I Can, but I Won’t- An Exploratory Study on People and New Information Technologies in the Military.

I CAN, BUT I WON’T: AN EXPLORATORY STUDY OF PEOPLE AND NEW

INFORMATION TECHNOLOGIES

I. Introduction

Overview

According to the National Institute of Standards and Technology (NIST), cloud

computing “is a computing model that enables convenient, on-demand network access to

a shared pool of configurable computing resources (hardware and software) that can be

rapidly provisioned and released with minimal management effort or service provider

interaction.” In September 2009, the Federal Chief Information Officer, Vivek Kundra,

announced the Federal Government’s Cloud Computing Initiative. This initiative laid the

groundwork for the future rapid deployment of technology solutions supporting the

Federal Government without developing stove-piped systems (GSA, 2009). Currently,

the Federal Government, considered the world’s largest purchaser of information technology,

spends over $76 billion per year on more than 10,000 systems (Kundra, 2009). Correctly

implemented, the cloud computing model has the potential to dramatically decrease

information system cost via virtualization, reducing related infrastructure, building,

power, and staffing expenditures (GSA, 2009) making it attractive to any cost-crunched

organization.

Cloud computing centers provide a foundation to run enterprise services securely

and reliably across the DoD. By leveraging cloud computing techniques, the Department

of Defense (DoD) can transform its infrastructure from its legacy system-specific

infrastructures to a shared infrastructure (CIO/DoD, 2010). Some recent examples of

recent cloud computing successes in commercial industry include Kenworth and Coca-

Cola Enterprises. Engineers at Kenworth, lacking organic high-powered computing

capacities, rented time on a supercomputer thousands of miles away to help find gas-

guzzling design flaws. The Kenworth engineers took advantage of cloud computing and

discovered the mud flaps were a major source of drag. Redesigning the mud flaps will

save about $400 on a typical truck’s annual fuel bill (Fortt, 2009). Coca-Cola Enterprises

uses a cloud-based logistics system to optimize operations with merchandisers in the

field. Using smart phones and a cloud computing setup, personnel responsible for

restocking Coke products in stores stay in constant contact with their bosses and the

company’s information storehouse. This allows the field personnel to provide automatic

inventory updates and last minute changes to merchandising schedules.

Cloud computing, while trendy as a new computing model, it is not new. Since

2009, federal agencies have been able to buy cloud computing applications and services

at Apps.gov. GSA’s Apps.gov storefront offers an array of business applications,

productivity software, and services ranging from social networking, to website hosting,

and data storage (Helft, 2009). The government has approved these cloud-based

applications and services to replace more costly and demanding computing services that

are owned and operated by federal agencies. Government organizations seeking to tap

cloud-computing benefits are not constrained to simply the GSA storefront. In July 2010,

Google announced the launch of Google Apps for Government; a suite of Federal

Information Security Management Act (FISMA) approved cloud services for "moderate"-

level security requirements (Google, 2010). With FISMA approval, government

organizations can hire Google to provide secure sensitive, but not classified information

systems, processes, and data. This blend of government and commercial providers

affords options for dramatically changing the way organizations answer their IT needs.

Early Successes of Cloud Computing

Williams F1, a Formula 1 race team based in Oxfordshire, U.K., turned to AT&T

Global Services for a cloud computing solution. The Williams F1 race team is on the

road for 9 months out of the year competing in 19 Formula 1 Grand Prix races that make

up the race calendar. Its headquarters contains a fully managed enhanced virtual private

network that allows engineers, mechanics, car designers and drivers to run the business

while they are on the road. Alex Burns, CEO of Williams F1, states that their IT strategy

used to be about speed and the speed of decision-making on testing and accessing the

results with little regard for security. While Williams F1 is a tech savvy race team in a

competitive and high profile environment replete with espionage, it knew little about

security. To correct this, the team turned to a cloud provider for a highly secure

infrastructure service, allowing their own IT department to focus on helping the

organization build winning racecars. AT&T’s cloud security services enables the

Williams F1 staff to receive encrypted data from the car to make decisions during the

race. Alex Burns said that by using cloud computing, the team saves time and money

since security services require constant management, patches, new hardware, and the IT

department's time (Del Nibletto, 2010). Burns states that Williams F1's core competency

is building racecars, not IT security, nor should it be.

