Cloud Computing: an

Overview

Executive Summary

An increasing number of businesses are discovering that cloud computing can save them significant
operating and capital expenses. However, cost is not the only reason companies are moving to
cloud computing. This paper examines the many strategic business benefits behind the adoption of
cloud computing in small, medium-sized, and enterprise companies. As with many new technologies
proposed for improving business processes, customers have legitimate questions about adapting
their critical applications and data to these technologies. Similarly, myths and misperceptions of new
technologies may easily spread with little basis or justification. This paper addresses potential
business concerns and prevalent misperceptions of cloud computing and offers interested
businesses concrete advice for making informed decisions. Finally, this paper broadly describes the
cloud computing architecture.

Introduction

Cloud computing can dramatically lower the cost and complexity of computing for large and small
businesses. However, before choosing a cloud computing solution, businesses should have a more
thorough understanding of cloud computing capabilities—and limitations.

Perhaps the best way to begin to appreciate the potential for cloud computing is through a
concise definition of the term:

“Cloud computing is a model for enabling convenient, on-demand network access to a

shared pool of configurable computing resources (e.g. networks, servers, storage,
applications, and services) that can be rapidly provisioned and released with minimal
management or to service providers interaction.”–The National Institute of Standards and
Technology, U.S. Department of Commerce; October, 2009.

The delivery of computing resources on-demand is key to the core concept and value of cloud
computing. Until recently, customers who relied on computing resources needed to buy, install,
and maintain their computer equipment and software, regardless of the type of computing they
needed to accomplish. This is still the prevailing model for data processing and networking at
most businesses around the world today. Under this “wholly owned” computing paradigm,
whether an application is used every minute of the work day by thousands of employees or
merely once a year by one user, the company must purchase, maintain, and provide the
infrastructure for accessing that computing asset. The idea of allowing on-demand access to
configurable resources has tremendous cost-benefit advantages, as this paper examines in
more detail later.

Cloud computing could not have become a viable business market offering, however, until the
second half of the cloud computing definition became a reality:

“…a shared pool of resources that can be rapidly provisioned and released with minimal
management effort or service provider interaction.”

Rapid provisioning of shared resources evolved out of the adoption of Internet computing. As
ecommerce and Internet businesses ramped up in the late 1990s, success depended on
serving a large number of online customers and processing thousands of transactions and
hundreds of applications simultaneously. To handle this type of workload online, Internet
companies developed more dynamic computing infrastructures and software applications.
These online tools rapidly evolved into sophisticated systems and ultimately served as the basis
for cloud computing. Where these online architectures and applications first served only
ecommerce business developers, next-generation infrastructures and applications can now be
dynamically allocated to any member of the Internet network community. In essence, what was
learned, tried, and tested through the rigors of successful online commerce, can now be offered
as a computer service end-product to the wider business community.

The second key component in the evolution of cloud computing was virtualization of computing
resources. (Virtualization is the creation of a virtual version of a resource—such as a server, a
disk, or CPU—inside a computing framework.) A large physical server, for example, could “run”
multiple copies of virtual servers, each with their own virtual memory, CPU and disk. Therefore,
this pool of virtual resources can be more rapidly provisioned and deployed over the Internet
than any physical resource.

Today, dozens of companies offer cloud computing products of varying sophistication and
capabilities. As a testament to the features and business benefits of cloud computing, this diverse
market segment has grown tremendously over the last three years. According to industry

analysts at Gartner Group, Inc., revenue for cloud computing totaled $56.3 billion in 2009, a
21.3 percent increase from the industry’s 2008 revenue of $46.4 billion. Gartner further predicts
that market revenue for cloud computing will exceed $150 billion by 2013.

While much of the growth in cloud applications and infrastructure is driven by cost-conscious
enterprise businesses, the cloud computing industry is increasingly developing products
specifically for mid-sized and small businesses. The goal is to exploit the unique advantages of
cloud computing to give their businesses a competitive edge. Even the smallest business can
realize immediate reductions in capital expenditures for application software, for example. In
fact, some industry analysts, such as Cambridge, Massachusetts-based Forrester Research,
explicitly advise large and small customers to adopt cloud computing as a cost-savings
strategy during the current economic downturn.
Although trimming costs is one advantage to cloud computing, it is not the only motivation for
adopting an on-demand, configurable pool of computing resources. Cloud computing can save
maintenance and downtime, reduce human capital required to manage a data center, and can
better accommodate business growth.

