Fs tech-journal-cost-management

Financial services technology journal
The cost management issue

Professional IT managers know that controlling costs and aligning
expenditures with overall business goals and spending are critical
to organizational health. But it’s a classic case of “easier said
than done.” With continuing overcast skies forecast for the global
economy, every step you take to infuse your IT organization with
fiscal agility now should be appreciated—at the bottom line and in
every line of business.

PricewaterhouseCoopers
The cost management issue
Letter from the executive editor 1
This issue at a glance 4
The business of technology
Using IT business management and cost transparency 6
practices to improve operating leverage
Becoming a cost-agile IT organization 11
How it (should) all start
Taking the first step toward software as a service 17
Controlling IT costs with selective sourcing 23
Data for decision
Creating financial flexibility through 27
service-oriented architecture
Below the code
Easing desktop management costs with 31
innovative new technologies
Supporting next-generation consumption-based 37
hosting services with cloud computing

1PricewaterhouseCoopers
Optimizing IT spending for cyclical financial markets
The fluctuations of financial services firms’ business volumes reflect the cyclical
nature of the overall financial markets. These dynamics are often caused by
specific crises—such as the most recent subprime mortgage problems—or
a slowdown in the economy. These fluctuations require that executives and
technology leaders have the ability to restrict spending levels in market downturns
and quickly scale up when business volumes rise again.
During previous periods of market turbulence, executives have demanded budget
cutbacks and cost savings from their IT organizations. However, because many
IT costs are fixed, IT executives have limited options for reducing expenditures.
Typical cost-saving initiatives entail rationalizing IT assets and resources and
renegotiating vendor contracts.
Fixed IT costs cannot be scaled back easily to react quickly and appropriately
to market downturns. Optimization of the “operating leverage,” which is defined
as the percentage of fixed costs relative to overall operating costs, increases a
company’s ability to lower its IT operating expenses quickly during an economic
slowdown.
This issue of the Financial Services Technology Journal discusses approaches to
optimizing operating leverage. We begin by defining fixed costs and discussing
key areas or “levers” that often transition well from a fixed- to a variable-cost
basis. Subsequent articles relate to these levers and provide key considerations
for defining and assessing how to better manage IT costs.
Fixed vs. variable costs
Every technology spending category can combine fixed and variable costs. Figure
1 illustrates the breakdown of costs that are typically fixed or variable in an IT
operating budget.
Fixed costs do not depend on business activity. Variable costs are proportional to
business activity.
Evaluating the risk by understanding the cost structure
By understanding the cost structure and demand drivers of IT spending, we can
model their effects on operating leverage. A simple modeling approach consists of
four steps:
1 Estimate demand variation (i.e., maximum, minimum, average) for any
given IT service
2 Document the current IT expense structure (e.g., fixed and variable expenses)
and assess costs related to IT services delivery
3 Evaluate these cost variances as demand fluctuates within the range from step 1
4 Identify service alternatives and create scenarios with built-in demand
variations— where related expenditures are mostly variable—and estimate the
new costs derived from these scenarios
Letter from the executive editor

2 Financial services technology journal
The difference between the costs calculated in steps 3 and 4 provides some
visibility into the mix of fixed and variable IT budget items:
• The cost difference between steps 3 and 4 with the lowest demand represents
the effect that fixed IT costs have on profit margins and earnings during a
downturn. This amount also represents a potential margin for action to convert
fixed to variable costs.
• The cost difference between steps 3 and 4 with the highest demand estimates
the cost scalability of the current asset mix. A positive figure might reflect
adequate or excess resources for this level of demand.
Large financial services firms may find it difficult to track all their assets and
obligations and, thus, experience significant variances in expenditures vs.
forecasts when demand fluctuates. Therefore, it is difficult and complex to
precisely and accurately evaluate the current mix of resources and variations
because of demand shifts. In addition, the past, limited focus on operating
leverage resulted in few tools to track fixed and variable costs, making it
more difficult.
Fixed investments to reduce unit costs
During past profitable cycles, business users expected IT organizations to
provide high-quality services at optimal costs based on business projections—
from conservative to optimistic. Their IT organizations made large investments
(e.g., large purchases of dedicated servers, hiring large application development
teams, building large data centers to host equipment) to provide these services
at the lowest unit costs. These investments resulted in fixed operating expenses
that became part of IT’s budget and, later, subject to the scrutiny of company
executives. Due to heavy capital investments, one of IT’s major fixed-cost budget
categories is depreciation.
IT capital investments that result in a greater proportion of fixed operating costs
seemingly allow IT to provide services for a lower cost per unit compared to other
solutions (e.g., outsourcing, leveraging virtualization to reduce capital requirements)
in the short term. Because IT organizations face pressure to pursue the lowest-
cost solutions, those with high fixed-cost percentages are implemented the most
to comply with cost-savings guidelines and because of their perceived simplicity.
While the high fixed-cost IT model meets short-term cost-savings goals, it does not
optimize operating leverage. And it might lead to future constraints when additional
cost savings become necessary. Furthermore, fixed operating costs cannot be
eliminated easily when corresponding IT services demands drop.
IT executive challenges
Each business has its own specifications and requirements when it comes to
fixed vs. variable costs. In most cases, IT executives have to tackle the following
challenges to protect company margins during economic downturns:
• Dealing with the past IT executives have to cope with fixed costs from past
investments and find innovative and economical ways to reduce them.
• Planning for the future IT executives need to make the right decisions on future
investments with limited market visibility.
Figure 1
Fixed vs. variable IT operating costs
Mostly
fixed
Mostly
variable
Employee compensation X
Employee other expense X
Contractor/professional services X
Facilities
Rent/lease
Utilities
Maintenance
Facilities depreciation
X
X
X
X
Hardware
Hardware depreciation
Hardware leasing
Hardware maintenance
X
X
X
Software
Software amortization
Software nonamortization
Software maintenance
X
X
X
Communication X
Outsourcing X

• Changing expectations and behaviors IT executives typically face resistance
to change because this requires a paradigm shift, although solutions are now
available to change the current IT operating environment and add capacity or
new functionalities while increasing operating leverage (e.g., utility computing,
software as a service).
To improve IT operating leverage, IT executives need to address the
following requirements:
• Managing change People and process changes are often the most difficult
to implement; new behaviors must be promoted/enforced to support new
measures to improve operating leverage.
• Implementing new metrics New ways of tracking will be required with the
adoption of any new standards and their positive effects on:
– Operating leverage
– Skillfully dealing with vendors – Agilely handling growth
• Providing transparency IT must deliver transparency in its services and cost
and price structures; when making investments in new products and services,
IT should provide the needed insight on how that improves operating leverage
over time.
Preparation is the best defense
Professional IT managers know that controlling costs and aligning expenditures with
overall business goals and spending are critical to organizational health. But it’s a
classic case of “easier said than done.” When dealing with long-term amortization
costs and new technology uses—where return on investment can be difficult to
calculate—there’s a need for information that takes effective IT cost management
leadership to a new level.
That’s exactly what you get in this issue of Financial Services Technology Journal.
We’ve painstakingly planned this issue to offer effective, up-to-date thought
leadership on IT cost planning. From the critical right-now technologies (e.g.,
service-oriented architecture, cloud computing) to practices that involve outsourcing
and out-tasking (i.e., outsourcing simple tasks) to new paradigms in service models,
you should find this to be an invaluable resource in your efforts to improve your
organizational operating leverage. With continuing overcast skies forecast for the
global economy, every step you take to infuse your IT organization with fiscal agility
now should be appreciated—at the bottom line and in every line of business.
But in the end, leadership determines the opportunity for and success of IT
operating leverage. So we have strived to support IT leaders by arming you with
the strategic and operational insights that our consultants have acquired while
tackling some of the most complex problems facing IT organizations.
Good reading!
Julien Courbe
Managing Director, PricewaterhouseCoopers

This issue at a glance
Now, a major IT question is, “How do we transform fixed
costs into variable costs?” In this issue of the Financial
Services Technology Journal, we survey multiple levers for
improving operating leverage, which involve changes to
operating models and technology implementations.
As in past issues, we have organized our articles along four
dimensions:
• The business of technology section explores how IT
organizations can meet enterprise requirements and
demands by providing businesses with visibility and cost
transparency to support better operating leverage. In our
first article, “Using IT business management and cost
transparency practices to improve operating leverage,”
we focus on the importance of IT business management
and how it can help bridge the gap between business and
technology. Our second article, “Becoming a cost-agile IT
organization,” explains how adopting a consumer-oriented
approach can better position IT to scale back during times
of economic turbulence.
• Our second section, How it (should) all start, begins by
examining the new software as a service (SaaS) model
for deploying applications such as Web-based services
hosted by service providers, as described in the “Taking
the first step toward software as a service” article. We learn
about the advantages of SaaS, its cost-reduction benefits
and which techniques a company can utilize to assess its
readiness. Our second article, “Controlling IT costs with
selective sourcing,” focuses on how selective sourcing can
help reduce IT costs during turbulent times.
• We look at the Data for decisions dimension in the
article “Creating financial flexibility through service-
oriented architecture,” which focuses on service-oriented
architecture (SOA) as a more dynamic and agile service-
based business model than was represented by its
past static incarnations. In this article, we explore the
benefits, strengths, weaknesses and strategic approaches
necessary for adopting SOA.
• In the Below the code section, we bring you the leading
practices for converging new technologies and processes
for improving desktop management and computing
technology costs while keeping users satisfied. In our first
article, “Easing desktop management costs with innovative
new technologies,” we examine current desktop practices
and provide strategic recommendations for achieving all
your organization’s goals. The second article, “Supporting
next-generation consumption-based hosting services with
cloud computing,” focuses on the technologies behind
cloud computing, the current industry service model,
its limitations, the disruptive nature of cloud computing
and the potential effects of cloud computing on the cost
structures of IT organizations and enterprise data centers.
Using IT business management and cost transparency
practices to improve operating leverage
One of the objectives of IT business management is to
provide businesses with visibility into and appreciation of
the costs of supplying IT solutions. Business management
also helps IT better plan and design for IT services,
anticipate business demand and bridge the gap between
business and technology. The net effect is to transform
traditional enterprise IT organizations into true IT service
providers that are managed like businesses. Business-
aligned IT organizations are in sync with the overall goals of
the enterprise and can react proactively to changes in the
businesses that they serve.
Becoming a cost-agile IT organization
Historically, IT organizations have focused on supply-
side economics in the management of their business and
operational costs. By focusing on service provisioning rather
than service consumption, IT organizations often lost sight of
how their services aligned with business requirements and
demand. This article discusses how adopting a consumer-
oriented approach to both understanding demand and
developing solutions consistent with that demand can better
position IT to scale back volumes and costs during times of
economic turbulence.
Taking the first step toward software as a service
Software as a service (SaaS) is a new model for deploying
applications as Web-based services hosted by service
providers. The SaaS model proposes that companies can
significantly reduce costs by shifting their IT costs from
a fixed- to a variable-cost model. SaaS could replace
traditional hosted applications with outsourced services
priced by actual usage. In the current economic climate,
such a variable-cost model is attractive enough that
corporations are investing in it. This article discusses
the merits of SaaS, its strengths and weaknesses and
a systematic evaluation technique that enterprises may
consider to assess their readiness for SaaS.