In the military arena, the armed services were assigned the task of finding more

than $100 billion in overhead savings over the next five years (Gates, 2010). Any

savings the services generate can be reinvested in war-fighting modernization. At least

one service, the U.S. Army, has recently stepped out and looked toward cloud computing

as a means to cut costs and increase efficiency. In a move to reduce its $10 billion IT

budget, the Army issued a moratorium on server purchases (Foley, 2010) to reduce server

counts and consolidate the 200+ Army data-centers. Spearheading one of these cost-

cutting efforts is the United States Army Recruiting Command. Their Army Recruiting

Information Support System was over 10 years old and in need of an upgrade (Kundra,

2010). Therefore, they embarked on a pilot program designed to explore new

technologies that the Army could leverage to improve efficiencies of its recruiting

operations. Initial bids from traditional IT vendors that met the required functions ranged

from $500,000 to over $1 million. Instead of accepting these solutions, the Army

Recruiting Command chose a customized version of a cloud-based Customer Resource

Management (CRM) tool from Salesforce. The CRM tool from Salesforce provided all

the functions without needing to acquire all the necessary hardware needed to operate a

traditional system. The Army Recruiting Command is currently piloting this cloud-based

solution at an annual cost of only $54,000.

In a broader example of the cloud computing movement, in 2011, the Army will

start migrating its Microsoft Exchange email users to an enterprise-wide email that will

leverage Army-owned Microsoft licenses and the DoD cloud managed by the Defense

Information Systems Agency (DISA). The base email service enables the Army to

modernize its Microsoft Server software and increase email capabilities while

simultaneously substantially reducing hardware and storage expenses. Estimated savings

may reach 40% (DISA, 2010). According to the DISA press release, the DISA-managed

Enterprise Email will employ fewer servers and administrators, increase security and

eliminate thousands of existing heterogeneous local networks. This change will expand

the email capability with email storage growing to 4 gigabytes for most users. In

contrast, the Air Force’s Air Mobility Command (AMC) revised its newly restrictive e-

mail limit policy in January 2009. The new limits for e-mail size are divided into three

Tiers. Tier I, which includes Group CC\CV and CMSgt has unlimited size of e-mail

boxes. Tier II includes CC, CV, CEM, First Sergeant, and organization accounts are

limited to 250MB. Finally, normal users, Tier III are limited to only 50MB for their e-

mail box storage. This is one of the many examples where the use of cloud computing

resources can reduce costs and increase efficiencies while maintaining performance and

security. These examples of the early success of cloud computing reinforce the need for

this research.

Motivations for Research

The motivations for this research stems from personal experiences as a Signal

Officer for an Infantry Battalion and a Company Commander for an Operational-base

Signal Company. While serving as a deployed Signal Officer at a remote Forward

Operating base in Operation IRAQI FREEDOM, there were many areas in which

communications efficiency and capabilities could be improved. During the pre-

deployment and post-deployment process, the Signal section had to transfer data to hard

drives and file servers in order to re-image computers so they could connect to the

respective local networks. While the imaging process was not difficult, the process of

backing up the users’ data multiple times over eighteen months due to training exercises

and deployments was time consuming. For a training exercise at the National Training

Center, users’ data had to be backed up so computers could be re-imaged for the network

used for training. This process was done after returning to the home station, before

deployment to Iraq, and upon returning to home station. By comparison, using a cloud-

based data service during the training exercises and deployments would eliminate the

need to repeatedly move 150 users’ data back and forth from individual computers to a

server. Instead, the data would reside in the cloud and always be available.

At another time, as a Company Commander serving in Alaska, I encountered

another problem with a potential cloud-computing solution. The Stryker Brigade soldiers

possessed multiple email addresses that were used during deployments. There was an

Army email address (xxx.us.army.mil), an email address at work (xxx.post.army.mil), an

email address for the Brigade that was used for training and deployments

(xxx.unit.army.mil), and SIPRnet email addresses for the Army, Post, and unit. Multiple

e-mail addresses used during garrison and deployed operations meant multiple accounts

to manage at one time; users had to check several accounts for messages, and IT

personnel had to provide support for each set of exchange servers. Using a cloud-based

email service, like the DISA Enterprise Email, would overcome this obstacle since doing

so would eliminate the need for separate post and unit email accounts. Since

implementing cloud-computing solutions does not rely on multiple exchange servers, it

eliminates the need to manage multiple email accounts. These examples lead to the

following question: If cloud computing can provide cost savings while still meeting the

computing needs of the organization, then why is it not readily implemented through the

Department of Defense?

