Migrating enterprise applications to cloud

2 The State of the Pre-Cloud Enterprise
2 The Cloud: What and Why
3 Getting to the Cloud
6 Which Applications Fit —
and Which Don’t

Table of Contents
MIGRATING ENTERPRISE..
APPLICATIONS TO THE CLOUD..
A solid strategy to help organizations reap the many benefits
offered by cloud computing.
Executive Summary
One of the most important strategic issues IT decision-
makers face is how to best leverage the rapidly evolving
cloud resources available today. Those who make the right
move could substantially improve the efficiency, reliability and
flexibility with which they deliver critical IT services to their
internal and external users.
Effectively migrating enterprise applications to the cloud
means considering a number of salient points, including:
•Challenges and opportunities motivating organizations to
consider cloud migration
•Key attributes and potential benefits of various cloud-
provisioning options
•Issues and concerns faced by IT decision-makers regarding
cloud migration
•The suitability of specific types of applications for
cloud migration
•How to best ensure the success of a cloud migration project
With homework and a sound strategy, IT leaders should be
empowered to make smarter decisions about migrating
applications to the cloud, to act on those decisions with
greater confidence and ultimately to achieve better results
for the business.

2
The State of the
Pre-Cloud Enterprise
IT organizations today are under intense pressure. On one
hand, IT can deliver more value than ever when it comes to
automating business operations, ensuring a great customer
experience, capitalizing on new opportunities and pinpointing
potential problems that require immediate attention. On
the other hand, no organization has an unlimited IT budget.
Every organization’s bottom-line performance relies, to
some extent, on how well it leverages its finite IT resources to
implement and run high-value enterprise applications.
A limited budget isn’t the only issue IT faces when it comes to
enterprise applications. Other issues include the need to:
•Scale applications in terms of number of users, volume of
data and other parameters despite the limited capacity of an
in-house data center
•Ensure operational resiliency even in the event of data
center disruptions such as hardware failures, regional power
outages or extreme weather
•Allocate more IT resources to highly innovative and strategic
projects — rather than simply keeping existing applications
and infrastructure up and running
While these issues have become critical for IT decision-
makers, cloud computing has emerged as a viable alternative
to conventional application provisioning. Cloud computing
decouples application provisioning from specific hardware
and sometimes even from a specific data center. In doing
so, it provides numerous benefits that align closely with the
concerns noted above — as well as several additional benefits.
Because the cloud-computing model is so attractive,
90 percent of companies are already using the cloud in some
form or another, according to a study by the Computing
Technology Industry Association (CompTIA). And 69 percent
of companies are planning to move at least some of their
critical applications to the cloud by the end of 2014, according
to research conducted by Virtustream.
Despite this powerful market momentum, many IT decision-
makers still have serious questions that they need to answer
before they move their applications to the cloud. These
questions include:
•Will the enterprise be able to maintain the same standards for
security and compliance?
•Will the performance of applications suffer as the result of a
move to the cloud?
•Can applications in the cloud integrate and interoperate
with each other as they do when they are running in the
data center?
•Could any other unintended consequences result from
a move to the cloud?
•How does the organization decide which applications to move
to the cloud first — and which ones not to move at all?
•If and when the entity decides to move an application to the
cloud, how can it do so with the least disruption and risk to
mission and IT operations?
Clear, accurate answers to these questions are obviously a
prerequisite for any successful move to the cloud.
The Cloud: What and Why
The term “cloud” refers to an alternative delivery model
for traditional IT that can be hosted at an external service
provider’s facility, on-premises in a company’s own data
center, or both. Regardless of where it resides, infrastructure
can generally be considered cloud if it possesses the
following characteristics:
•Metered: Cloud computing is a metered technology that
offers accurate calculation of CPU use. The process also
serves to increase awareness among departments and users
of the true costs associated with IT services. Keep in mind,
while metered usage is part of the formal cloud definition,
failure to charge for that isn’t necessarily an internal private
cloud deal-breaker.
•Elasticity: Cloud environments should offer the ability to
dynamically add or reduce computing capacity based on
actual need and on demand. Typically, historical and real-
time reporting provides stakeholders with visibility into
fluctuations in both demand and the corresponding allocation
of capacity.
•Automation: Virtualization offers the agility needed to speed
up IT operations and trim costs by boosting infrastructure
utilization. The self-managing aspects of autonomic
computing automates the process by which the user can
provision resources on-demand. Environments with these
characteristics offer a wide range of benefits when it comes
to the delivery of enterprise applications.
•Reducing or aligning costs: Cloud computing combines
economies of scale with the ability to allocate computing
capacity on an as-needed basis. This drives down the cost
of provisioning enterprise applications — while also keeping
those costs closely aligned with actual mission needs, even
as those needs fluctuate over time. This reduction and
alignment of IT costs is becoming increasingly important
as organizations become more dependent on a growing
portfolio of enterprise applications.
•Scaling on demand: Entities often experience periods
of peak demand for their enterprise applications. Some
peaks are predictable — such as those associated with
end-of-quarter reporting or seasonal sales cycles. Others
occur suddenly and unexpectedly due to the success of
a marketing campaign or a disruption in the supply chain.