Controlling IT costs with selective sourcing
By leveraging selective outsourcing and out-tasking
opportunities, financial services managers can benefit by
moving from a largely fixed-cost IT model toward a partial
variable-cost strategy. The objective of this article is to
review how selectively sourcing one or more subsets of a
financial services company’s IT resources may provide new
levels of flexibility. In this way, managers can incrementally
reduce their companies’ computing capacities and costs
during a business downturn.
Data for decisions
Creating financial flexibility through service-oriented
architecture
Service-oriented architecture (SOA) implementations are
showing companies that the traditional corporate IT model
needs a makeover. Today, many corporate IT models simply
cannot address the financial requirements of an SOA-
supported enterprise. Not surprisingly, IT organizations are
eagerly embracing the opportunities offered by SOA. The
evolution of SOA transforms it from being a static function
into a more agile, service-based business function. That is
an important competitive differentiator. However, it means
tweaking the corporate IT model to realize the benefits of
SOA.
SOA-supported enterprises are based on eco-systems
of business functions that represent the interests and
needs of multiple lines of business (LOB). However, many
responsibilities must be determined before implementing
SOA. For example, who should pay for certain services?
How are fees broken down if services are reused across
an enterprise? How should SOA projects be financed? The
current financial model in many companies is appropriate
only in budgeting vertical LOB applications. SOA requires
a different vision and execution. Success in SOA initiatives
means organizations will have to adopt a strategic approach
that better realizes the associated financial benefits.
Below the code
Easing desktop management costs with innovative new
technologies
As demands from technology-savvy employees increase,
companies are finding it more difficult to cut desktop
management costs while keeping users satisfied. Now,
the convergence of new technologies and associated
management processes is helping transform desktop costs
into variable costs. The benefits of this approach include
meeting user expectations while leveraging data center
and IT investments. This article examines current desktop
support practices and the factors transforming them into an
on-demand, variable-cost model service. It also provides
strategic recommendations on achieving these goals.
Supporting next-generation, consumption-based hosting
services with cloud computing
Cloud computing is a key enabler of the software-as-a-
service model that supports the deployment of applications
as Web-based services hosted in “clouds” of technology
assets (e.g., servers, storage, network devices). The open
secret of cloud computing is its ability to split application
workloads across shared resources, which can be
geographically dispersed, while supporting application
performance and reliability as computing demand fluctuates.
Cloud computing also eliminates the need to invest in
hardware, software or facilities necessary to run applications
and supports billing for services on a pay-as-you-go basis.
This article reviews the technologies behind cloud
computing, the current industry service model and its
limitations. Then it discusses current trends and capabilities
in cloud computing. Today’s risks and benefits are also
detailed. Finally, it discusses the disruptive nature of cloud
computing and its potential effects on the cost structures of
IT organizations and enterprise data centers.

Despite multimillion-dollar investments, many information technology (IT)
organizations today cannot quickly respond to evolving business needs and
adapt to large-scale shifts like mergers, sudden drops in sales or new product
introductions. Moreover, visibility into true IT costs, usage and chargebacks is
frequently poor while vulnerability to costly downtime increases.
In today’s tough economic climate, those situations are unacceptable. IT
organizations need better control of their cost structures and operating leverage.
They need to be more agile. One solution is IT business management (ITBM).
Effectively meeting business demands requires IT to be run like a business. ITBM
pushes costs out of cost centers into profit centers where demand originates, so
those who require services pay for them. When implemented properly, ITBM can
reduce IT’s percentage of organizational fixed operating costs. ITBM internalizes
management disciplines, including process, organization, technology and facilities
within IT operations. As a result, IT becomes a value center rather than a cost
center.
ITBM also can help shift an IT organizational cost structure and reduce its
operating leverage in two ways. First, it gives end users visibility into actual IT
services costs and business decision-making abilities, based on anticipated IT
services costs. With ITBM, the full cost of IT services is factored into business
decisions at the corporate and business-unit levels. For example, if a business
unit decides to offer a new revenue- generating service for its customers, an
effective ITBM program shows the IT-related costs of developing and maintaining
this service. Then, the business unit can decide whether to pursue this business
opportunity.
Second, ITBM provides users both the ability and responsibility to adjust IT
service levels based on true demand and associated costs. Just as consumption
will differ at an all-you¬can-eat buffet vs. a restaurant where the same dishes
are served a la carte, the same is true of consumption of IT services. It is easy to
understand why a business that does not pay for its IT services has the potential
to use more than it needs, knowingly or unknowingly. Factoring these costs into
profit center operating budgets is a solid method for promoting reasonable IT
usage.
At the core of these concepts lie the challenges of IT cost transparency.
Using IT business management
and cost transparency practices to
improve operating leverage
Abstract
One of the objectives of IT business
management is to provide businesses
with visibility into and appreciation of
the costs of supplying IT solutions.
Business management also helps IT
better plan and design for IT services,
anticipate business demand and
bridge the gap between business
and technology. The net effect is
to transform traditional enterprise
IT organizations into true IT service
providers that are managed like
businesses. Business-aligned IT
organizations are in sync with the overall
goals of the enterprise and can react
proactively to changes in the businesses
that they serve.
1
The following framework, as described, is part of PricewaterhouseCoopers’ proprietary Technology Cost Transparency solution.

Six pillars of IT cost transparency
Six major pillars that any effective IT cost transparency framework1
must address
include:
1 Costing — determining the actual cost of IT services
2 Chargeback— accurately and equitably charging back for services
3 Data/infrastructure—the data, assets and application mapping tools supporting
IT services delivery and management
4 Service catalog—establishing a menu of services and a bridge between costing
and chargeback
5 Management reporting—providing essential regular and ad hoc reporting
capabilities, predictive modeling and visibility into the organization
6 Governance—including the communications plan, program management and
leadership buy-in supporting the IT cost transparency framework
The maturity level of each pillar (Figure 1) can vary from reacting to orchestrating. In
a reacting state, ad hoc actions are taken without proper planning or visibility. This
often results in fixed-cost investments that may not realize their potential over time.
In an orchestrating state, IT spending decisions are based on increased visibility
into their effect on operating leverage.
Supporting an ITBM strategy that emphasizes variable costs or promotes services
that can take advantage of fixed-cost investments means examining each pillar
with an eye toward improving the current maturity level (Figure 2). It also means
identifying the required people, tools and processes to progress from reacting
mode to orchestrating mode when using an ITBM strategy to control fixed costs.
Understanding your environment
When addressing your IT cost structure, an adequate system of demand
monitoring and modeling is required. The inability to distinguish between a blip
and a sustained rise in required service levels often causes investments in excess
capacity. Experience shows that proper high-level management of IT services
growth includes dashboard functionality such as trend modeling and providing
quick snapshots for executives.
When modeling trends in which IT services are in demand, a dashboard should
allow for intuitive estimates of sustainable service levels. It also should include
reliable estimates of future investments and should effectively relate this
information to specific business units and demand drivers.
Also, a dashboard must be easily understood by non-IT executives and business
managers in order to create awareness that some business uses of IT services are
not financially justifiable. It should disclose upcoming investment plans, including
spending levels and whether costs are fixed or variable. This information helps
business managers understand and better appreciate IT organization attempts
to decrease operating leverage. It also creates more-informed decision makers
willing to work with IT in critically evaluating the business logic of IT service
demands.
Once an understanding is established about your environment, IT can leverage
data to design an ITBM strategy that can drive organizational change toward
lower fixed costs. The result is a more agile cost base.
ITBM and cost transparency
tools and practices can transform
an IT organization’s operating
leverage. These tools and practices
provide critical information and
transparency to those who are
demanding IT services and shift
demand toward services more
conducive to a variable cost base.
Reacting
Coping
Anticipating
Orchestrating
Figure 1.
ITBM pillar maturity level vs.
organizational benefits
ITBM pillar maturity
From reacting to orchestrating...
Benefitstoorganization

Figure 2.
Information technology business management maturity levels vs. six pillars of IT cost transparency
ITBM pillar Reacting Coping Anticipating Orchestrating
Costing • Ad-hoc reports are
developed manually to
address business-unit
inquiries
• IT cost elements are not
fully identified
• Limited availability of IT cost
information
• Inconsistent compute
methodology across IT
domains
• No activity-based costing
model
• Consistent IT costing
framework and methodology
• Activity-based costing
methodology that provides
some transparency and drill-
down capabilities
• IT costing framework and
methodology that provide
chargeback information to business
units on a business transaction basis
• IT costing methodology that drives
demand management of IT resources
and allows for accurate forecasting
and predictive modeling to better
forecast for incremental investments
requirements
Chargeback • No chargeback mechanism
in place to recoup IT costs
from business units
• IT costs remain in IT
department/budget
• No visibility between
IT organizations (e.g.,
between application
development and
infrastructure)
• High-level chargeback
methodology allocating costs
across business units using
estimated drivers
• Business units have difficulty
understanding IT charges
• Variances are unexplained
based upon shifts in
business volumes
• Mix of rate x volume and
fixed charges
• Some visibility into upcoming
demand, allowing IT
organizations to adjust
chargeback rates and avoid
one-time adjustments
• Mostly rate x volume based
chargeback
• Chargeback information is
posted to general ledger and
used in financial reporting
• IT chargeback methodology that
provides transparency into IT
costs and linkages with changes
in strategic direction or business
volume (e.g., opening or closing a
new branch)
• Visibility into chargeback can help
organizations promote IT behaviors,
improving operating leverage
Data/
infrastructure
• Driver data is not readily
available, extracts not
defined
• Lack of IT resource-
consumption measurement
tools
• Mapping of applications
across shared servers is
manual and inconsistent
• Driver data is not readily
available, extracts not defined
• Lack of IT resource
consumption measurement
tools
• Mapping of applications
across shared servers is
manual and inconsistent
• Cost-driver data is stored
in multiple mediums (e.g.,
Microsoft Excel spreadsheets
and source systems) and
aggregated manually
• Some IT resource consumption
measurement tools are in place
• Cost-driver data, IT asset data,
IT resource-consumption data is
collected and stored in a central
repository (or virtual repository)
• Actual usage of infrastructure assets
by shared applications is measured
and metered
Service
catalog
• No service catalog in place
• Business units have no
common point of reference
for IT services
• No service catalog in place
• Business units have no
common point of reference
for IT services
• A paper-based IT service
catalog is in place
• List of revised IT services is
updated annually
• IT service costs are not well
understood
• An online IT service catalog is in
place with real-time updates
• IT service costs structure and
variation are understood by
management and users
Management
reporting
• Limited management
reporting
• No predictive modeling
or forecasting based on
business trends
• Limited management
reporting
• No predictive modeling
or forecasting based on
business trends
• Management reporting
performed manually in
Microsoft Excel or Access
• Limited forecasting and
predictive modeling, mostly
done manually
• Costing/chargeback system provides
robust operational and financial
reporting to support transparency
goals, highlighting fixed vs. variable
components
• Predictive modeling and forecasting
is in place
Governance • Components in place do
not have formalized written
documentation
• Some components may
occur on an ad-hoc basis
• Formalized governance
program is documented,
however, no mechanisms are in
place to track or measure the
governance program’s effects
• Formalized written governance,
however, results of changes
implemented are not measured
• Program is comprehensive, but
no continuous improvement
plan is in place
• Governance program structure
is in place and measures the
effectiveness of the components in
improving operating leverage
• Continuous improvements are
planned to further control fixed costs