The Business Case for Cloud Computing

The primary motivation for businesses to move to cloud computing is to save money.
Specifically, companies will spend less on computer hardware and software, reduce the number
of personnel required to maintain their data, and significantly save on the space and energy
needed to run their computer equipment. However, cloud computing also arms companies with
cutting-edge tools to become more competitive, including flexible and agile computing platforms,
more scalable and high-performance resources, and highly reliable and resilient applications
and data.

Flexibility, Scalability, and Resilience Using Cloud Computing

While cost savings may be the prime motivation for moving corporate resources to a cloud
computing platform, companies are discovering that cloud computing also offers business
advantages that are difficult to realize in any other fashion. Because of the ease with which new
resources can be added when using cloud technologies, companies can respond quickly to
unexpected increases in demand rather than wait for infrastructure upgrades—a wait that could
cause lost business and create lasting damage to the perceived value of a service. Similarly,
businesses can throttle back their consumption of computing resources during off-peak hours or
during slower business seasons, rather than paying for capacity they will not use. This rapid,
elastic response to business requirements through cloud computing gives businesses a
competitive advantage in the marketplace over companies clinging to a fixed computing asset
model.

Similarly, companies using a cloud infrastructure can develop and deploy new applications in a
fraction of the time required in more traditional computing environments. Testing and development
resources can be quickly allocated and isolated from production processes, and can just as easily be
removed when they are no longer required. Development can be conducted safely and under nearly
identical environments as final production deployments, insuring more reliable and predictable
application performance. Therefore, cloud computing delivers scalable and high-performance
capabilities on-demand without significantly increasing operating expenses or capital outlays.

Finally, companies that deploy applications on high-performance cloud computing platforms are
guaranteed a level of business continuity and resilience nearly impossible to duplicate on owned
computing systems without a significant investment in disaster recovery sites, replication software,
and highly trained disaster recovery and storage management

personnel. Because sophisticated cloud computing platforms keep redundant copies of client
data at multiple cloud computing data centers, customers are protected from local physical
disasters such as fire and flood, accidental data loss or corruption, and malicious data
destruction. If data or applications become unavailable at one site, systems can switch over to
the backup data center.

In a recent study, one accounting firm found typical

Server utilization rates as low as12percent. In stark

Economic terms, this means that roughly $4400 of

Every $5000 server investment is idle.

Hardware, Software, and Maintenance Savings

Computer hardware such as CPU, disk, memory, and network I/O is notoriously under-utilized
in business data centers. Typical server utilization, for example, may be from 15 to 30 percent
of capacity on average, with only intermittent spikes of activity pushing the server environment
to approach its full capacity. In a recent study, one accounting firm found typical server
utilization rates as low as 12 percent. In stark economic terms, this means that roughly $4400
of every $5000 server investment is idle. As a result of all these factors, business data centers
are wasting a huge percentage of their hardware money.

In an ideal cloud computing environment, companies would pay only for the hardware
resources they use, when they use them. In addition, the cloud platform should easily adapt to
the customer’s business profile, seamlessly adding more resources during peak computing
periods or during unexpected spikes in activity.

While tracking the initial expense of buying and installing network applications such as email, SQL,
or other productivity software at a typical business is easy, the true cost of ongoing maintenance and
support of that software is often hidden and poorly understood by many companies. Upgrades, data
backup, configuration, and technical

support multiply the true cost of software well beyond the original purchase price. Support and
productivity losses such as downtime during upgrades or recovery of lost data are rarely figured
into the bottom line cost of software, yet are significant cost multipliers. Delivery and configuration
of software can be significantly streamlined under cloud computing, and the reliability of cloud
computing data centers ensure that downtime and data loss are minimized.

Hidden Personnel Costs

According to a recent study published by Government Computing News, companies could

potentially save 70 percent of their operational costs using a cloud computing provider versus
managing the same resources in house. A significant portion of these operating expenses are
related to IT personnel. For example, a typical network administrator’s yearly salary is $69,000,
while network security specialists cost on average $87,000 per year. Most mid-sized networks
require dozens of IT professionals and technicians to maintain operations. But salary and
benefits are just one component of IT personnel costs. Ongoing training of personnel is also
essential for maintaining a high level of expertise among IT employees that require time off for
retraining, potential travel to IT seminars and workshops, and tuition and fees associated with
continuing education. Other personnel expenses include auditing for compliance with
government regulations, such as HIPPA and Sarbanes-Oxley.