3
Under conventional provisioning models, organizations face
a difficult choice when it comes to this peak demand: Either
burn limited capital budgets on excess capacity to handle
those peaks or endure subpar application performance until
those peaks pass. Cloud computing eliminates this choice by
making capacity available when it is actually needed without
requiring the organization to pay for unnecessary capacity
year-round.
•Shifting Capital Expenditure (CAPEX) to Operational
Expenditure (OPEX): Chief financial officers have become
increasingly aggressive about preserving capital. Cloud
computing supports this financial imperative by allowing
companies to avoid capital expenses on IT — to instead “rent”
infrastructure on an as-needed basis from service providers
as an operating expense.
•Reallocating IT resources to strategic innovation:
Most organizations devote a high percentage of their
IT budgets to maintaining operation of IT equipment,
especially when it comes to ensuring the reliability of critical
enterprise applications. This spending severely limits their
ability to allocate resources to new, innovative IT projects.
Cloud computing can free up these precious resources
and promote innovation by reducing infrastructure costs
and allowing operational burdens to be offloaded to
service providers.
•Leveling the playing field: Due to their relative lack of
capital IT budget and skilled IT staff, small to medium-
sized organizations often struggle to implement the same
technologies as their larger competitors. Cloud computing
levels the playing field by giving smaller organizations access
to the computing resources they need. In today’s market —
where success often depends upon the effective use of
state-of-the-art technology — this “democratized” access
has proved to be a game changer.
•Increased time to market: With cloud computing, IT server
provisioning, which can take six-to-eight weeks, can
be completed in a matter of hours. While cost savings
and the ability to scale up and down are attractive, some
organizations are pushed to the cloud by the ability to get
products to market much more quickly.
Cloud computing offers other benefits as well. For example,
economies of scale allow cloud service providers to build
failover facilities more cost-efficiently than individual
organizations can. This built-in infrastructure resiliency
often makes the cloud an attractive way to safeguard
business continuity. Those same economies of scale also
often make it easier for service providers to engineer secure
remote access to enterprise applications — making them
an attractive choice for supporting mobile users as well as
contractors and partners.
Getting to the Cloud
While businesses can gain significant benefits by running their
enterprise applications in the cloud, several key issues must be
taken into consideration before doing so.
Application Suitability
Not all applications are architected to run in virtual or
multitenant environments. In some cases, an application has
intensive CPU or input/output needs that require a dedicated
multithread processor and therefore don’t lend themselves to
virtual machines that share physical CPU and I/O capacity with
other applications.
In other cases, it may be because the application vendor has
not structured the licensing of the application for a virtualized
or multitenant environment.
Whatever the reason, IT decision-makers must first determine
whether the application in question can run on a modern
virtualized infrastructure “as is,” whether it must be modified
to do so, or whether the cloud is totally out of the question
because of specific technical limitations.
Flavors of Cloud: Which Is Right for You?
While “cloud” generally describes any computing architecture
that leverages virtualization and the resulting ease of
provisioning resources, cloud implementations vary significantly
in their suitability for different application workloads — and in
their economics — based on how and where they are deployed.
Here are the key cloud “flavors” to consider:
TYPE DESCRIPTION PROS CONS
Hosted
Public/
Multitenant
Service
providers build
large-scale cloud
environments on
infrastructure
that is securely
allocated to
the application
workloads of
many different
organizations.
Infrastructure is
highly scalable.
Operational
burdens are
offloaded.
Economies of
scale result in
lower costs.
Multiple tenants
bring potential
for compliance
and performance
issues.
Organization
has less
direct control
of data and
infrastructure.
Hosted
Private
Service
providers
host cloud
environments on
infrastructure
dedicated
to specific
applications,
customers,
industries, etc.
Combines
scalability and
operations
offload with
enhanced
compliance,
security and/
or control.
Private cloud
may not
offer same
economies of
scale as public/
multitenant
cloud offerings.
Hybrid Application
workloads run
in a combination
of private- and
hosted-cloud
environments.
Allows for mixing
and matching of
different cloud
flavors based on
varying business
and technical
requirements.
Requires more
sophisticated
management
and monitoring
of application
behaviors and
multiple cloud
environments.