Key strategic initiatives
The specific initiatives that help transform the ratio of fixed to variable costs for an
IT organization represent each of the six pillars:
Costing
Initiatives focus—auditing/overhauling accounting data attributes and identifying
fixed vs. variable costs based on historical trends
Benefits—costs categorized in terms of activities, allowing for unpooling and
itemization of fixed and variable components
Chargeback
Initiatives focus—enhancing transparency of internal invoicing with IT service
consumers
Benefits—line-of-business (LOB) visibility into breakdown of charges (i.e., direct/
indirect, fixed/variable); socialization and widespread ownership of operating
leverage/risk reduction regarding IT expenditures
Service catalog
Initiatives focus—establishing consistent service catalog technology and
implementing it in request centers; creating cost models based on cost objects
that identify fixed vs. variable costs
Benefits—provisioning up-to-date, readily available service prices, influencing
current demand for IT services; informing requests for future service designs,
positively affecting operating leverage
Management reporting
Initiatives focus—defining key performance indicators; forecasting baseline
demand necessary to determine acceptable levels of fixed investment; setting
up a program management office to manage business requirements, tools and
technologies; managing business intelligence
Benefits—top-down control of an acceptable threshold of fixed investment,
based on management assessment of sustainable demand; defending against
overly optimistic LOB leads
Data/infrastructure
Initiatives focus—auditing/consolidating IT asset use data; device and
application mapping; virtualization and metering
Benefits—hardware costs, fixed by their nature, helping to prevent usage
increases, effectively avoiding unnecessary investments; collecting essential
data for validating costs charging back to LOBs for services, since estimating not
immune to variance or error

Governance
Initiatives focus—articulating roles, authorities and communications protocols
for IT managers and service consumers; implementing reports training; publishing
common metrics
Benefits—sophisticated, well-designed management controls wasted when there
is unclear ownership and responsibility; overly engineered processes, delaying
sensible investment unnecessarily; training current and future decision makers
on content and interpretation of service catalogs and management reports for
effective IT service execution, including appropriate investment decisions
Conclusion
ITBM and cost transparency tools and practices can transform an IT
organization’s operating leverage. These tools and practices provide critical
information and transparency to those who are demanding IT services and shift
demand toward services more conducive to a variable cost base. An effective
ITBM strategy must address each of the six pillars of IT cost transparency at a
sufficient level of maturity in order to change behavior and raise awareness among
business IT services users.
Figure 3
Addressing all six pillars to effectively change organizational operating leverage
Strategy
transforming
operating leverage
Service
catalog
Coating
ITBM
Chargeback
Data/
infrastructure
Governance
Management
reporting

A new approach to managing IT costs
Worried about turbulent market conditions and a possible recession, financial
services organizations are searching for opportunities to reduce IT costs.
Traditionally, these organizations focused on reducing discretionary spending,
tightening IT budgets and reprioritizing or deferring IT projects. However, these
bottom-line approaches to cost reduction rarely factor in the effect such steps
have on the quality of IT services provided to their businesses. Initiatives to
reduce IT costs can have profound effects on the availability, performance and
improvement opportunities of IT services.
An emerging approach to managing costs in IT organizations that operate as
true service providers is to manage the demand for business services. Forward-
thinking financial services organizations are implementing demand management
models based on service catalogs that define and capture the economics of the IT
services provided (i.e., which services, at what service levels, at what costs). This
approach increases organizational control of IT operational costs. The challenge
for business-oriented IT organizations is to closely align their goals, processes,
IT supply chains and cost structures with changes in the businesses they serve.
The current economic crisis has led financial services companies to reevaluate
short-term business goals and requirements and accelerate their move toward a
demand management model based on IT service consumption.
Managing service consumption
Traditionally, IT organizations have relied on supply-side economics to manage
their business and operational costs, focusing on managing service provisioning
rather than service consumption. These organizations are so intent on how
they create services and make them available—by developing IT solutions—
that they often forget why they do this—to align services to business drivers.
Service portfolio management practices are often not well ingrained in such
organizations, resulting in decisions that cannot be rationalized through
rigorous cost-benefit analysis. As a result, IT organizations that rely on a service
provisioning management model often incur large fixed costs that cannot be
attributed to specific services or lines of business. Because they are so focused
on provisioning services, these organizations have little incentive to analyze and
respond to marginal variable costs that can be attributed to specific services and
service levels.
On the other hand, IT organizations that operate under an IT service consumption
management model perform tasks related to:
• Managing service levels
• Managing the supply chain
• Managing delivery costs
Managing service levels
Becoming a cost-agile IT organization
Abstract
Historically, IT organizations have
focused on supply-side economics
in the management of their business
and operational costs. By focusing
on service provisioning rather than
service consumption, IT organizations
often lost sight of how their services
aligned with business requirements
and demand. This article discusses
how adopting a consumer-oriented
approach to both understanding
demand and developing solutions
consistent with that demand can
better position IT to scale back
volumes and costs during times
of economic turbulence.

IT departments driven by service provisioning used performance metrics such as
mean time between failures or mean time to repair to measure their technology
components. Although this model helped IT departments offer technology
support, manage IT vendor supply chains and support personnel, it did not
provide decision makers with transparent information on how their business
systems were affected by changes in IT resources and performance levels.
Service-oriented IT organizations began measuring the performance of the
services they provide using critical success factors and key performance
indicators. These metrics provide business units visibility into the effect IT
resource-level adjustments have on the services provided by IT departments.
However, services were still being managed at the point of provisioning rather
than at the point of consumption, with metrics being tracked for IT services
but not for business drivers and goals. As a result, organizations could not
easily determine the implications of their business goals or the potential loss of
productivity and/or revenue due to changes in IT resources.
Business-oriented, process-mature IT organizations leverage service management
tools and methodologies to align business goals with IT services and components
(Figure 1) by mapping business processes to them and measuring the near real-
time financial, business and operational effects of technology outages. New
products from leading vendors have had success in developing these concepts into
dashboards with metrics at the component, service and business-process levels.
Managing the supply chain
Generally, IT organizations that manage service provisioning are concerned with
the back end of IT services supply chains (i.e., developing and/or purchasing IT
solutions from vendors). Although this model has led to comparatively mature
development of vendor and service-level management processes, they are not
always aligned with business requirements on the demand side. IT organizations
that do not systematically focus on demand and service portfolio management
Figure 1
Evolution of business-oriented IT organizations
• Manage costs by managing service consumption
• Dynamic service portfolios
• Variable cost models are aligned with business services
• Focus is on preventing and resolving
technology component outages
• Manage to the “point of provisioning”
• Fixed-cost models difficult to align
with business services
• Define clear services and service targets
• Service delivery and support processes
are streamlined
• Focus is on service demand management, planning
for future capacity and reducing the number of
service incidents and outages

cannot correlate changes in business demands and adjustments to IT operational
processes and resources.
The emerging approach to IT service portfolio management offers “a dynamic
method for governing investments in service management across the enterprise
and managing them for value.”1
Service portfolio management helps companies:
• Define inventory services and business cases
• Analyze the value of their service portfolio
• Align supply and demand
• Approve services, resource allocations and chartering decisions
• Communicate service definitions and/or modifications
Organizations with well-defined portfolio management functions (i.e., front-end
supply chain management for IT services) influence demand by introducing
return-on¬investment concepts and infusing IT decision making with business-
driven goals.
One step in IT services supply chain management is to define and manage service
catalogs, subsets of the services offered in IT services portfolios. These end-user
services are implemented with several different “views” to meet the functional
requirements of their target audiences at each link of the service supply chain.
For example, the buy view represents functions to select, customize and request
IT services. The sell view reflects product/service manager views and organizes
IT services into services families, such as hosting, storage, communications and
middleware. Each services family reflects logical technology groupings used for
reporting and billing. The sell view also includes mappings to IT functions that
support specific services. The operate view maps IT services to the units that
support them in terms of people, technologies, vendors and/or financials. Tight
integration between these three views creates a well-managed IT services supply
chain that is responsive to changes in business requirements.
Managing delivery costs
The service provisioning approach to IT services has resulted in IT budgets with
large fixed costs and IT organizations that cannot adjust services and costs
to meet changing business demands at specific performance levels. On the
other hand, managing service consumption incorporates more granular costs
into IT budgets and aligns IT organizations with specific services/performance
levels demanded by the business. Financial services organizations that want
to implement service-aligned cost structures should use activity-based costing
methods or other top-down models suitable for cost allocation based on usage
and/or performance tiers.
Once organizations establish cost-allocation methodologies, they can begin
to manage the cost of delivering IT services through service consumption
management. However, to make these initiatives successful, companies must
significantly redefine the relationship between their IT and business organizations.
The dialogue between IT organizations and business units must change from
an annual budget review to an economic negotiation about providing specific
business services at specific performance levels.
1
Information Technology Infrastructure Library v3.0, United Kingdom Office of Government Commerce, May 2007.

Information Technology Infrastructure Library v3.0 and
cost-agile organizations
Information Technology Infrastructure Library®
(ITIL®
) v3.0 provides a path
for IT organizations to migrate from process-oriented service provisioning
management to business-services driven service consumption management. ITIL
3.0 was introduced when IT organizations were attempting to augment their IT
operational processes with industry-leading practices to become more cost agile
in response to changing business demands. With ITIL 3.0, IT organizations have a
more viable and focused framework to manage service consumption in line with
their overall business cost structures. This framework provides cost agility by:
• Aligning IT goals with business direction—whereas ITIL 2.0 advises how to
improve what IT does already, ITIL 3.0 tells IT organizations what they should be
doing, saying that “for services to provide true value to the business, they must
be designed with business objectives in mind. Service design is the stage in the
life cycle that turns service strategy into the blueprint for delivering business
objectives.”2
• Focusing on service life cycle management—the service strategy, design,
transition, operations and continuous improvement life cycle stages all provide
key activities and metrics for managing the portfolio of IT services, managing
the service supply chain through business consumption of IT services and
measuring and demonstrating business value.
Business service
management
IT service
Utility
Name, description,
purpose, effect, contacts
Assets/resources
Systems, assets,
components
People/partners
Resources, staffing, skills,
suppliers
Warranties
Service levels, targets,
service hours, assurances,
responsibilities
Processes
Processes
Applications
Service-level agreements/
service-level requirements
IT processes
Data
Suppliers
Support services
Infrastructure
Support teams
Operational-level
agreements
Contracts
Business processes
Services
Business processes
Figure 2
Required configuration management database service consumption mappings

Operating IT departments as businesses is an ambitious initiative for most
organizations. Managing service consumption rather than service provisioning
allows IT organizations to align business goals with IT goals and provides the
cost agility they need to run their departments as businesses.
Resurgence of the configuration management database
The configuration management database (CMDB) has been around at least since
ITIL 2.0. However, CMDB has taken on strategic importance for organizations that
want to manage IT service consumption. The introduction of new technologies
by leading vendors in IT service management has also made CMDB much more
feasible to build and maintain.
Usually, CMDB has contained all the configuration items/technology components
in an IT environment as well as their interdependency mappings. In the new
service consumption model, CMDB has grown into a representation of business
processes mapped to business services and technology services/applications,
which are mapped to supporting technology components, as required (Figure 2).
IT organizations that are successful in establishing these mappings can begin to
manage service levels, service supply chains and delivery costs in conjunction with
changing business demands.
Conclusion: becoming more responsive to changing business demands
Operating IT departments as businesses is an ambitious initiative for most
organizations. Managing service consumption rather than service provisioning
allows IT organizations to align business goals with IT goals and provides the
cost agility they need to run their departments as businesses. Financial services
companies that are weathering current industry storms can leverage new business
services management technologies and CMDB tools to develop IT organizations
that are more responsive to their needs and can more easily adjust to changing
business demands.

