Non-Bursty No Dynamic
Streaming 4 Yes Varies Yes Rate
Audio / Video
Web Browsing 5 Varies Large / Bursty No Rate
Instant 0 Varies Varies No Discard
Messenger

Baselining / Trending Traffic Patterns
Baselining and Trending Traffic involves the use of traffic analysis information from
various tools the organization has employed (such as CISCO IOS NetFlow, health
reports from routes, etc.) to develop and maintain baselines about the WAN that include
things like:
• Top Talkers and Conversations on the WAN
• Performance and Utilization Statistics
• Projecting Traffic Trends and Usage Patterns


WAN Bandwidth /

Monitoring and Tuning
As a result of all the data that the organization has now gathered, knowing what
applications are running and how they use the bandwidth that is available, the
organization is better able to optimize and run the WAN more efficiently.

With monitoring, the organization is able to:
• Dynamically allocate WAN resources to business-critical applications through
bandwidth management,
• Contain rogue traffic,
• Forecast potential congestion points,
• Identify unauthorized traffic quickly, and
• Develop better Cost / Benefit and Return on Investment models.

As a result of this monitoring, along with the Baseline and Trend information, the
organization can fine-tune the performance of the WAN even further. Thanks to the
insight and control resulting from the processes discussed in this paper, the organization
may be able to:
• Deploy compression technologies
• Create virtual bandwidth out of existing resources
• Avoid costly bandwidth upgrades or time upgrades to arrive “just-in-time” to
minimize performance issues

Reporting
As with any process, reporting is a requirement in Bandwidth / Application Impact
Analysis. Reporting should involve the development of reports that are useful to key
stakeholders including: network managers, application managers, business managers,
security officers, and business continuity executives.


WAN Bandwidth /

Conclusion
While Bandwidth / Application Impact Analysis can be overwhelming and complex at
times, it is a crucial aspect of Network Capacity Planning and Network Design. In the
long run, instituting Bandwidth / Application Impact Analysis can help an organization
save costs and provide a high performance level for the applications that run across the
WAN. Bandwidth / Application Impact Analysis provides visibility to the organization
into applications running on the WAN, their requirements for bandwidth, and utilization
patterns. Without it, an organization is blind to the true nature of the traffic on the WAN
and can end up taking “shots in the dark” when resolving performance issues and/or
adding capacity.

Having this type of information available before deploying a new application that will
transverse the WAN can assist the Federal Student Aid WAN / Virtual Data Center
Provider in ensuring that routers, firewalls, and WAN Circuits are configured correctly,
help prevent implementation communication issues, and provide a more secure WAN by
being able to secure ports and protocols that are not used.


WAN Bandwidth /

Appendix

The 7 Layers of the OSI Model
The OSI Model (also referred to as the “protocol stack”) is a hierarchical structure of
seven layers that defines the requirements for communications between two computers.
The model was conceived to allow interoperability across various platforms, and since
the 1980’s has been the model used as the standard for network communications. The
seven layers of the OSI Model are defined in Table A-1 below.
Layer is
Layer Layer # and
Concerned Definition Examples
Supports Name
With
Host Layer

Everything at this layer is application-specific.
This layer provides application services for file HTTPi,
7–
transfers, e-mail, and other network software FTPii, E-
Data Application
services. Tiered application architectures are part Mail
Layer
of this layer. Telnet and FTP are applications that (SMTP)
exist entirely in the application layer.
The presentation layer works to transform data
6– into the form that the application layer can accept.
Data Presentation This layer formats and encrypts data to be sent SSLiii
Layer across a network, providing freedom from
compatibility problems.
The session layer sets up, coordinates, and
5 – Session terminates conversations, exchanges, and
Data TCPiv
Layer dialogues between the applications at each end. It
deals with session and connection coordination.
This layer provides transparent transfer of data
4–
between end systems, or hosts, and is responsible
Segments Transport TCP, UDPv
for end-to-end error recovery and flow control. It
Layer
ensures complete data transfer.
Media Layer

This layer provides switching and routing
technologies, creating logical paths known as
virtual circuits for transmitting data from node to
3 – Network IPvi, IPSecvii,
Packets node. Routing and forwarding are functions of this
Layer BGPviii
layer, as well as addressing, internetworking, error
handling, congestion control and packet
sequencing.
This layer furnishes transmission protocol
knowledge and management and handles errors in
the physical layer, flow control and frame
Ethernet,
synchronization. The data link layer has two sub
Wireless,
2 – Data layers, the Media Access Control (MAC) layer
Frames Frame
Link Layer and the Logical Link Control (LLC) layer. The
Relayix,
MAC sub layer controls how a computer on the
ATMx
network gains access to the data and permission to
transmit it. The LLC layer controls frame
synchronization, flow control and error checking.
This layer defines all the electrical and physical T1 Line,
1 – Physical specifications for devices. This includes the layout 10Base-T,
Bits
Layer of pins, voltages, and cable specifications. 100Base-
TX, DSL
Table A-1 - 7 Layer OSI Model


WAN Bandwidth /

OSI Model Control Flow
With the OSI Model, control is passed from one layer to the next, starting at the
application layer in one station, and proceeding to the bottom layer, across the physical
connection (or cable) to the next station and back up the protocol stack. Diagram 1
illustrates how control is passed between each layer.