4
Network Requirements
When considering the use of hosted-cloud services,
the network requirements of an application are a key
consideration. This is because application traffic will have
to travel a much different path than it does when the
application is running in an enterprise data center.
For some applications — such as those designed to run as
pure web browser–based services — this can be a nonissue.
For others, network latency or jitter can be highly problematic.
So the network tolerances of an application must be well
understood before a cloud migration.
These tolerances are often complex. Many applications
entail connectivity between multiple components such as
a database, a middleware server, a presentation server and
an authentication server.
These multiple network paths can all have implications when
it comes to the end-to-end performance of an application.
IT decision-makers must therefore factor in all of these paths
when determining which application components, if any,
can reside in the cloud and which ones need to remain in the
enterprise data center.
That said, tight network tolerances are not necessarily a deal-
breaker for cloud migration, since some cloud providers offer
better network capabilities than others.
Security
Stakeholders — including line-of-business managers,
compliance managers and customers — have legitimate
concerns about the security of data that is allowed to reside
outside the control and confines of an enterprise data center.
IT decision-makers, therefore, have to carefully consider the
security implications associated with running applications
on a hosted-cloud infrastructure.
In particular, the security of cloud providers should be
evaluated based on several criteria, including:
•Encryption of data, both in motion over the network and at
rest in the storage environment
•Access control in terms of how users are authenticated and
how granularly permissions can be granted or denied for
different application-related privileges
•Perimeter security such as firewalls, intrusion detection,
periodic penetration testing and other best practices
•Physical security to ensure that unauthorized personnel
cannot gain direct access to servers and storage arrays
where sensitive data resides
IT decision-makers should bear in mind that the real
question at hand is not whether a given cloud provider’s
environment fulfills some theoretical ideal of perfect
security — but rather whether it is more or less secure
than other provisioning options.
Keep in mind, no IT environment is perfectly secure, but cloud
provider environments are often much more secure than
their enterprise counterparts because they are newer, more
homogeneous and operate at economies of scale that allow
for more intensive security operations.
Compliance
Many organizations operate under regulatory constraints
that mandate how certain types of data must be managed
and protected. This obviously has implications when it
comes to provisioning applications in the cloud, because that
provisioning will directly affect both where data is stored and
where it moves while in use.
In some cases, regulations may make it impractical to move
an application to the cloud. Companies operating in several
countries, for example, may be subject to local regulations that
forbid certain types of data from crossing national borders.
This may make it necessary to keep data in separate, dedicated
physical servers in each country where the company operates.
These data centers may leverage virtualization — and may
even individually utilize cloud-like computing infrastructure.
But they will not be able to dynamically share computing
capacity across all territories where the company operates.
The Need for Speed
When it comes to accessing public-cloud services, not all
network connections are created equally. IT decision-makers
therefore must carefully consider the tolerances of the
application in question with regard to latency, jitter and other
network characteristics — and balance those against the
cost of enhanced network connectivity. This is true whether
public-cloud services are being accessed directly from an in-
house data center or from a service provider’s facility.
Three main types of network connections can be used to
access public-cloud services:
TYPE OF
CONNECTION
ATTRIBUTES COST
Public Internet
VPN
The lowest common
denominator for
connections to a
public-cloud service,
subject to the variations
associated with generic
Internet connectivity.
Low to none, depending
on bandwidth
requirements
Multiprotocol
label switching–
enhanced
Enables “class of
service” connectivity
that insulates the
performance of cloud
connections from
other less-critical
Internet traffic.
Additional costs
charged by service
provider. May require
some reconfiguration
of internal network.
Direct
connection
Access speeds
substantially
improved by dedicated
connectivity to
public-cloud service
that bypasses the
public Internet.
Premium pricing