Introduction
The current economic slowdown—fueled by difficulties encountered in 2007
and 2008 in financial services—is pushing companies to venture into often-
uncharted territory to control costs. That’s why more businesses are intrigued
with the prospect of reducing costs associated with commoditized applications
by purchasing them as a service. Because companies only pay for their actual
usage, software as a service (SaaS) is cost-effective, particularly since it allows
IT to scale up and down with business conditions and transaction volumes.
But the first major challenges of SaaS lie in determining which methods reduce
costs appropriately. SaaS differs from other commoditized applications based
on quality, maturity, features and the set of traditional software it is supposed to
replace. There is not a one-size-fits-all methodology for assessing SaaS. Those
with experience in leading SaaS assessments—a critical function—are scarce
because SaaS only recently gained increased attention from chief information
officers (CIOs) and application owners.
Companies that believe SaaS remains immature will wait for it to gel. But forward
thinkers are building experience so they can better evaluate SaaS offers and
cherry-pick those that are appropriate for their companies before integrating them.
By creating advisory groups, companies can better guide application owners and
business managers as they evaluate whether moving to SaaS makes sense and
how it could be implemented. PricewaterhouseCoopers refers to this advisory
group concept as the SaaS Technology Experience Pool (STEP).
IT applications as a training ground
An effective way of implementing the first phase of SaaS is having IT become an
early adopter. That gives IT an opportunity to provide the skeleton framework for
other STEP advisers to be able to hone their skills on migrating IT applications
to a SaaS model. Identifying the business applications that are potentially good
for SaaS comes in phase two. Later, discussions with business users will help
ascertain which applications on their side of the operation fit the SaaS model.
STEP should include both dedicated members and part-time subject-matter
experts. Together, they can identify which IT application domains to investigate
when seeking SaaS cost efficiency:
Commoditized or standard application areas—it makes sense to look at
network monitoring, help desk, IT asset management and messaging because
they are examples of commoditized or standard application areas, which are
typically nonbusiness-specific and often designed to comply with technical
standards or industry-leading practices. Many SaaS vendors offer standardized
functionality for services in these domains. A STEP team should also look at
whether other IT teams are using different commoditized tools or practices that
might be better if they were standardized.
Abstract
Software as a service (SaaS) is a new
model for deploying applications as
Web-based services hosted by service
providers. The SaaS model proposes
that companies can significantly
reduce costs by shifting their IT costs
from a fixed- to a variable-cost model.
SaaS could replace traditional hosted
applications with outsourced services
priced by actual usage. In the current
economic climate, such a variable-
cost model is attractive enough that
corporations are investing in it. This
article discusses the merits of SaaS,
its strengths and weaknesses and a
systematic evaluation technique that
enterprises may consider to assess
their readiness for SaaS.
Taking the first step toward
software as a service

Practices that require IT skills or specialization exceeding those of the IT
organization—antivirus, spam filtering, and search and security suites are some
areas that require extensive experience, which may be difficult to achieve for
many IT organizations. For example, having dedicated IT staff designing spam
filtering rules or searching algorithms is usually not a good use of personnel,
especially when IT is producing results that could more easily be provided by
a company that specializes in these areas. When examining IT practices, the
STEP team should look for those that consume many company resources for
little visible gain, areas where keeping skills up to date is becoming increasingly
difficult because there are no structured training road maps or tasks that require
research and development beyond the capabilities of the IT organization.
In-house applications that do not provide an advantage—but that require
dedicated development, test and maintenance teams. Good examples are in-
house-developed applications such as configuration management databases,
software deployment tools, inventory and procurement applications, and time and
expense reporting tools intended for consultants and other third parties. Typically,
these applications were developed to satisfy a need the market could not fulfill
with off-the-shelf tools. Today, it is likely vendors have software that performs the
same functions as the homegrown applications. STEP teams should look at SaaS
offerings and find replacements for these applications, reducing costs and freeing
developers for more business-critical applications.
When IT leads the charge toward SaaS by showing examples of the improvement
and cost efficiency it provides, business units are easier to bring onboard,
extending SaaS benefits to more of the corporation.
In the first phase of a SaaS implementation, CIOs and their IT staffs identify
applications that could create cost efficiency if they were replaced by SaaS. They
will evaluate SaaS providers and the feasibility of SaaS, in general, which helps
move projects from theory to reality while working in a controlled environment.
Assessing select IT applications early has the benefits of:
Experience acquisition—by performing a SaaS assessment, strategy, design and
implementation, CIOs and their organizations can handle potential problems in
an environment they know and control, as opposed to the business environment
in which applications belong to their clients. And they can evaluate SaaS without
adding in the complexity of dealing with requirements from business users. In
this environment, errors become learning opportunities rather than finger pointing
exercises.
Believability acquisition—CIOs who have overseen their own SaaS
implementations have an easier time convincing business units to assess and
move their applications to SaaS. Even unwilling business units can be reassured
through real-life case studies and by making key SaaS skills available.
Skills development—as the IT organization performs the application assessment,
strategy design and replacement by SaaS, the skills acquired by IT staffers help
increase the probability of success when applying SaaS to business applications.
Other benefits include reducing implementation times for subsequent initiatives
and understanding the practice of including business professionals in the process
to get their buy-in early on for these transformational efforts.
By performing a SaaS
assessment, strategy, design and
implementation, CIOs and their
organizations can handle potential
problems in an environment they
know and control, as opposed to
the business environment in which
applications belong to their clients.

Assessment
The table (Figure 1) and workflow (Figure 2) on page 16 detail critical aspects for
analyzing potential applications for their fit with the SaaS model. The workflow
also details PricewaterhouseCoopers’ recommended decision sequence.
Each item in Figure 1 requires an initial discovery and data collection phase.
IT should keep the assessment coordinated with its other SaaS activities.
For example, a discovery phase for data security in the assessment should
be coordinated with the selection of SaaS providers that offer security and
regulatory certifications. It also should be phased in over time to define the
requirements for a SaaS provider. The need for coordination will likely exceed
available IT resources, making it a good idea to connect with the company’s
project management office (PMO) for help in planning and conducting project
management activities when needed.
Area Description Importance
Application
available as SaaS
If the application is not available as SaaS, there is no need to
continue the analysis.
“Go/no go”
Data security The data processed by the application is subject to regulation (e.g.,
e-mails) or special security measures (e.g., customer private data).
High; could become a no go if the SaaS
service provider cannot comply with
regulations or risks are deemed too high
Data migration This offers ease of extracting data from the current application and
migrating it to the SaaS application.
High; could become a no go if there
is no cost-sensible way to extract
and export data or if the SaaS service
provider does not allow initial data loads
Application
dependencies
Numbers and types of dependencies between potential applications and
other applications. These dependencies create the following constraints:
• Data constraints: other dependencies receive or send data to the
potential application; the operations are not feasible in a SaaS
model for technical, cost or security reasons
• Time constraints: other applications interact with the potential
application in a time-sensitive manner; the SaaS model would exceed
the current tolerance for latency and transaction completion times
• Technology constraints: the SaaS model cannot accommodate
interfaces with other business applications cost efficiently due to
incompatibilities among application technologies
High; could become a no go if the
identified dependency group of
applications cannot be separated or
migrated completely to a SaaS model
Application
functionality
Available functionality in the SaaS application must match existing
functionality. If applications being evaluated for migration to a SaaS
model belong to areas performing standard operations, the biggest
challenge is talking business units through perceived gaps.
Medium/high
Business-critical
status
How critical is the effect on the business if the SaaS service provider
cannot offer the same application availability or performance as the
current IT organization?
Medium/high
Figure 1
Critical assessment areas for potential software as a service (SaaS) applications

Because assessing potential applications crosses organizational boundaries, it
requires a cross-functional approach. CIOs and their organizations must also be
ready to cross borders to seize opportunities that send a message to corporate
leadership: Assessments provide opportunities for IT to lead SaaS transformation
by priorities—predominantly cost efficiency—that align with business needs.
In addition to experience acquisition, believability acquisition and skills
development, IT organizations also develop abilities that ease future migrations of
business applications to SaaS:
• Conducting assessments—IT organizations gain experience through contact
with subject-matter experts and stakeholders and incorporate their input into
the assessment. Workflows and activity sequences improve with experience.
The benefit comes when implementing SaaS for business applications, as the IT
organization will understand the majority of its tasks.
• Estimating costs and risks—IT applications do not necessarily have the
same cost and risk profiles as business applications, but assessments of IT
applications still require risk and cost models as bench¬marks for comparison
of future models.
• Defining changes to support processes and the user experience effect—
switching to a SaaS model requires a review of IT support processes around
potential applications. The review allows the IT organization to reevaluate its
support practices and lessen their effect on business users. This will alleviate
user concerns about their support and service levels if applications are replaced
by SaaS.
Figure 2
Sequence of main decisions for assessing potential SaaS applications
Is potential
application available
as SaaS?
Dependency
group can be
separated/migrated
to SaaS?
Is application data
subject to regulation or
security?
SaaS application
matches existing
functionality?
Provider
complies with regulation/
security needs?
Can the functionality
gaps be overcome?
Is data
migration
feasible?
Provider offers
application availability
and performance?
Potential
application
suitable for
SaaS model
Yes
Yes
Yes
No
Yes
Yes Yes
No
Yes
Yes
No
No No
NoNo
No
End of process

Evolving the organization toward STEP
The organizational setup for assessing SaaS for IT applications becomes the
blueprint for STEP. It should reflect the need for cross- pollination of ideas and
a commitment to having the appropriate members. Because IT and business
personnel are likely in short supply, their time must be used carefully. That
means a STEP group has to be agile. PricewaterhouseCoopers proposes such
an organization (Figure 3) with few fixed roles and a majority of ad hoc functions
whose input can be called upon when needed.
The definitive design of STEP will likely incorporate IT leader, ad hoc IT, business
leader and ad hoc business functions. As the assessment progresses, knowledge is
distributed among ad hoc functions. STEP leaders perform coordination and process
improvement functions. Organizing this way makes staff availability more resilient to
change, but close supervision by the leaders and the PMO is required. Otherwise, a
loose structure like this could dilute accountability and command chains.
It is the role of CIOs to sponsor projects and provide vision and guidance. CIOs
articulate STEP team findings and success, helping to extend SaaS initiatives to
business applications and achieving buy-in from executive management.
Figure 3
The assessment organization and basis for a Software as a service Technology Experience Pool
CIO (project sponsor)
Project management office
Business leader
Finance
Legal
Application development
End users
Functional analysis
IT leader
IT vendor management
IT financials
Systems integration
Systems architecture
IT operations
IT infrastructure
Data management
IT security
Ad hoc IT roles
Ad hoc business roles
Fixed roles
Legend

Conclusion
SaaS promises meaningful opportunities for CIOs to lead cost-reduction
initiatives within their organizations. By supporting more manageable and scalable
applications—and shifting from fixed-cost to variable-cost models for the
underlying IT organizations—CIOs can better model future technology services.
Cost-efficient SaaS is realized by thoroughly identifying potential applications and
assessing their characteristics and suitability for SaaS. PricewaterhouseCoopers
recognizes the need for quick results and suggests CIOs blaze the trail to SaaS
with IT applications. In doing so, they enjoy a sandboxed environment while
building critical skills within their teams. They also gain valuable experience that
can then be applied to other areas of the business. And they have the opportunity
to share their SaaS success with corporate management, helping gain buy-in for
business SaaS implementations.