Research Goals

In the commercial industry, Kenworth, Coca-Cola Enterprises, and Williams F1

are just a three of the success stories of how cloud computing has helped organizations

save time and money. Whether saving money by redesigning the aerodynamics of a

truck, or optimizing the workload of merchandisers working in the field, or collaborating

on data gathered from a car during a race, cloud computing is providing advantages in

commercial industries.

The same advantages are only beginning to be seen in the government sector.

The Army Recruiting Command is saving money on a pilot cloud computing solution for

the aging Recruiting Information System. Similarly, on a larger scale, the Army is

transitioning to a cloud-based Enterprise E-mail system that is supposed to increase

capabilities while reducing costs. If there are numerous successes using cloud computing

technologies in commercial industries, then why has it not been implemented as quickly

in the government sector? The answer to this question is the thrust of this research.

There are two main research goals in the thesis. The first goal is to determine if

there is a perceived difference of opinion between commercial and government sectors on

issues concerning cloud computing. By comparing the results from a 2009 IDC

Enterprise survey with the results from a military IT Personnel survey, the thesis will

determine if there is a significant difference in the perception of cloud computing issues.

Commercial and government sectors might have a different perception on cloud

computing issues due to the nature of their organizations. Government and commercial

industries have different customers, stakeholders, and goals driven by the nature of the

two industries. For one example, expenses and income do not drive the government

sector, whereas the commercial sector pays attention to profit of the organization. The

differences in public and private industry can mean a difference in technology needs and

opinion about a new technology. In fact, studies on information systems (IS) projects in

the commercial sector may have limited relevance due to differences between the

government and commercial industries (Bozeman and Bretschneider 1986; Bretschneider

1990) and any recommendations derived from commercial sector studies might not

necessarily apply to the government sector (Coase 1937).

The second goal of the thesis is to understand the relationship between the

perceived viability of cloud computing and perceived willingness to implement it in the

organization. Is it possible that cloud computing is not seen as a viable technology in the

military environment? Conversely, is it viable and IT professionals in the military are

just unwilling to make the changes needed to implement it in their organizations. The

previous discussions on the success of cloud computing leads us to the following

questions.

Research Questions

With the benefits of cloud computing seemingly apparent, the reasons why it has

not yet been adopted may lie in the human element of the people, process, and

technology equation. This thesis is an exploratory examination of the opinions military

IT personnel of cloud computing in the in the Department of Defense and the Army. The

research presented is motivated by a desire to understand two overall research questions:

(1) Is cloud computing perceived as a viable technology in the DoD/Army?

(2) Is there a perceived willingness to implement cloud computing in the

DoD/Army?

The research will provide an understanding of whether DoD IT personnel think

cloud computing is a viable technology for their organization as well as examine the

connection between the perceived viability of new technologies and the willingness of IT

personnel to implement them. Answering these questions will provide a framework for

determining whether new technologies are viable for the organization and the perceived

willingness to implement the technology.

Thesis Overview

The remainder of this thesis includes four more chapters and supporting

information found in the appendices. This next chapter introduces cloud computing,

defining and explaining the varying characteristics, service models, and deployment

models. Chapter three then discusses the research strategy and survey methodology used

to collect the data required to address the research questions. Then, chapter four presents

the analysis of the survey data and tests the research model to answer the primary

research questions. Chapter five discusses the results, providing recommendations along

with possible limitations and future research ideas.

II. Background

History of Cloud Computing

Recently, cloud computing has become one of the most talked about topics in

information technology (Gartner, 2010). The Gartner Group (2010) identified cloud

computing as the primary source of growth in IT spending with global revenues surging

to over $130 billion in 2013. Cloud computing is on-demand access to virtualized IT

resources that are located outside of your own datacenter (Marks and Lozano, 2010).