Cloud computing significantly reduces the number of IT personnel at a business by shifting

maintenance, IT expertise, and many compliance requirements to the cloud infrastructure
provider. In fact, depending on the competency of the cloud computing provider, most
companies will receive support and expertise that far exceeds what they could possibly afford to
recruit on their own.
Energy and Space Savings

With fewer servers, storage sub-systems, and network routers on the premises, companies that
move a significant portion of their computing infrastructure to a cloud provider save money on both
the electricity to power those systems and the air conditioning systems required to keep them cool.
Floor and rack space for equipment is also reduced. Because the most sophisticated cloud
computing data centers are more energy efficient, this energy savings is not just shifted to the cloud
provider but constitutes actual savings and a net reduction in total energy consumption, resulting in
a positive impact on the environment.

Cloud Computing Concerns

While the business benefits of cloud computing are clear, some businesses are still reluctant
to move critical data and applications to cloud computing, and especially to a third-party
cloud infrastructure. Some common concerns include the following.

Lack of direct control. With data hosted by a cloud provider and not on the company’s
premises, management may feel that data is no longer under the direct control of the business
and somehow more vulnerable. However, once business managers and IT directors are familiar
with the management tools and data structures of the cloud-based applications, many of these
control issues can be dismissed. The cloud provider should give the customer tools for data
management, and the provider should have ample disaster recovery and fault tolerant
measures in place to protect the data 24X7.

Uncertain security. Once customers become familiar with the security procedures and measures
that most cloud computing providers have deployed, they find there is little concern about security.
In fact, depending on the cloud provider, customer data in most cases is far safer at a cloud
computing facility than on typical network LANs, where
decentralized security management can introduce any number of ongoing security
vulnerabilities.

Utility pricing. Unpredictable pricing is a frequent concern of many companies considering

cloud computing. Because many cloud providers operate on a utility pricing model, customers
are charged by usage—similar to cell phone or electricity use. Therefore, many companies are
concerned about the potential for wildly fluctuating computing costs. However, not all cloud
computing providers use utility pricing, but instead offer fixed price plans, keeping computing
costs predictable and affordable.

Data lock-in. Many companies fear that once they commit to cloud computing, their data and
businesses are locked in, making them ”hostage” to a particular provider of cloud services. For
this reason, companies should ensure that any hosted applications or services they use are
based on open standards and open data formats in the event that they need to quickly and
cleanly migrate their resources to another vendor or system.

Supplier viability and reliability. Customers should be concerned about the viability of any IT
supplier, especially one tasked with hosting critical applications and data. For this reason, potential
cloud customers should check not only the viability of the host’s infrastructure but also check its
financial health and market standing. Aside from the cloud provider’s long-term market viability, what
measures do they take to ensure their data centers are reliable and available 24X7?

Cost of converting to cloud computing. Many customers are understandably concerned about the
cost of converting from their current applications to cloud-based applications. In many cases, such
as SQL databases, spreadsheets, and word processing documents, the cost will be nominal as
cloud-based applications can preserve the formats most businesses are currently using. Custom
applications may

require more time and expense to convert, however. For these reasons, many companies
moving to cloud computing often do so incrementally. They may first offload their email, word
processing, or data storage tasks to offsite cloud computing resources on a trial basis before
committing larger aspects of their business infrastructure to a new platform. A recent study
found that 70 percent of companies that utilized cloud computing on an incremental or test
basis planned to further deploy cloud computing more broadly. However, other companies find
that just one or two applications delivered on cloud computing can make a big difference in
their bottom line and productivity, and may never see a need to move any other data or
applications to a cloud provider.

Myths and Misperceptions about Cloud Computing

Cloud computing, like many new products, is subject to misperceptions and myths. These myths
probably arise primarily from a poor understanding of the technology or the capabilities of the
providers.

Myth: Cloud security and compliance is impossible to achieve.

Cloud computing security is no different than any heavily secured ecommerce or wide area
network. To be sure, not all cloud computing vendors utilize the same technology measures,
and for that reason it is up to the customer to thoroughly scrutinize the cloud provider’s security;
however, cloud computing in itself does not introduce any new or unforeseen vulnerabilities or
weaknesses. Also, cloud computing companies can provide a unified platform for conducting
and verifying compliance audits.

Myth: All clouds scale on demand. Not all cloud vendors have the resources or architecture to
adequately scale applications and traffic on demand. While all try to maintain a certain number of
extra resources to accommodate fluctuations, many cannot dynamically scale operations when
demands exceed predicted thresholds. For example, in most

cloud computing environments, scaling to add more users (horizontal scaling), requires manual
reconfiguration and allocation of additional CPU resources. As a result, companies should carefully
select their cloud computing provider to ensure it can dynamically accommodate increased user
demand; otherwise applications and users may be subject to frequent and disruptive slow-downs
and system bottlenecks.