5
In other cases, regulatory mandates may affect the choice of
cloud service providers. Examples of such regulations include:
privacy requirements of the mandate and assume liability for
any breaches. This shields healthcare companies from the
financial and legal risks associated with data security breaches
— although it does not necessarily mitigate the reputational
risks associated with such an event.
Service Level Management
When applications run on conventional infrastructure within
an organization’s own data center, service level management
— while it can be somewhat challenging technically — is
nonetheless relatively simple in principle. Applications run over
known infrastructure components that are under the direct
control of the company’s IT staff. So if a problem arises, that
infrastructure can be quickly assessed to determine the cause
of the problem.
Things get a bit more complicated with in-house cloud
environments because the infrastructure underlying an
application is more fluid. This fluidity can make problems more
difficult to diagnose and resolve.
The challenge is ensuring the level of service which changes
dramatically when an application is running on public-cloud
infrastructure. For one thing, the organization and the
service provider must both clearly define what constitutes
satisfactory performance in a complete and unambiguous
service-level agreement.
The SLA should define not only expectations regarding the
technical performance of the application, but also response
times promised by the service provider in the event of a
problem. A well-designed SLA will also define the roles and
responsibilities of each party in resolving problems — as well
as the consequences if conditions are not met.
Health Insurance
Portability and
Accountability
Act (HIPAA)
A federal law that includes
provisions that companies
involved in healthcare must
follow regarding the security
and privacy of patient data
Payment Card
Industry Data Security
Standard (PCI DSS)
A set of requirements designed
to ensure that all companies
processing, storing or transmitting
credit card information
maintain adequate security
Federal Risk and
Authorization
Management Program
(FedRAMP)
A program that standardizes
how government agencies
and contractors address
cloud security assessment,
authorization and monitoring
These regulations affect application deployment in the cloud
differently. PCI and FedRAMP, for example, provide criteria
for certifying cloud service providers. An organization
operating under either of those mandates, and looking to move
applications to the cloud, could readily fulfill its compliance
requirements by choosing a certified partner.
HIPAA, on the other hand, does not specifically define any
technical certification for cloud service providers. Instead,
cloud service providers confident of their ability to fulfill the
requirements of HIPAA offer customers in the healthcare
industry a Business Associate Agreement (BAA), through
which the service provider agrees to both fulfill the data
Service models and skills
Cloud service providers offer several models for running enterprise applications on their hosted, virtualized infrastructure. These models
differ significantly in many ways, including how they impact the skill sets required to use them. The following chart offers a general
overview of service models available and their comparative impact on IT staff skills.
SERVICE MODEL ACRONYM DESCRIPTION BENEFITS SKILLS IMPACT
Software as
a Service
SaaS Service provider manages all
aspects of application provisioning,
from underlying hardware to
ongoing software updates.
All responsibilities for the
application are offloaded to the
service provider.Customization
of the application is limited to
configuration capabilities made
available from the provider.
Little to none: May need to learn
some configuration tools.
Platform as
a Service
PaaS Service provider manages the
complete underlying technology
“stack,” up to but not including
an application itself: hardware,
operating system, hypervisor, etc.
Provides access to scalable
capacity and offloads bulk of IT
management responsibilities
while still allowing full control
over enterprise applications.
Minor to significant: The
service provider may use
different technologies (such
as a different hypervisor) with
which IT staff are unfamiliar.
Infrastructure
as a Service
IaaS Service provider manages only
the hardware infrastructure.
Provides access to scalable
capacity and offloads some IT
management responsibilities
while still allowing full
control over enterprise
applications, hypervisors,
operating systems, etc.
Minor: Requires use of remote
management and monitoring tools.