These virtualized IT resources are accessed over the Web, shared by others, easy to use,

and paid-for through subscriptions. Significant innovations in virtualization and

distributed computing, as well as improved access to high-speed Internet and a weak

economy, have accelerated interest in cloud computing (Lee, 2010).

The idea behind cloud computing dates back to the 1960s when John McCarthy, a

computer scientist, stated in a speech given at MIT, “That computing may someday be

organized as a public utility” (Biswiss, 2011). Cloud computing gets its name from a

metaphor for the Internet (Velte et al., 2010) and was probably copied from internet

diagrams (Biswiss, 2011) where data is depicted as traveling from one computing device

to another through a nonspecific cloud (Figure 1).

Figure 1. The Internet Cloud

(http://singularityhub.com/)

In such a model, cloud computing providers allow users to remotely access

hardware, software, and data resources for a fee. Users essentially store data and

applications in the “cloud” for easy access (Figure 2). There are several commercial

applications utilizing cloud computing. For example, OnLive, an online gaming service,

provides access to games instantly over a broadband connection using a browser.

Similarly, Google Docs provides a web-based document editing and management service,

while Bing Maps provides an on-line mapping service. The breadth of commercially

available services available via this model continues to grow (Geelan, 2009).

Figure 2. Cloud Computing (Velte, 2010)

What is Cloud Computing?

Commercial industry leaders have many variations of the cloud computing

definition and its characteristics. Accenture, a consulting firm, defines cloud computing

as the dynamic provisioning of IT capabilities (hardware, software or services) from third

parties over a network. In another view, Kevin Fogarty, a contributing editor at CIO

magazine (2009), states that the cloud-computing model has all applications, services and

networks available to IT and end users via the Internet. Meanwhile, Jeff Kaplan from

ThinkITStrategies views cloud computing “as a broad array of web-based services aimed

at allowing users to obtain a wide range of functional capabilities on a ‘pay-as-you-go’

basis (Geeelan, 2009).” While different, at the core these definitions all have

organizations or users accessing and purchasing IT capabilities from a third party.

Nonetheless, they are not the only ones that show the variety of perceptions about what

cloud computing is.

According to TechTarget, a publicly traded IT marketing company, cloud

computing is a general term for anything that involves delivering hosted services over the

Internet. Under such a vision, cloud computing has all three distinct characteristics that

distinguish it from traditional IT hosting: services are sold on demand (usually by the

minute or the hour), services are elastic (so a user can have as much or as little of a

service as they want at any given time), and the services are fully managed by an external

provider allowing consumers access with nothing more than a personal computer and

internet. One consulting firm Lexnet Consulting Group, defines cloud computing as a

delivery of services that replace the need for an organization to incur infrastructure costs;

it can be thought of as an “outsourced” IT network (Chipman, 2010). This cloud

represents a shift away from computing as a product to computing as a service delivered

over the Internet (Khajeh-Hosseini, Sommerville and Sriram, 2010). Knorr and Gruman

(Infoworld, n.d.) state cloud computing comes into focus when one considers what IT

departments and organizations always lack: a way to increase capacity or capabilities

when needed without investing in new infrastructure, training, or software licenses.

Thus, the term cloud computing can be used to cover any subscription-based or pay per

use service that extends existing IT capabilities (Knorr and Gruman, n.d.).

There is one commonality for almost all papers and definitions of cloud

computing: they all refer to or reference the National Institute of Standards and

Technology (NIST) cloud computing model. For example, see: Bambacus, 2008,

Biswas, 2011, Chipman, 2009, DoD/CIO, 2010, Foarty, 2009, Geelan, 2009, Khajeh-

Hosseini et al., 2010, Kundra, 2008 and 2010, Knorr and Gruman, 2010, Lee, 2010,

Lewin, 2009, Marks, 2010, Rhoton, 2009, Rittinghouse and Ransome, 2010, and Wyld,

2009 and 2010. Since the NIST model is fundamental to so many commercial cloud

computing visions, this thesis will use the government’s definition of cloud computing

which was published in January 2011 under the Federal Information Security

Management Act (FISMA of 2002 in Special Publication 800-145). The NIST defines

cloud computing as a “model for enabling ubiquitous, 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 and effort or service provider interaction.” The cloud model

promotes availability and is composed of five essential characteristics, three service

models and four deployment models that are explained below.