In today’s difficult economy, financial services managers—whether at banks or
other organizations—face new and unprecedented challenges. These managers
must struggle with fewer new checking accounts, fewer trades and transactions,
reduced savings and reduced customer activity in other areas, as well. Ideally,
to control expenses, IT costs for these managers should drop, too. But at many
institutions, that’s simply not the case. Instead, the fixed costs of many IT
resources deliver a double blow: Business is down, yet IT costs remain high.
Many financial services firms accept the idea that business expansion requires a
large increase in IT investment; however, for these same organizations, business
contraction does not yet translate into reduced IT costs. IT assets—including
servers, networks, software and associated real estate and staff—remain fixed
costs. Increases and decreases tend to come in massive step movements rather
than incremental changes. As a result, these IT assets require large upfront
investments that remain on the financial books even during business contractions.
In essence, many financial services firms simply are not geared toward
incrementally increasing or decreasing IT capacity, and related costs, to keep
synchronized with the economy.
IT models: fixed costs vs. variable costs
A different, innovative approach is needed. Financial services managers can
benefit by moving away from the industry’s largely fixed-cost IT model and moving
toward a partially variable-cost strategy. The primary objective of this article is
to review how selectively sourcing one or more subsets of a financial services
company’s IT resources can provide new levels of flexibility. In this way, managers
can incrementally reduce their corporate computing capacities and costs during a
business downturn.
For example, imagine that an investment bank handles four million trade-
execution orders in the first quarter of a year but only two million such orders in
the second quarter. Ideally, as the number of trades is halved, the investment
bank’s IT expenses over these two quarters would be cut by half, too. But since
many banks use a largely fixed-cost IT model, they would be unable to enjoy
these savings.
Controlling IT costs with selective sourcing
Abstract
By leveraging selective outsourcing
and out-tasking opportunities,
financial services managers can
benefit by moving from a largely
fixed-cost IT model toward a
partial variable-cost strategy. The
objective of this article is to review
how selectively sourcing one or
more subsets of a financial services
company’s IT resources may provide
new levels of flexibility. In this
way, managers can incrementally
reduce their companies’ computing
capacities and costs during a
business downturn.
By moving partially to a variable-
cost model, managers can lower
their IT expenses incrementally.
Five variable sourcing options
offer innovative choices in today’s
challenging economy.

Five sourcing opportunities for variable costs
For financial services managers looking to shift their IT costs from a fixed-cost to
a variable-cost structure, PricewaterhouseCoopers has identified five alternative
sourcing strategies. Because each company is different, one size will not fit all.
Instead, managers will need to decide which of the five options are appropriate for
their own situations. Managers also will need to assess the degree to which these
options can be applied to gain the greatest flexibility and savings. Finally, because
these strategies involve maintaining at least some IT operations on a fixed-cost
basis, managers will need to decide what the ideal proportion is.
Real estate
Data centers represent a major fixed cost for many financial services providers.
Companies typically purchase buildings or land on which to build; during a
downturn, these fixed costs cannot be adjusted. Instead, managers may want to
convert a portion of their real estate to leased property, with an option to increase
or decrease space on a quarterly or semi-annual basis.
However, this may not be an ideal strategy during financial booms. When
businesses are growing, a partial-lease strategy could cause real estate costs to
escalate rapidly. So maintaining the ability to switch between fixed and variable
costs—depending on the market climate—is an important aspect of this model.
A financial services firm might secure an option in its leasing agreement to buy
space, thus converting the data center to a fixed-cost basis. Also, a company
considering this model should maintain a base (e.g., 50 percent of its real estate
holdings) that the company owns no matter what. Then the remaining portion of
real estate holdings can fluctuate between fixed and variable costs, depending on
market conditions.
Another option involves working with companies that provide “data center in
a box” services. Essentially, managers subscribe to these services rather than
buy or lease a data center. The services not only contain the data center space
but also the utility and facility infrastructure, including power, cooling and racks.
Managers can use these services in a flexible manner by contracting their
subscriptions, for example, or allowing them to expire during business downturns.
Enterprise license agreements
Commonly, financial services firms sign enterprise license agreements (ELAs) with
their large software vendors. Essentially, these agreements require companies
to pay a fixed license fee for software. If the number of users changes, the fee
remains the same. That’s beneficial during an upturn: As a company expands
dramatically, it still pays the same software-licensing fee. But this arrangement is
unfavorable when financial services firms downsize due to reduced business. In
these cases, companies are unable to reduce their software costs, although they
have lowered the number of users through staff reductions.

For such times, a hierarchical approach may be better. Under this plan,
companies would sign an ELA for a fixed base amount of their software licensing
needs. They would also agree to a variable-cost model, in which the number of
software licenses in use would be recalculated quarterly. In this way, financial
services companies can reduce software-licensing fees when downsizing.
Staff and contractors
Staff represents a large cost component for any financial services IT department.
During an economic downturn, many companies rely on layoffs, attrition and other
forms of downsizing to lower staff costs.
Maintaining a baseline staff, plus a proportion of independent contractors,
provides variability in the staffing model that can be leveraged during business
downturns. The baseline might represent the staffing needed to keep IT systems
operating. Then, as business demand requires, contractors can be brought on or
let go without the need for severance packages and other benefits.
Outsourcing
Many financial services companies outsource some IT functions, such as
networking, desktop support or application development. However, contracts for
these traditional outsourcing arrangements tend to be inflexible. Changes can
involve a good deal of renegotiation, additional fees or a combination of both
renegotiation and extra fees. Yet, during a business downturn, change may be
exactly what smart managers need.
As an alternative, managers may want to consider outsourcing IT complexity
rather than outsourcing IT functions. Companies would continue to provide
the median technology setup needed by most employees, but they also would
outsource more complex technologies required only by a comparatively small
number of employees. This strategy can save money during a business downturn,
as demand for these more complex applications tends to drop during such times.
Limited demand is precisely what makes these more complex applications so
expensive to run. Because complex applications are provided on a limited basis,
financial services companies find it difficult to reach economies of scale.
As an example of outsourcing complexity, managers at investment banks might
outsource low latency trading such as index arbitrage. These applications
essentially trade futures contracts during the fractions of a second—or
hundredths of a second— when the market changes but the underlying stock
prices don’t. Applications able to take advantage of these short-lived trading
opportunities require a good deal of IT infrastructure, and they must be
appropriately tuned at all times. During a business downturn, a financial services
firm’s appetite for operating complex financial products usually diminishes, so this
just becomes a matter of exiting the outsourcing arrangement.

Ideally, such outsourcing arrangements will be made before a vendor is selected
and a contract is signed. According to a 2008 survey conducted by Brown-
Wilson Group, more than 90 percent of outsourcing IT managers said that it’s
vital to have a global outsourcing strategy in place before searching for a vendor.
Amending a contract based on IT complexity is far easier than amending multiple,
function-based contracts. It also allows a company to focus on its competitive
advantage—provisioning financial products. When possible, astute financial
services managers will let others—those with core competency in technology—
provide these complex applications.
Making the transition
Clearly, there is no prescribed path between fixed-cost and variable-cost models.
Therefore, financial services firms that want to move to selective sourcing may
consider turning to consultants or other trusted third-party advisers to help
them pick an effective strategy. For example, as a leading technology consulting
company, PricewaterhouseCoopers conducts workshops that help financial
services managers identify an appropriate path for moving to one or more of the
five variable-cost models.

Overview
Service-oriented architecture (SOA)-supported enterprises are based on
ecosystems of business functions that represent the interests and needs
of multiple lines of business (LOB). However, many responsibilities must be
determined before implementing SOA. For example, who should pay for
certain services? Or how are fees broken down if services are reused across an
enterprise? How should SOA projects be financed? The current financial model
in many companies is appropriate only in budgeting for vertical LOB applications.
SOA requires a different vision and execution plan. To achieve success in SOA
initiatives, organizations will have to adopt a strategic approach that better
realizes the associated financial benefits.
SOA and agility
Understanding why financing SOA initiatives is different from the traditional IT
finance model requires a look at what actually encompasses an SOA-supported
enterprise. There are five basic types of “encounters” with SOA (Figure 1):
• SOA encounter of the first kind. At this maturity level—which is where many
large organizations reside today—the corporate IT group simply implements
reusable components across its enterprise services. This level is essentially
an iteration of enterprise application integration (EAI), making it easy to build a
business case and secure financing for its implementation. At this stage, the
investment into new licenses and skills is relatively small, and IT can provide the
business with the ability to reuse functionality in real time. Business users receive
information and execute functionality only when required—limiting the reuse of
functionality between multiple applications. A key benefit of the SOA encounter of
the first kind is that it drives down the costs of IT support. The real-time or near-
real-time integration allows for spreading the processing loads over extended
periods, which improves infrastructure utilization. The benefit is particularly
marked when compared with the batch processes traditionally used for EAI that
require intensive use of infrastructure during the batch processes. In many cases,
the batch processes leave resources under utilized during normal work hours.
Consequently, IT must over- procure infrastructure and computing hardware for
those few hours of batch processing. The rest of the time, those resources sit idle.
• SOA encounter of the second kind. Companies successful with the previous
encounter often find themselves at this next level. With encouraging results,
organizations attempt to expose more services to SOA processes. However,
these services might have little enterprise reuse if IT does not spend enough
time understanding the business operations of the enterprise. The corporation
might then decide to halt its SOA efforts because it lacks a clear view of
business needs and budgetary requirements.
• SOA encounter of the third kind. This level is an attempt to increase the agility
of existing business applications. To support changing business processes,
existing application components are exposed as services and are resequenced
to fulfill new business needs (e.g., composite applications). This level supports
reusing what already exists and reducing investments in new solutions.
Data for decision
Abstract
Service-oriented architecture (SOA)
implementations are showing
companies that the traditional
corporate IT model needs a
makeover. Today, many corporate
IT models simply cannot address
the financial requirements of an
SOA- supported enterprise. Not
surprisingly, IT organizations are
eagerly embracing the opportunities
offered by SOA. The evolution of
SOA transforms it from being a static
function into a more agile, service-
based business function. That is an
important competitive differentiator.
However, it means tweaking the
corporate IT model to realize the
benefits of SOA.
Creating financial flexibility through
service-oriented architecture