…in most cloud computing environments, scaling to add more users

(horizontal scaling)—takes manual reconfiguration and allocation of
 additional CPU resources. As a result, companies should carefully select
their cloud computing provider to make sure it can dynamically
accommodate increased user demand; otherwise applications and users
may be subject to frequent and disruptive slow-downs and system
bottlenecks.

Myth: Virtualization = cloud computing. Virtualization makes dynamic, scalable cloud

computing possible, but does not constitute a cloud architecture on its own. Virtual machines
deployed without intelligence or dynamic scalability can be nearly as inefficient and costly as
physical resources they replace. In fact, many enterprises have identified this inefficient use of
virtualization and dubbed it “virtual sprawl.” This unchecked proliferation of virtualization, in turn,
can easily translate into a decrease in network performance according to a 2008 report by the
Aberdeen Group.

Myth: Performance is worse in the cloud. If the cloud infrastructure and applications are poorly
managed and deployed, this might be true. But when properly configured, most users notice no
difference when using cloud-based applications. In some cases, cloud computing provides
noticeable improvements in performance as better provisioned machines with access to more
resources can better handle more complex processes than poor-performing, local desktop
computers. The most significant potential bottleneck for cloud computing is access

to the network itself. If users are connecting over high-speed connections, performance
should never be an issue.

Myth: The cloud requires more IT management. In the beginning of a deployment, as users
and applications are carefully migrated to cloud infrastructure, this may be true. But once
conversion is completed, local IT no longer must manage the software, the hardware, or the
data protection of the applications and associated data. In fact, every application migrated to a
cloud infrastructure cuts maintenance costs and management overhead.

Myth: Cloud computing is only good for low end applications and software as a service.
Many vendors have jumped into the cloud computing marketplace with simple software
applications and declared themselves “cloud computing” experts. While their applications may
certainly fill a need, delivering a simple application over the Internet does not approach the
sophistication or complexity of delivering massive computing infrastructures on demand. Cloud
computing is the backbone on which businesses worldwide perform thousands of transactions a
second, transfer massive amounts of data across the globe, and securely process trillions of
dollars of financial transactions. The most robust, secure, and scalable business applications
available today operate using cloud computing.

Myth: Cloud computing is less reliable than in-house systems.

Some of the most secure and reliable installations in the industry are cloud computing data
centers. The best cloud computing centers are built from the ground up with multiple layers of
redundant components, power, physical and cyber security measures, and staffed by highly
trained personnel. Using an economy of scale impossible to duplicate except at the largest
enterprise IT centers, the best cloud computing centers deploy reliability, security, and failsafe
measures that most businesses simply cannot afford.

View of Cloud Computing

Business computing is constantly evolving as business needs change and technology advances.
The first great wave of business computing was centrally managed. This mainframe-centric world
meant that all data was kept, processed, and managed from one central location. Users could
access and enter data, but only through strict central control. This central control caused problems
with productivity as users struggled for adequate access to produce and share timely work.

With the advent of PC and LAN computing, data and computing resources became
decentralized, spread across numerous servers, and on individual desktops spread throughout
the business. Now data was easier to access, productivity increased, yet the chaotic dispersal
of data caused problems with security, data reliability, and wasted computing resources. Finding
and retrieving data in a decentralized system is often difficult as well, causing numerous
inefficiencies. Similarly, computing resources increased dramatically, but they were not easily
accessible throughout the network.

With a distributed computing architecture based on cloud architecture, users are looking for the
security, protections, and reliabilities of the centralized mainframe computing systems combined with
the access and connectivity of a decentralized LAN environment. Smart computing software enables
a complex network of data centers to work together as a unified computing platform on which
companies can build their next-generation applications.
Summary

Cloud computing has rapidly evolved into a major tool for businesses. Companies can save
capital expenses, software maintenance, personnel, and energy costs by shifting significant
portions of their business processes to cloud computing platforms. The flexibility of cloud
computing also allows them to more quickly develop and deploy innovative software
applications.

Despite fears of lack of data control or application security, cloud computing has proven itself to
be a safe, reliable platform for enterprise computing throughout its evolution. In many cases,
next-generation cloud-based applications are more secure and perform better than traditional
data center-based solutions.

Sophisticated cloud computing technologies can provide a full range of computing resources for
customers.