The five essential characteristics are on-demand self-service, broad network

access, resource pooling, rapid elasticity, and a measured service. The three service

models are Software as a Service (SaaS), Platform as a Service (PaaS) and Infrastructure

as a Service (IaaS). The four deployment models outlined by the NIST are private cloud,

public cloud, community cloud, and hybrid cloud. Further details are provided below.

Five Essential Characteristics

Characteristic #1: On Demand Self-Service.

On-demand self-service occurs when a consumer can automatically and

unilaterally provision computing capabilities, such as server time and network storage,

without requiring contact with the providers’ customer service representative. It means

the consumer can use the cloud service as needed without any customer service

interaction with the cloud provider (Ruggles, 2010). One example is from VMware.

Their vCloudTM Express offers an on-demand service, where developers can use the

service at their convenience to address various infrastructure and programming needs

such as experimentation, prototyping and testing. A vCloudTM customer can create a

virtualized server, add storage, configure a firewall, and scale additional capabilities in

and out according to what and when they need it.

Characteristic #2: Broad Network Access.

Broad network access is an essential characteristic defined by the NIST as

capabilities that are available over the network and accessed through standard

mechanisms that promote use by heterogeneous thin or thick client platforms (e.g.,

mobile phones, laptops, PDAs). This is the most vital characteristic of cloud computing

since it is network based, and accessible from anywhere, from any standardized platform

(Velte, 2010). This does not necessarily mean Internet access. A private cloud is

accessible only behind a firewall, regardless of the type of network (Ruggles, 2010). The

main point of this characteristic is that the service is accessible from anywhere.

Characteristic #3: Resource Pooling.

Resource pooling is the next characteristic. In resource pooling, numerous clients

use the same set of resources at the same time (Rhoton, 2009). It works on economies of

scale: users do not have their own resources, so a provider gives multiple parties access

to a large pool of shared resources at efficiencies no one user could match. The cloud

computing service providers make their resources available to anyone who is willing to

pay for access.

Characteristic #4: Rapid Elasticity.

The characteristic of rapid elasticity is defined as the ability to scale computing

resources both up and down, as needed. To the consumer, the cloud appears to be

infinite, and the consumer can purchase as much or as little computing power as needed

(Ruggles, 2010). This characteristic not only allows the services to be scaled both up and

down (scalability), but per-usage billing is common with this characteristic and this leads

to direct cost savings.

Characteristic #5: Measured Service.

The last characteristic is called a measured service. NIST defines this as the

leveraging of a metering capability; usage can be monitored, controlled and reported,

providing clear usage details for the provider and consumer, like a utility service model.

Measured services allow the cloud provider to charge for exactly what the customer is

using. The customer can track usage and costs and align them with their specific

business units or functions for cost accountability.

Three Service Models

Besides containing the five essential characteristics, cloud computing

implementations are available in three service models: Software as a Service (SaaS),

Platform as a Service (PaaS), and Infrastructure as a Service (IaaS). These models

describe the level of functionality offered by the cloud provider. Figure 3 depicts the

relationships between the models as a pyramid. At the highest or software as a service

level, the consumer receives more functionality and customization and knows less about

the implementation details. At the lower levels, the consumer receives more components

in the service and a higher degree of control over the infrastructure, platform, and

software.

Figure 3. Cloud Service Models

Service Model #1: Software as a Service.

Software as a Service (SaaS) provides the customer the ability to use applications

running on a cloud infrastructure (NIST, 2011). SaaS offers fully functional applications

on-demand to provide specific services such as e-mail, customer relationship

management, web conferencing, and other applications. The applications are accessible

from various devices though a thin client interface (e.g., web browser) allowing

consumers to avoid upfront investment in servers or software. Salesforce.com is the most

well known example among enterprise applications (Knorr and Gruman, 2010). Onlive,

Google Docs and Bing Maps are other well-known examples the growing number of

SaaS providers. For example, Google Docs allows you to upload files, then edit, and

view the documents from any computer or smartphone. This fosters real-time

collaboration with other authorized users, while shielding the company from

implementation details such as software installation, updates, and patches.