Unfortunately, traditional applications are created with hard coded component
interactions. Simply exposing these components as services does not make
them ready for SOA. SOA lies on loosely coupled components—so components
should exist and work independently. Resequencing outside the application’s
core requires complete reverse engineering and redesign, which can be as
expensive as coding a new application. While it’s a good idea to have composite
applications, trying to create composite applications on top of existing
applications is questionable. The more resequencing needed, the more expensive
the exercise becomes. Chief information officers and architecture teams will have
a tough time building a sound business case for this level of SOA.
• SOA encounter of the fourth kind. This level, also known as SOA 2.0, brings
EAI to the next level. It adds intelligence in the way applications interact with
each other. Because event-driven architecture1
(EDA) improves the workflow
of value chains, the enterprise gains a reliable and agile integration framework
that provides incremental value to processes that span multiple departments,
such as sales or fulfillment. When executing a customer order involves multiple
organizations, EAI is not enough. The entire fulfillment process must be monitored
and pushed for optimal completion. With many companies organized into value
chains, few have a single group responsible for the overall value chain. But EDA
can be used to enforce proper execution flow for horizontal processes. EDA can
also be used to monitor—and in some cases enforce—broad and asynchronous
processes, raising alarms if something goes wrong and providing executive
managers with information on overall performance. Companies can achieve the
same results without using a workflow, but that means IT must hard code the
1
This approach is called event-driven architecture because a particular business event starts a value chain. For example, a customer
request for a loan starts a fulfillment process.
Figure 1
Maturity scale for service-oriented architecture (SOA)
Create a library of elementary
business processes and
implement business rules engines
to increase business agility
The ideal of SOA for
organizations
Agile
enterprise
Increasinglyagileenterprise
Expose existing application
components as services and
resequence them to fulfill new
business needs
Reuse of application
components to fulfill
new business needs
Composite
applications
Add intelligence to interactions
between applications by
improving workflows
SOA 2.0 or event-
driven architecture
Event-driven
Increase the integration level
and evolve the previous SOA
implementation
Increased number of
services exposed
through SOA
Service
integration
Centralized governance and
IT management
Reusable components
implemented across
enterprise services
Component
integration

business logic, which limits the organization’s ability to implement changes at the
speed required to keep up with dynamic market conditions.
• SOA encounter of the fifth kind. This level represents what organizations should
strive to achieve with SOA. This case extends resequencing, which hypothesizes
that any business can be decomposed into smaller business processes. Steps
in business flows located between decision points are called elementary
business processes (EBPs). In many cases, the difference between how business
works today and how it will work tomorrow lies in the order in which EBPs are
sequenced rather than in the internal workings of each EBP. A library of EBPs and
a business rules engine (workflow) allows the implementation of sequencing and
provides business agility. Agility is supported for vertical organizations (business
units) and for horizontal processes (integration within value chains). While this
concept is not new, what has been missing until the year 2000 is enterprise-
class business rules engines. Adopting this approach promises changes in
how applications are designed and created and in the governance of corporate
IT organizations. But IT will be challenged to build reusable agile business
frameworks. Only extensively reusable frameworks can provide the ability to react
swiftly to changing market needs.
Financial effects of SOA
At high levels of maturity, changes required by business units are largely
concentrated around changing the order of execution of elementary business
processes rather than the processes themselves. This concept transforms
the roles of IT and the business units in the software development cycle and
significantly affects finances. Typically, traditional software development follows a
recurring pattern:
1. Business requirements are developed by the business community and then
translated into
2. Technical architecture for this particular business domain, which is translated into
3. Unified Modeling Language design, which is translated into
4. Detailed code design, which is translated into
5. Code
Too many translations significantly increase the probability of coding errors.
Improving the process requires direct translation of the business architecture into
executable code artifacts, which increases the level of modeling abstraction. The
concept is simple: the higher the level of abstraction, the closer we are to the real
human language of business design. That means spending less time developing
applications. Higher levels of abstraction provide these benefits:
• Greater developer productivity
• Fewer moving parts, which reduces the possibility of coding errors and
concomitant quality assurance efforts
• Fewer coding errors resulting from multiple levels of business architecture
translations
• Increased precision for future development efforts
• Shortened development times and reduced emergency business changes on
ongoing projects
The introduction of the business
process execution language for
process design and rules engine
environments provides business
users with a human¬like language
for designing applications from
end-to-end. As the level of
abstraction increases significantly,
productivity increases as well—up
to six times when compared with
the traditional model used today.

Enterprise-class business rules engines provide a high-generation language
used for resequencing EBPs, supporting the benefits of the appropriate level of
modeling abstraction. In this new environment, IT focuses on the following:
• Providing and supporting a computing infrastructure (i.e., databases, servers,
networks) independent of application functionality
• Managing the business rules engine
• Creating a library of EBPs and maintaining underlying application services
The new roles of the business will include:
• Defining business products, business process flows and business rules
• Identifying gaps in the library of EBPs and issuing the corresponding
requirements to IT
• Using the business rules engine to configure designed functionality
• Running simulations and publishing the configuration into production
These changes drive significant financial benefits, making software development
costs variable based on the level of business activity and more predictable:
• Over time, IT will amass EBPs, causing a decline in the demand for new ones
• Slow business cycles translate into slackened demand for new EBPs, reducing
software development efforts
• More accurate estimates of the development efforts for new EBPs
• IT can calculate the number of CPU cycles required by each EBP and its other
attributes, making chargebacks easier
• Available information on how much an EBP is used and by whom allows IT to
attach a dollar value to it and issue chargebacks for its use
• EBPs are simple components, helping increase quality and reducing time and
efforts for testing and fixing coding errors
• The overall enterprise architecture is simplified because EBPs (and their related
application domains) are independent from the integration framework, reducing
complexity and associated costs
Conclusion
The introduction of the business process execution language for process design
and rules engine environments provides business users with a human-like
language for designing applications from end-to-end. As the level of abstraction
increases significantly, productivity increases as well—up to six times when
compared with the traditional model used today. Over time, as EBPs increase,
users will rely less on IT for this work-product. The result will be a well- performing
SOA organization where IT more closely resembles assembly-line management.

The need to transform the desktop cost model
Delivering desktop environments to large, distributed user bases is challenging
and costly. But delivery is just one of today’s many desktop management cost
drivers. Others include maintenance, support and depreciation of desktop personal
computers (PCs); testing and deployment of new application releases; patches and
upgrades to operating systems (OSes); and desktop repairs. With more workers
demanding higher-quality support services, coupled with growing security and
privacy concerns, it is not surprising desktop expenses keep increasing.
However, combinations of existing and breakthrough technologies are
transforming the desktop cost model (Figure 1). New solutions—such as desktop
virtualization, application streaming and blade PCs— promise significant
change in the way IT organizations deliver and support desktops. They also help
transform the desktop cost model by:
• Lowering hardware and software expenses through better resource utilization
• Increasing operational efficiency in day-to-day desktop management
• Delivering more choices in technologies that can offer hardware, OSes and
applications separately, leading to more options for greater flexibility and
cost efficiency
• Reducing the risk of data loss
• Decreasing the risk of privacy and security breaches
The evolution of desktops and users
Today, desktop computers are the primary interfaces between IT organizations
and end users. The experiences office workers— who spend so much time at
their workstations—have using PCs strongly influences how they perceive their IT
organizations. As employees’ working habits change, their expectations of IT and
desktop requirements change, too:
• They covet mobility options, even if they travel infrequently
• They want always-on connectivity or at least the ability to access applications
and data when offline
• They desire access to personalized desktop environments from anywhere,
including their homes
While users lead the major shifts in desktop requirements, rapidly changing
business drivers also are producing:
• More reliance on outsourced labor that requires cost-effective ways to deliver
standardized, secure desktops and applications to remote, third-party users.
• Workforce reductions that require rapid provisioning and decommissioning of
desktop computers.
• Mergers and acquisitions that require integration of different desktop standards,
vendors and support models.
Below the code
Abstract
As demands from technology-savvy
employees increase, companies
are finding it more difficult to cut
desktop management costs while
keeping users satisfied. Now, the
convergence of
new technologies and associated
management processes is helping
transform desktop costs into variable
costs. The benefits of this approach
include meeting user expectations
while leveraging data center and IT
investments. This article examines
current desktop support practices
and the factors transforming them
into an on-demand, variable-cost
model service. It also provides
strategic recommendations on
achieving these goals.
Easing desktop management costs
with innovative new technologies

• Major OS upgrades and application releases with stringent hardware
requirements, which pose complex problems for IT organizations. Decoupling
hardware, OSes and applications is a creative way of tackling these major
undertakings without having to redesign the entire desktop stack.
These challenges and demands are pulling IT organizations in two directions.
First, they must manage the need for mobile, distributed desktops that access
applications and data from the Internet or network cloud. Second, they must
centralize control and security by confining computing and data resources within
a secure data center.
Approaches to transforming the desktop cost model
New technology enablers and low-cost solutions can help IT organizations
address changing business drivers and cut costs from their desktop operations.
First, look for cost takeouts. Later, consider reaching for more ambitious savings
and disruptive cost models.
There are many way to transform the desktop cost model:
Reducing capital costs and depreciation. Reducing the company’s overall
desktop inventory, extending the life of equipment or both helps reduce capital
costs and depreciation.
As we near the reality of U.S.$300 laptops, replacing expensive laptops with these
less-costly—but also less-powerful—variations is increasingly attractive. Low-cost
units will not suit all users and workloads, but they provide a novel approach to
shifting capital costs.
Reducing operational costs. IT organizations strive to standardize desktop OS
images, automate software patching and streamline desktop management. Today,
though, technologies provide innovative ways to deploy a standard OS by cloning
virtual images or other options. Technologies also provide ways to develop, test,
Figure 1
Information Technology Infrastructure Library-based desktop cost model
Desktop technologies
Hardware Software Staffing Accommodation Transfer
• Servers
• Storage
• Network
• Workstations
• Operating
systems
• Applications
• Connection
brokers
• Management
tools
• Service desk
• Deskside
support
• L2/L3 support
• Warehouse/
storage of
inventory
• Outsourced
desktop
support
• Vendor
support and
maintenance
Cost elements
Direct costs Indirect costs absorbed by service Unabsorbed indirect costs
• Workstation hardware
• Desktop software licenses
• Back-end license costs
(physical and virtual)
• Deskside support
• Service desk
• L2/L3 support
• Data center power/cooling
• Network
• Facilities

package and distribute software. One example is through application streaming
over the network.
These approaches create operational efficiency and affect software license costs.
Previously, licensing options included installing the application and paying for
licenses for each desktop or installing it only as users requested it. Now, a third
option is available: as-needed application streaming. This new approach makes
better use of license fees and allows IT organizations to negotiate licenses. Rising
license and support costs also can be mitigated with open source software.
Enterprise-level open source software has been anticipated for years, but as
adoption rates have gone up among consumers, a phased-in introduction of
alternatives like open source office productivity suites or operating systems is no
longer far-fetched.
Moving to variable-cost models. Desktop virtualization and application
streaming technologies support many capabilities. One is the pooling of desktop
images, in which users share limited numbers of images. Once a user logs off, the
image is put back in the pool, optionally scrubbed clean and made available for
the next user. In companies where users have varying work hours, it is possible to
create less than one image per user. That translates into reductions in storage and
possibly software licenses through:
• On-demand creation of new desktop OS images from a template—also referred
to as cloning—which means desktop pool costs can shrink or expand based on
user needs
• Legacy technologies, such as terminal servers, that allow multiple user sessions
to share computing resources on a single server
With these capabilities, the desktop is becoming a good pay-as-you-go
candidate, where users are charged—symbolically or literally—based on actual
computer use. When a user needs a desktop, it can be provided on demand and
priced accordingly, without the need for heavy upfront capital costs. This makes
company desktop costs more predictable.
Also, the desktop-as-a-service model is emerging. Hosting companies are
entering the space and building on the pervasiveness of the on-demand
infrastructure to provide desktops on demand.
Reducing exposure to risk. Moving to a centrally managed desktop model
reduces security and privacy risks associated with large distributed pools of
workstations containing business and customer data.
By moving data, OSes and applications back into a corporate data center,
companies can better leverage standard security policies, backup mechanisms and
even disaster recovery plans. Financially, a centrally managed desktop model might
mitigate the risk of multimillion-dollar fines for loss of sensitive data. It also can
increase employee productivity through improved data and desktop availability.
The next-generation desktop
In the past, concepts like diskless workstations and network terminals promised
radical new ways of addressing desktop challenges. But they relied heavily on IT
elements like network bandwidth and servers. They also had short- comings in
processing power, limiting their applicability to narrow, specific uses.
Today, however, a next-generation desktop (Figure 2) can help deliver the right
desktop experience to the right user in a cost-efficient manner. Virtualization