Diagram 1 – OSI Model Control Flow


WAN Bandwidth /

OSI and Letter Communication Parallel
To help better understand the OSI Model and translate it into plain English, Diagram 2 –
RM – OSI and Letter Communication Parallel below draws an analogy (or parallel)
between the 7 Layer OSI Model and a US Postal letter sent from a manager in one
division or company to another manager in a different division or company.

Original parallel drawing done by Josef Sábl. This drawing was modified to include the color bands.
Diagram 2 – RM – OSI and Letter Communication Parallel


WAN Bandwidth /


i
Endnotes associated with Table 1 found in the Appendix
E

HTTP – Hyper Text Transfer Protocol (HTPP) is a request/response protocol between clients and servers. The originating
client, such as a web browser, spider, or other end-user tool, is referred to as the user agent. The destination
server, which stores or creates resources such as HTML files and images, is called the origin server. In between
the user agent and origin server may be several intermediaries, such as proxies, gateways, and tunnels.

An HTTP client initiates a request by establishing a Transmission Control Protocol (TCP) connection to a
particular port on a remote host (port 80 by default; see a list of well-known ports). An HTTP server listening on
that port waits for the client to send a Request Message.

Upon receiving the request, the server sends back a status line, such as "HTTP/1.1 200 OK," and a message of its
own, the body of which is perhaps the requested file, an error message, or some other information.
ii
FTP – FTP or file transfer protocol is a commonly used protocol for exchanging files over any network that
supports the TCP/IP protocol (such as the Internet or an intranet). There are two computers involved in an FTP
transfer: a server and a client. The FTP server, running FTP server software, listens on the network for connection
requests from other computers. The client computer, running FTP client software, initiates a connection to the
server.
iii
SSL – Secure Sockets Layer (SSL) provides endpoint authentication and communications privacy over the
Internet using cryptography. In typical use, only the server is authenticated (i.e., its identity is ensured via a digital
certificate that can be verified by the client out of band) while the client remains unauthenticated; mutual
authentication requires both the client and server to maintain digital certificates. The protocols allow client/server
applications to communicate in a way designed to prevent eavesdropping, tampering, and message forgery.
iv
TCP – The Transmission Control Protocol (TCP) is one of the core protocols of the Internet protocol suite. Using
TCP, applications on networked hosts can create connections to one another, over which they can exchange data or
packets. The protocol is connection-oriented in that it guarantees reliable and in-order delivery of sender to
receiver data. TCP also distinguishes data for multiple, concurrent applications (e.g. Web server and email server)
running on the same host.
v
UDP – The User Datagram Protocol (UDP) is one of the core protocols of the Internet protocol suite. Using UDP,
programs on networked computers can send short messages known as datagram’s to one another. It is a
connectionless or best effort protocol.
vi
IP – The Internet Protocol (IP) is a data-oriented protocol used for communicating data across a packet-switched
network.
vii
IPSec – IPSec (IP security) is a standard for securing Internet Protocol (IP) communications by encrypting and/or
authenticating all IP packets. IPSec is a set of cryptographic protocols for (1) securing packet flows and (2) key
exchange.
viii
BGP – The border gateway protocol (BGP) is the core routing protocol of the Internet. It works by maintaining a
table of IP networks or “pefixes” which designate network reachability between networks.
ix
Frame Relay – Frame Relay (also found written as "frame-relay") consists of an efficient data
transmission technique used to send digital information quickly and cheaply in a relay of frames to one or
many destinations from one or many end-points.
x
ATM – Asynchronous Transfer Mode, or ATM, is a cell relay network protocol which encodes data traffic into
small fixed-sized (53 byte; 48 bytes of data and 5 bytes of header information) cells instead of variable sized
packets (sometimes known as frames).

Note: The definitions for the list above were derived from Wikipedia, the free online encyclopedia.

Wide Area Network (WAN) Bandwidth / Application Impact Analysis

More Related Content

Similar to Wide Area Network (WAN) Bandwidth / Application Impact Analysis (20)

Wide Area Network (WAN) Bandwidth / Application Impact Analysis