Service Model #2: Platform as a Service.

NIST defines Platform as a Service (PaaS) as another service model. PaaS allows

the consumer to deploy their own applications that are created using approved software

tools, onto the cloud. PaaS falls between SaaS and Infrastructure as a Service since it

balances functionality with control. Consumers use PaaS to generate custom applications

using software development languages and tools offered by the vendor. PaaS offers an

operating environment that includes the operating system and application services. These

services are constrained by the vendor’s design and capabilities. PaaS solutions are

development platforms in which the development tool itself is hosted by the provider and

accessed through a web browser. Due to this, developers can build web applications

without installing any tools on their computer and can deploy those applications without

any special system administration tools. Examples include Force.com by Salesforce.com,

Google App Engine, and Microsoft’s Azure.

Service Model #3: Infrastructure as a Service.

Infrastructure as a Service (IaaS) is where the consumers are provisioned with

processing, storage, networks, and other fundamental computing resources on which to

deploy and run any software, including operating systems and applications (NIST, 2010).

The IaaS vendor provides a virtual machine to the consumer allowing them to manage

applications and data while the vendor manages which physical machine executes the

code. Virtualization enables IaaS providers to offer almost limitless instances of servers

to customers and make cost-effective use of the hosting hardware. IaaS users enjoy

access to enterprise-grade IT infrastructure and resources that would be prohibitively

costly if purchased on their own. Commercial IaaS providers include Amazon’s Elastic

Compute Cloud (EC2), Rackspace, and GoGrid.

Four Deployment Models

Each of the preceding characteristics and platforms are accessed through different

types of clouds, also known as deployment models. The NIST lists four different

deployment models in SP 800-145: private, public, community, and hybrid (Figure 4).

As explained below, their differences are primarily in the way the consumer or

organization access services.

Deployment Model #1: Private Cloud.

Under NIST, one deployment model is the private cloud infrastructure, which is

operated solely for the use of one organization. It may be managed by the organization

or a third party and may exist on premise or off premise. The private cloud is operated to

maintain a consistent level of security, privacy, and governance control. Organizations

use their own infrastructure and private cloud to operate and protect mission critical

systems. By operating and maintaining a private cloud, the organization has to buy and

manage the infrastructure, and does not benefit from lower upfront capital costs (Foley,

2008).

Deployment Model #2: Public Cloud.

The public cloud infrastructure is made available to the general public or a large

industry group and is owned by an organization selling some type of cloud services. In

such a situation, the cloud service provider makes resources available to the general

public over the Internet. These services may be free or on a pay-per-usage model. The

benefits of a public cloud service are that they are easy and inexpensive to set up since

the provider absorbs the hardware, application, and bandwidth costs. In addition, they

are easily scalable to meet the needs of the consumer and are paid per usage.

Deployment Model #3: Community Cloud

A community cloud infrastructure is one shared by several organizations and

supports a specific community that has shared requirements (e.g., mission, security

requirements, policy, and compliance considerations) (NIST, 2011). This type of model

can be managed by the organizations (members of the community) or a third party and

may exist on premise or off premise. For example, a financial services community cloud

would bring together cloud-based services needed to assist their customers while still

meeting industry-specific security and auditing requirements.

Deployment Model #4: Hybrid Cloud.

The last deployment model is a hybrid cloud. NIST states that the hybrid cloud

infrastructure is a composition of two or more clouds (private, community, or public) that

remain unique entities but are bound together by standardized or proprietary technology

that enables data and application portability. For example, an organization that keeps

their data in a private cloud while using the processing resources of a community cloud is

one type of a hybrid cloud.