technologies, development of streaming technologies that reduce bandwidth
requirements and the emergence of new challenges to desktop support
organizations all play important roles. The mission is accomplished using a
combination of technologies, tools and cost modeling:
• Technologies that deliver OSes, applications, user data and settings to users
provide the means for aligning user experiences with current expectations in
areas like personalization.
• Streamlined management tools and processes deliver and support next.
While many are already in place—at various levels of maturity—the heavy
reliance on data center and network services to deliver the next-generation
desktop means the desktop support model will have to cross an IT organization’s
functional boundaries.
• A costing model looks at sourcing options, chargeback models and technology
solutions, favorably shifting the balance of fixed vs. variable costs.
Technology framework
The technology framework for a next-generation desktop (Figure 3) must include
an end-user-facing access device. There are many from which to choose:
Physical workstations are composed of traditional laptops and desktops running
OSes and applications that meet company policies and standards. Endpoint
virtualization applies traditional virtualization techniques to run multiple isolated
OSes on a single workstation. This can either replace multiple workstations for
users who need more than one or provide added security.
Thin clients are simplified workstations with limited computing power. They can
display remote sessions of OSes running on a central server and transmit user
Figure 2
Examples of next-generation desktop technology solutions
Application Application Application
Operating system Operating system Operating system
Operating system
Connection software
Blade PC Hypervisor
Server
Client PC/thin client
Application Application Application
Datacenter
Blade PC
Internet connection
End user
Desktop 1 Desktop 2
Virtual desktop pool
Office/remotelocation

mouse or keyboard inputs back to the server. Typically, thin clients feature a
minimalistic hardware configuration.
Unmanaged or light endpoints can be inexpensive or ultraportable laptops used for
providing a full desktop experience for mobile users with limited computing needs.
They also can act as mobile thin clients. This category includes noncorporate assets
such as PCs owned by employees or firms contracted by the organization.
Unmanaged endpoints can be used in combination with portable media, such as
standard or virtualized OSes stored on CD-ROMs or USB devices, which deliver
the desktop on inexpensive media and require less workstation management.
Mobile devices are typically personal digital assistants or smartphones. They often
lack the processing power to run resource-heavyOSes and applications but can be
used as remote clients to software components running on a central server.
Various techniques can help deliver OSes and applications to any of these access
devices:
• Remote display and application virtualization—OSes or applications run on a
central server and are remotely displayed on the user’s access device
• Streaming—OSes or applications are delivered over the network but run locally
on the access device as soon as it has received enough data to start them
• Traditional application access methods—these include web front-ends, such
as AJAX and other Web 2.0 technologies, and legacy fat clients
Connection brokers are important middleware components that help connect
users to remote desktops. Typical uses include user session management,
entitlements and access rules, and handling virtual desktop pools, which use
existing data center technologies to store OSes and applications in a secure,
central facility. Technologies that can be used include physical servers, often
through blade PCs (designed to provide workstation-grade computing power in
a single, easy-to-provision blade), virtual machines hosted on physical servers,
Figure 3
Next-generation desktop technology framework
Thin clients
Keyboard, screen, mouse
Physical workstations
Remote display
OS streaming and
network boot
Virtual OS images
Storage of OS images
FC SAN, Infiniband, etc.
Desktop onportable media
USB drive, CD-ROM, etc.
Mobile devices
Smartphones, ultraportable laptops
Application streaming
Connection broker
Shared or multisession
servers
Storage of user profiles and data
NAS, CAS, collaboration tools
Unmanaged/light endpoints
Budget laptops, non-company owned
Application access
HTTP/AJAX, fat clients, etc.
Application virtualization
Blade PCs
AccessDistributionInfrastructure

and multi-session servers such as terminal servers that support the delivery of
applications to multiple concurrent users.
Storage challenges in these various delivery methods include the location for user
profiles (standard or roaming) and user data (e.g., using network-attached storage)
but focus largely on the storage required for handling hundreds or thousands
of OS images. Back-end storage deduplication or software solutions are often
prerequisites to large-scale virtual desktop deployments.
Desktop management
Desktop management remains a source of both challenge and prohibitive cost.
But by introducing technologies like those described earlier, companies can get
closer to achieving operational efficiency and related savings, including:
• Inventory reduction and improved inventory management through discovery or
management tools used in the data center today (e.g., the IT organization now can
use its chosen virtual machine management tool to track virtual desktop OSes).
• Provisioning and decommissioning through data center automation and
simplified end-user devices that can be installed by nontechnical staff.
• Software distribution, through application virtualization and streaming, that helps
companies deploy software. It also helps push software updates transparently,
with the new version being deployed as users access it.
• Asset management, through better use of software licenses, increased sharing
of infrastructure—such as higher utilization of storage and computing via
virtualization—and leveraging existing tracking tools to facilitate metrics.
• Break/fix can now be rolled into existing data center support processes (e.g.,
with virtual server provisioning, users can reimage a failed desktop OS). Moving
to simpler end-user devices also can help mitigate hardware failures.
Conclusion
Many benefits are rising up through next-generation desktop technologies. As
organizations struggle to become more agile while controlling costs, they need
technologies that can transform desktop delivery and user experiences. But no
single technology is a one-size-fits-all solution. Virtual desktops running on thin
clients may be a good fit for an outsourced call center, but road-warrior executives
might require a different solution. As with any technology introduction, the key lies in
a phased-in approach. In this case, the approach includes three important steps:
• Understanding end-user behaviors, computing needs and service expectations
• Defining a strategy and associated technology support for delivering a desktop
service that meets end-user requirements
• Mapping client requirements, support needs and cost objectives for each
desktop service to an appropriate mix of technology solutions
Successfully executed, this phased-in approach to next-generation desktop
technology can significantly change how workstation functionality is delivered
while increasing its cost-effectiveness for your organization.
Many benefits are rising up
through next-generation desktop
technologies. As organizations
struggle to become more agile
while controlling costs, they
need technologies that can
trans¬form desktop delivery and
user experiences. But no single
technology is a one-size-fits-all
solution.

Cloud computing: an emerging technology
In today’s economic slowdown, financial services firms are reducing costs to
improve their bottom lines. These measures include IT budget cuts at a time when
IT departments face increasing demands for improved reliability and performance,
quicker deployments and more transparent financial reporting.
Before the current economic rough patch, the trend for media-rich applications
and analytics engines resulted in a proliferation of different server and storage
platforms in enterprise data centers. On average, large companies are
experiencing a twofold increase in their server, network and storage resources
every one to five years to meet rising demands. Consequently, IT organizations,
particularly those that resulted from mergers or acquisitions, manage varied
and complex technology portfolios that are riddled with operational inefficiency,
platform fragmentations and a lack of process and tool integration across the IT
infrastructure. This complexity increases application failure rates1
and does not
support rapid deployment times or cost transparency. Companies with complex,
fragmented environments face monumental challenges to improve application
performance and reliability during periods of varying demand and adopt a
chargeback structure based on resource consumption.
On the other hand, mature IT organizations regularly assess emerging
technologies and their abilities to address these challenges. In larger companies,
architecture groups typically monitor and experiment with new technologies,
usually on a no-cost basis negotiated with product vendors. Other companies
maintain a watch list of technologies and monitor the market through analyst
reports and other pertinent data. They use a filtering function that analyzes market
information and weighs their company’s readiness to adopt technologies based
on several criteria.
Although they are still maturing, cloud computing and the software as a service
(SaaS) business model are currently grabbing the industry’s attention and setting
off alarms throughout corporate filtering functions.
SaaS is a new business model for deploying applications as Web-based services
hosted by a service provider. A critical support of SaaS is cloud computing, which
is characterized by its ability to split application workloads across shared pools
of resources, support fast provisioning, provide on-demand application scaling
and offer billing based on resource consumption. Companies that leverage SaaS
and cloud computing don’t have to invest in in-house applications hosting. These
may help them significantly reduce costs and shift their IT cost structures from a
fixed-cost model to a variable-cost model. IT organizations should consider this
innovative approach to their IT challenges and help meet increasing user demands.
Supporting next-generation, consumption-based hosting services
with cloud computing
Abstract
Cloud computing is a key enabler
of the software-as-a-service model
that supports the deployment of
applications as Web-based services
hosted in “clouds” of technology
assets (e.g., servers, storage,
network devices). The open secret
of cloud computing is its ability to
split application workloads across
shared resources, which can be
geographically dispersed, while
supporting application performance
and reliability as computing demand
fluctuates. Cloud computing also
eliminates the need to invest in
hardware, software or facilities
necessary to run applications and
supports billing for services on a pay-
as-you-go basis.
This article reviews the technologies
behind cloud computing, the
current industry service model and
its limitations. Then it discusses
current trends and capabilities in
cloud computing. Today’s risks and
benefits are also detailed. Finally, it
discusses the disruptive nature of
cloud computing and its potential
effects on the cost structures of IT
organizations and enterprise
data centers.
The secret to cloud computing
is its ability to split application
workloads across shared pools of
geographically dispersed resources
while maintaining application
performance and reliability as
demand for computing resources
fluctuates.
1
Industry figures suggest that 80 percent of infrastructure failures are caused by human error, often due to fragmented and
siloed change management processes (e.g., one isolated process for each technology domain that does not take into account
dependencies on other domains) and an overall lack of standardized operating environments.

Utility computing: the technology behind cloud computing
The infrastructure behind cloud computing is managed
through utility computing technologies. Utility computing
supports dynamic provisioning and repurposing of server,
network, storage and desktop resources. Configuration files
for applications, operating systems (OSes) and servers are
packaged into images or instances and maintained in data
stores. The images can be dynamically loaded or unloaded
based on predefined policies for demand, time, events,
priorities or other criteria. Policies also define boundary
values for the operating environment (i.e., maximum RAM in
use, maximum CPU in use). In addition, utility computing can
be used to manage physical and virtual servers.
Utility computing dynamically scales applications, on demand,
across shared infrastructures. This eliminates the need to
over-provision resources to keep up with processing needs
during peak use. It also allows idle servers to be used for
other purposes or powered off to reduce overall data center
energy usage. For example, utility computing lets companies
use servers dedicated to disaster recovery for development
or testing during normal operations. Because applications
are dynamically resized and varying levels of resources are
allocated at different times, utility computing allows computing
and storage resources to be billed based on consumption. In
addition, utility computing can help companies:
• Reduce costs through efficient use of resources, greater
control over hardware inventory growth, lower staffing
needs and reduced complexity due to lower power
consumption and improved process automation, with the
added benefit of fewer opportunities for human error
• Improve quality of service through infrastructure
provisioning automation, driven by business policies and
simplified capacity allocation, based on demand, to meet
varying computing loads
• Increase agility by standardizing operations and
infrastructures to allow rapid service provisioning and
scaling to meet demand
The goal of utility computing is to rearrange infrastructure
resources currently locked in silos across business lines
and environments (e.g., development, testing, disaster
recovery) into a pool of resources that are automatically
allocated and deployed based on business policies and
service requirements.
Implementing utility computing can be challenging
because of the requirements for integrated real-time
operations and appropriate instrumentation. Figure 1
identifies the key instrumentation functions required to
support utility computing.
Figure 1
Key instrumentation functions required for utility computing
Workload management
Self-heal by monitoring
and adding resources
dynamically as required
Dynamic workload
management
Charge based on usage
and allocation
Metering
Resource
management
Provisioning
(including software
distribution)
Build, update and
patch all software
components
Server provisioning and configuration
Audit and remediate
for greater
standards
compliance
Fault monitoring
Monitor faults
from a business
perspective
Resource management,
monitoring and fault
management
Discovery
Configuration
Fault monitoring
Resource monitoring
Provisioning
(includingsoftwaredistribution)
Audit
Resource management
Metering
Workload management
Utility computing
instrumentation
Assign CPU,
memory, disk, I/O,
licenses and power
Configuration
Resource
monitoring
Asset management
Discover and track
all hardware and
software assets
Configure
all hardware
and software
components
Configuration
management
Audit