6
Of particular importance is how problems are reported,
tracked, resolved and then reviewed periodically to determine
if both parties’ expectations are being met. Ideally, these
processes should be sufficiently automated, which may entail
some integration between the two parties’ IT help desks or
trouble-ticket systems.
Pricing
Cost is obviously a primary consideration for any company
engaging with a cloud service provider. But determining
the cost of a cloud engagement can be tricky given the fact
that service providers offer pricing structures that differ
considerably.
In addition, the long-term cost of any cloud engagement can
be difficult to estimate, because it is by definition tied to usage,
which can be difficult to predict. Organizations should also take
steps to ensure that they do not accidentally run up excessive
charges due to unnecessary or unauthorized usage.
IT shops should consider other costs related to the cloud as
well. For example, while an organization retains ownership of
the data it generates over time in the cloud, cloud providers
may charge a fee for the bulk transfer of that data. This cost
should be factored into any evaluation of a cloud provider’s
pricing structure.
On the other hand, a cloud strategy can cut maintenance costs
by having an outside resource monitor, patch and maintain the
infrastructure. It can also boost levels of efficiency as multiple
virtual servers can be provisioned on the same physical
hardware. This often results in reduced use of datacenter
space, and therefore a reduction in the demand for power
and cooling.
Which Applications Fit —
and Which Don’t
In addition to understanding the general principles associated
with cloud-based application provisioning, IT managers should
consider the suitability of specific types of applications for
the cloud.
Software Development, Testing and
Quality Assurance
Entities with active internal software development teams can
benefit substantially from the cloud. Developers and quality
assurance testers often need significant computing capacity
for relatively short periods of time. Internal IT operations
teams may not be able to provide this capacity quickly or cost-
efficiently, which can delay important development projects
and consume developers’ time.
The cloud offers developers near-immediate and relatively
inexpensive access to the computing capacity they need. It
can also give them access to the different environments they
may need to thoroughly test how well their code runs under
a variety of potential real-world conditions. This can help
accelerate the delivery of new high-quality applications
and services.
Just as important, with the cloud, this capacity can be
rented only for as long as it is needed. This eliminates the
waste associated with buying extra hardware for
development teams.
Collaboration Apps
Email, web conferencing, document sharing and other social/
collaborative applications are increasingly important to
organizations seeking to increase productivity, accelerate
business processes, improve decision-making, foster
innovation and leverage institutional knowledge.
Cloud computing allows organizations to provide communi-
cation and collaboration tools from anywhere. This offers the
ability to support work from any location, at any time and on
any device. Furthermore, it provides credence to the bring-
your-own-device (BYOD) phenomenon and the resulting
incremental productivity offered to the organization.
Added security comes in the way of authentication and
security protocols offered by the cloud service provider for
access outside of the firewall. The CSP also provides ongoing
maintenance efforts needed to keep protocols secure and
data compliant.
Personal Productivity Apps
Desktop word processing, spreadsheet and presentation
apps — in particular Microsoft Word, Microsoft Excel and
Microsoft PowerPoint — have become staples of knowledge
worker productivity. They have also become more expensive
to own and manage as users increasingly work on multiple
devices from diverse locations. Licensing can be a particularly
thorny issue as many employees work from a varying number
of devices.
Microsoft has responded to the changing needs of knowledge
workers with Microsoft Office 365, a cloud-based version of
the company’s popular Office suite that provides anytime/
anywhere access — along with other useful features such as
calendar syncing and collaboration tools. By migrating users
to this type of solution, companies can reduce their software
ownership costs while better supporting today’s workers.
Big Data, Analytics and Other Computing-
Intensive Apps
Every minute of every day, users and computing endpoints
generate a wealth of data. From Facebook posts to utility
meter readings, this data offers tremendous potential for
actionable insights about changing market dynamics, sales
opportunities, operational problems and exposure to risks.