Figure 4. Types of Cloud Deployment Models

(Sam Johnston, 2009)

The Silver Lining of Cloud Computing

In an effort to reduce costs during tough economic times, cloud computing offers

a way to significantly reduce waste, increase data center efficiency and utilization rates,

and lower operational costs (Kundra, 2010). Cloud computing benefits are analogous to

public utilities (Kundra, 2010). Public utilities provide access to clean water and

electricity; just turn on the faucet or the light switch and the service is there, and the

consumer pays for only what they use. Cloud computing services can be turned on and

off, as the organization needs them, and provide a pay-as-you-go capability. In such an

approach, only a low initial investment is required to get started (Kundra, 2010). This

lowers the barrier of entry for new or small organizations (Jaeger et al., 2008). Besides

the direct economical benefit, cloud computing can save maintenance and downtime,

reduce human capital required to manage a data center, and better accommodate business

growth (Joyent, 2009). Organizations can get its cloud service operational in record time

at a fraction of the cost of an on-premise solution (Waxer, 2010).

Reducing risk is another benefit of cloud computing. Cloud computing allows

individuals and businesses to reduce some risk by stating in the service contract that data

protection and disaster recovery provisions are maintained and the provider is liable in

the case of failures (Rhoton, 2009). Furthermore, by using a cloud provider, an

organization can reduce the likelihood of not provisioning enough resources for

fluctuating demand and reducing the risk of lost revenue due to unplanned downtime.

Elasticity, which can also be called scalability (Waxer, 2010) allows IT departments to

have an abundance of resources for peak demand can add or subtract capacity as its

network loads dictate (Waxer, 2010). This elasticity allows for the organization to pay

for only what is used (Kundra, 2010) rather than purchasing, installing, and configuring

new equipment (Velte et al, 2010).

Cloud computing also offers firms potential security. System security under

cloud computing promised to be as good as traditional systems since cloud providers can

devote resources to solving security issues that smaller business cannot afford (Marks and

Lozano, 2010). Google (2010) states cloud computing provides improved security

through multiple levels of redundancy across multiple datacenters, thereby ensuring data

integrity while obscuring it from tampering. Cloud providers can offer services that

include disaster recovery, monitoring, forensic readiness, password assurance, and

security testing (Rhoton, 2009).

There are several different types of efficiency gains realized through cloud

computing. Energy efficiency is gained through the use of higher utilization rates, fewer

servers and less energy consumption (Kundra, 2010). Organization and IT department

efficiency is gained by allowing IT personnel to concentrate on mission-critical tasks and

less time on IT operations and maintenance (Kundra, 2010; Golden, 2008). Moreover,

cloud computing allows organizations to convert fixed costs to variable costs, which are

only paid by usage and can be tracked by the departments using the service (Marks and

Lozano, 2010). Cloud computing allows for the increased efficiencies in several areas

that would not be possible using traditional computing methods.

As discussed above, cloud computing offers many potential benefits. From low

initial investment, to reducing maintenance and down time costs, to increasing security

and monitoring, cloud systems allow the organization to be more flexible. Even though

cloud computing implementations offer these potential benefits, no implementation is

risk free.

The Dark Side of the Cloud

There are also some risks involved in implementing cloud computing. Despite

vendor claims of improved security through improved expertise and redundancy, security

remains a sticking point for this new model. Information Security Magazine (2009)

states that information security is the most crucial risk associated with cloud computing.

Making intellectual property, trade secrets, personally identifiable information, or other

sensitive information available on a network requires a large investment in security

controls. In highly sensitive situations, security and other business requirements may

dictate using something other than a public cloud regardless of vendor assurances.

One of cloud computing’s characteristics, broad network access, is a both benefit

and a risk. The benefit, mentioned above, is that employees can access data from

anywhere. The risk involved is that network connections are susceptible to outages and

subject to bandwidth issues. If employees try to access their data during “peak hours”

(e.g., 0900 hours on Monday), then the connection speed could suffer. Should the

organization be driven to purchase more bandwidth to support cloud access, the

expensive upgrade would reduce the cost benefit of cloud computing.

Finally, data privacy is another issue. Users are giving their data to a cloud

provider. Cloud service providers are the holders of very large amounts of sensitive data

and law enforcement officials only need a subpoena to access a user’s data (Zittrain,

2009). When users place their data and applications on centralized servers, they lose

direct control of that information. Sensitive information that was once stored on

organizational computers now resides on the servers of cloud service companies.