Instrumentation tools related to resource management,
monitoring and fault management support a standardized
framework for real-time monitoring and management. Key
features include:
• Integrated infrastructure discovery capabilities and
configuration management database linked to asset
management and service-level management tools
• Management and monitoring of resources (e.g., CPU,
memory, input/ output, storage, licenses, power)
• Fault and change detection and reporting in real time
• Application health and resource utilization monitoring
Instrumentation tools for server provisioning and
configuration are primarily used to create images that
combine baseline application and server configuration files
and deploy those applications in appropriate computing
environments. Some vendors refer to baseline images as
distributions. These tools include a highly standardized suite
of processes around configuration and change management
that allows applications to be provisioned within minutes.
Key benefits of server provisioning and configuration include:
• Automated change management and control, including
configuration management, patch management, application
packaging and rollbacks, promotion and deployments
• Configuration compliance through enforcement of
policy-based audits of infrastructure resources and
application configurations
• Bare metal server provisioning
• Software distribution, patching and updates
• Automated deployment and configuration of OSes
• Application release management, including application
discovery, configuration and certification of software,
creation and packaging of images, application promotion,
and deployment
• Job scheduling
• Physical and virtual server management
Instrumentation tools for dynamic workload management
allow rapid provisioning of applications across shared
servers in response to load conditions, failovers and
business policies. Key characteristics include:
• Horizontal scaling, adding or subtracting an entirely new
server image based on demand to meet service levels
• Capturing and hosting application images
• Delivering high-availability capabilities through rapid
reconfiguration of available resources
• Starting and stopping applications
• Monitoring and measuring resource usage of applications
and services and performance metrics against business
policies and service levels
• Harvesting of servers (dynamically turning off lower-priority
applications and repurposing servers for higher-priority
applications) when no more capacity is available
• Powering servers up and down based on power
management policies (i.e., time, event, capacity, demand)
• Bare metal provisioning
• Physical and virtual server management
Depending on the solutions selected, functionality
overlaps may exist between tools for server provisioning
and dynamic workload management, particularly in
application packaging, bare metal provisioning and physical
and virtual server management.
Figure 2 illustrates the roles of server provisioning and
dynamic workload management tools throughout an
application’s life cycle.
Figure 2
Functional roles of server provisioning and workload management tools
Preproduction
Server provisioning
and configuration
• Bare metal
provisioning
• Software and
patch distribution
• Configuration
management
• Application
Production
Dynamic workload
management
• Image capture
• Dynamic
provisioning
• Usage metering
• Server harvesting
• Application
Decommission
Dynamic workload
management
• Server shutdown
• Removal from
pool
Server provisioning
and configuration
• Change control
• Configuration
Cloud computing service models
From a service delivery perspective, cloud computing
service models (Figure 3) consist of the following elements:
1. A service catalog is presented to users when they access
a cloud computing provider’s portal. Users can request
predefined services, such as enterprise e-mail, and select
business policies and service tiers.
2. System and capacity management tools interpret user
requests and confirm available resources.
3. Capacity management tools invoke dynamic server
workload tools to provision and activate predefined
packaged images within identified server resources.

4. Dynamic workload management tools reconfigure infrastructures to meet
varying demands and business policies.
5. Monitoring tools verify that services perform according to committed service
levels and track usage.
Candidate applications for cloud computing
Applications best suited for cloud computing are not dependent on OSes or
hardware. Examples include Web-based applications that can scale horizontally
by relying on web farms that have front-end load balancers. Additional servers
help meet increased service demands. When demand decreases, infrastructures
can scale down by turning off servers.
Current cloud computing services
Key attributes of current services offered by cloud computing providers include:
• Pay based on resource consumption (i.e., CPU hours, amount of bandwidth
used) rather than a fixed cost for servers
• No long-term contracts—some vendors require only one hour of commitment
Cloud computing is compelling because costs related to computing resources are
no longer fixed but vary according to consumption.
Pricing models are still evolving, and many different types may emerge. For
example, pricing for e-mail appears to be moving toward a fixed charge per user
with discounts available based on volume.
However, the pricing model for Web-based business applications, which will likely
become the most popular services of cloud computing providers, may have fixed-
price components and variable-cost elements similar to the access charges and
traditional minutes used in cell phone rate plans. According to research, prices for
Web-based applications are just a fraction—typically 25 percent—of what it would
cost to run those applications in an enterprise data center.
Application hosting services have been available commercially since the advent of
the Internet; cloud computing is simply the most recent incarnation. Historically,
Figure 3
Cloud computing service architecture
User portal System and
capacity
management
Dynamic workload
management
Cloud of
servers
Build image Server provisioning
and configuration
Service
catalog
Select predefined
service and policies
Application
monitoring and
metering
Certification and
image packaging
process
New
application

Internet service providers branched into collocation or application service provider
services and launched cloud computing services. New players in the cloud
computing market include Google, Microsoft and Yahoo.
Today’s market offerings do not yet align with cloud computing’s vision. In fact,
current cloud computing service offerings present significant risks:
• Service-level agreements (SLAs) are not yet in line with enterprise-class SLAs
for many cloud computing providers.
• Most cloud computing providers do not offer consistent support to help
companies design application architectures, define sizing requirements or
configure business policies.
• To keep costs down, most cloud computing providers leverage open source
software. Linux applications are the only ones that work well in this environment,
while Windows- and UNIX-based applications have limited support.
• Licensing in a virtualized and dynamic workload management environment has
not been adequately addressed.
• Most companies are not yet confident with the level of privacy and security
offered for their data.
• Some European countries require financial data to be hosted in-country, a
difficult compliance requirement for cloud computing because it is essentially
geography independent. However, some providers are deploying services with
built-in geographic constraints.
• Most cloud computing providers do not allow security monitoring.
• Most cloud computing providers are not compliant with the Payment Card
Industry standards and cannot process credit card payments.
• Porting existing enterprise applications is difficult due to their dependencies
on middleware, hardware and software residing within corporate data centers.
Currently, most major business applications would not qualify as candidates
to be hosted with cloud computing providers unless their architectures and
dependencies could be simplified.
Cloud computing providers
There are five types of cloud computing providers:
• Traditional application service providers (e.g., Akamai, Amazon, Salesforce.com)
• Hosting providers (e.g., Layered Technologies, Terremark, XCalibre)
• Large integrated companies (e.g., IBM, Sun Microsystems)
• Web giants (e.g., Google, Microsoft, Yahoo)
• New entrants (e.g., iCloud, Mosso, Zimdesk)
A representative example of today’s cloud computing service model is illustrated
by Amazon.2
It offers two key services—Elastic Compute Cloud (EC2
) and Simple
Storage Service. Amazon’s customers can create from scratch or select from a
virtual server template which images to deploy. The company uses a third party,
RightScale,3
to help customers port or create web applications to run on EC2
.
Application hosting services have
been available commercially since
the advent of the Internet; cloud
computing is simply the most
recent incarnation.
2
Amazon.com, Amazon Elastic Compute Cloud (Amazon EC2), beta available at http://www.amazon.com/gp/browse.
html?node=201590011, retrieved August 14, 2008.
3
RightScale, Cloud Management Platform and Support, retrieved from http://www.rightscale.com, Oct. 27, 2008.

Although the industry is still in its infancy, Google, Microsoft and Yahoo—which
have vast data center infrastructures that they can leverage into grid computing—
as well as other players plan to enter the market soon. This could change the data
center computing paradigm dramatically.
For example, today Google offers e-mail and collaboration services as part of
its Google Apps service.4
It includes Web-based e-mail (with 25 gigabytes of
storage per account), calendaring, document management and other collaborative
services for U.S.$50 per user per year, with 24/7 support. These costs are
significantly lower than the costs of an enterprise mailbox. In addition, Google has
previewed its Google App Engine,5
which allows developers to develop and test
new applications hosted with Google’s cloud computing infrastructure, offering a
development sandbox and publishing functions to deploy newly developed web
applications onto Google’s servers.
Conclusion: creating a paradigm shift for IT and financial services companies
Today, cloud computing services and their infrastructure technologies are still
maturing and considered too risky for major business applications. However,
financial services companies can benefit from an increasing array of cloud computing
services, including enterprise services like e-mail and Web- based collaborative and
hosting services for nonmission-critical applications—such as instant messaging and
Linux-based J2EE applications—at a fraction of the cost of enterprise data centers.
The more commoditized and standardized the applications, the more appropriate
it is to use cloud computing to host them instead of a corporate data center. In
addition, cloud computing is better suited for:
• Low-priority business applications for data mining and/or that have few if any
hardware or software dependencies
• Hosting facilities for development and testing, leveraged during times of
increased demand such as marketing campaigns, or when enterprise computing
resources are unavailable
• Running product pilot projects without committing to purchase servers until the
tested products or applications prove viable
Until cloud computing matures, it will continue to be a disruptive force and can
change computing and enterprise data center paradigms in significant ways.
In the future, companies may require little or no infrastructure to run corporate
applications, reaping the benefits of a variable-cost structure for IT and increased
levels of responsiveness to fluctuating computing demands.
This paradigm shift began when web conferencing applications such as Webex
were introduced. Because applications no longer have to be hosted behind
enterprise firewalls, companies are reluctantly accepting the fact that an increasing
number of managed applications can run outside their corporate infrastructures.
As cost advantages increase to favor cloud computing, the relevancy of enterprise
data centers6
may come into question. Finally, the consumption- based cost
approach inherent in cloud computing will support cost transparencies demanded
by business users. These transparencies and the significant cost benefits of using
cloud computing will help build support among organizations.
4
Google, Welcome to Google Apps: choose the edition that fits your needs, retrieved from http://www.google.com/a/help/intl/en/admins/
editions.html, August 14, 2008.
5
Google, What is Google App Engine?, retrieved from http://code.google.com/appengine/docs/whatisgoogleappengine.html, August 14, 2008.
6
A twist to the impending vanishing physical data center paradigm is the potential for companies to run their virtual data centers in a cloud.

Contacts
Julien Courbe
Managing Director
Tel: (646) 471-4771
julien.courbe@us.pwc.com
pwc.com
© 2010 PricewaterhouseCoopers LLP. All rights reserved. “PricewaterhouseCoopers” refers to PricewaterhouseCoopers LLP (US), a Delaware limited liability partnership, or, as the context
requires, the PricewaterhouseCoopers global network or other member firms of the network, each of which is a separate and independent legal entity. This document is for general information
purposes only, and should not be used as a substitute for consultation with professional advisors. DH-10-0217

Fs tech-journal-cost-management

More Related Content

What's hot (20)

Similar to Fs tech-journal-cost-management (20)

More from Karthik Arumugham (20)

Recently uploaded (20)

Fs tech-journal-cost-management