7
To realize this potential, companies must be able to rapidly
process massive volumes of Big Data, perform sophisticated
analysis of it and deliver results to users in a visually intuitive
way. That kind of computing requires a lot of processing
power, storage and throughput, which can be expensive to
acquire and maintain. Plus, Big Data workloads are often highly
sporadic, requiring massive capacity for short periods of time
and then extended periods of inactivity.
Here again, the cloud is an obvious fit. With the right cloud
service, organizations can get access to the Big Data
processing capacity they need for as long as they need
it — without massive capital expenses for capacity they
don’t need. Many cloud providers also specialize in analytic
applications and offer advanced and highly customizable
analytic algorithms that help find the high-value “needles”
in Big Data “haystacks.”
One major caveat: Organizations should seriously consider
cost and time factors when moving large amounts of data to a
cloud service provider’s environment. Additionally, significant
costs may be incurred if data has to be moved out of that
environment. However, in most cases, the capital savings and
elastic scalability of cloud environments more than offset
these costs.
Disaster Recovery, Business Continuity
and Records Retention
As organizations become increasingly dependent on
applications and other IT services on a moment-to-moment
basis, they also must safeguard those applications and
services from all types of disasters — including extreme
weather, regional power outages, terrorist attacks and even a
large-scale failure of an enterprise data center.
The cloud can clearly play a central role in provisioning
recovery and continuity. With the cloud, replicated versions
of critical applications and their associated data can be
maintained in one or more locations sufficiently distant from
a main enterprise data center that they are unlikely to be
vulnerable to the same disaster. Users can then access those
resources from other locations.
Organizations can also leverage cloud service providers to
comply with regulations that govern records retention. The
cloud can provide economies of scale that reduce overall
storage costs — while also providing rapid access to data over
the network when necessary.
Apps That May Not Work: ERP and Other
Core Systems of Record
Most organizations have a set of core applications, such as
enterprise resource planning (ERP) systems, that function
as their primary systems of record. It may be challenging or
impractical to migrate these applications to the cloud for a
variety of reasons:
•Performance issues: Because these applications tend to
be transaction-intensive, they may require high network
throughput and very low latency. This performance may be
difficult to achieve in a cloud environment.
•High-availability requirements: While the cloud
offers excellent overall continuity and availability, any
application accessed over the Internet is susceptible to the
occasional “hiccup.” For systems where even a relatively
brief interruption can have significant consequences,
organizations may be better off investing in high-availability
infrastructure in an in-house data center.
•Sensitive data: Systems of record are almost always
repositories for some of an organization’s most sensitive
information — including customer data and financial records.
Regulatory compliance and governance best practices may
therefore prevent movement of this data to the cloud.
•Application architecture: For companies that have been
running their businesses on the same ERP system for many
years, the architecture of the application itself may preclude
migration to a virtual environment.
The cloud is not a panacea. It won’t eliminate every IT challenge
an organization faces, and it isn’t appropriate for every
application. But with the right insight and the right partners,
companies can leverage the evolving range of cloud solutions
to drive down costs, improve productivity, empower managers
to make better decisions and drive the kind of innovation that
dramatically improves business performance in dynamic and
competitive markets.
The Best of Both Worlds: Leveraging
the Hybrid Cloud
To move some applications to the cloud, organizations may
be best served by adopting a hybrid model, which combines
multiple types of cloud or conventional infrastructure.
Hybrid-cloud models include:
•Own the base, rent the spike: Under this model, normal day-
to-day workloads run on in-house IT infrastructure (which
can be either conventional or a private cloud). Off-premises
hosted-cloud resources are used only during periods of
peak demand to maintain service levels to users.
•Cloud as firewall: Under this model, an off-premises cloud
environment is used to run customer-facing processes
that may support web or mobile applications. Back-end
application components such as databases and transaction
processing systems remain in-house.
•Compliance segmentation: Here, data is split between
in-house and off-premises environments. This enables
compliance with data-related regulatory mandates, while
still allowing the benefits of cloud computing to be applied
where appropriate.

Migrating enterprise applications to cloud

More Related Content

What's hot (20)

Viewers also liked (9)

Similar to Migrating enterprise applications to cloud (20)

More from Sougata Mitra (20)

Recently uploaded (20)

Migrating enterprise applications to cloud