Examples include user email, banking information, and backups of individuals’ hard

drive. This creates a risk for the users since storing data in the cloud could increase the

possibility that unwanted third parties will access this data. Some cloud computing

providers store data in clear text, leaving it vulnerable to a security violation. By stating

in the contract with the cloud service provider, data can be encrypted in transit and during

storage at a specified provider location.

The State of Commercial Industry Cloud Computing

Spending on worldwide cloud services is expected to surpass $130 billion in 2013

(Gartner, 2010) and the Cloud Expo in Santa Clara, CA, drew more than 5,000 delegates

and over 100 sponsors and exhibitors in November, 2010. Indeed, many organizations in

the commercial sector are using cloud computing to attain tremendous savings and

reorganize their operations (Kundra, 2010). For example, the NASDAQ is using cloud

computing to give customers a snapshot of information about market conditions at the

time of the trade (Crosman, 2009). Further, a list of some companies that have saved

money by using Google Apps include: Genetech (biotech industry), Virgin America

(airline industry), Salesforce,com (customer relationship management industry), and

Heinz (U.S.-based food industry). Related to this, Morgans Hotel Group deployed

Google Apps for messaging and collaboration needs to its 1,750 employees (Google,

2009). Additionally, JohnsonDiversey, a global provider of commercial cleaning and

hygiene products and solutions chose Google Apps. Google helped JohnsonDiversey

migrate its 12,000 employees to one communications platform, lowering its IT costs and

furthering its commitment to sustainability through the elimination of energy-intensive

email servers. These are just a few examples of commercial industry using a cloud

service provider to obtain cost savings and reorganize their operations.

The State of Government Sector Cloud Computing

Cloud Computing is currently being implemented in government sectors all

around the world. In some instances, government will be the leading sector in the

development of cloud computing (Wyld, 2010). In January 2011, the NIST published a

draft document on the Guidelines on Security and Privacy in Public Cloud Computing

that provides an overview of public cloud computing and the security and privacy

challenges involved with implementing cloud computing in the government sector.

Even though moving to a cloud computing environment can reduce costs,

standards must be in place that maintain the security of government information, protect

the privacy of the citizens, and safeguard national security interests. The Federal

Information Security Management Act (FISMA) of 2002 requires each federal agency to

develop, document, and implement a government-wide program to provide information

security for the information and information systems that support the operations and

assets of the government, including those provided or managed by another agency,

contractor, or other source.

Under FISMA, the NIST was tasked with developing the standards and guidelines

for categorizing all information and information systems, recommending the types of

information and information systems to be included in each category, and developing the

minimum information security requirements for information and information systems in

each category. FISMA defines three security objectives for information and information

systems that serve as the basis for NIST’s analysis: confidentiality, integrity, and

availability. Confidentiality means, “Preserving authorized restrictions on information

access and disclosure, including means for protecting personal privacy and proprietary

information.” Integrity means “guarding against improper information modification or

destruction, and includes ensuring information nonrepudiation and authenticity.” Finally,

availability means “ensuring timely and reliable access to and use of information” (44

U.S.C., Sec. 3542). Overall, the adoption of cloud computing in the government sector is

still in its infancy though. NIST’s Guidelines on Security and Privacy in Public Cloud

Computing and other government organizations are still just beginning to examine the

concerns and issues of cloud computing.

The General Services Administration (GSA) is one of the larger cloud computing

users. Recently, the GSA moved their primary information portal (USA.gov) to

Terremark’s Enterprise Cloud Service (Staten et al., 2009). In doing so, GSA migrated

all core resources for the USA.gov web portal to an IaaS platform giving them the ability

to deploy on-demand resources as web traffic increases. Migration to the cloud has

brought benefits and savings, such as avoiding idle server costs while still

accommodating web traffic spikes, acting on users’ requests in real time, and applying

security constraints on top of this platform (Staten et al., 2009).

The GSA and NASA are already using cloud implementations to their advantages

and realizing some benefits or using the services. As the success stories become more

prevalent, the future of cloud computing in the government sector will likely grow.

U.S. Government Public Sector Cloud Providers.

NASA recently launched NEBULA, a cloud computing based service that

provides highly-scalable, high performance, on demand infrastructure, platform, and

software as a service (Bambacus, 2010). Nebula is an open-source cloud computing

project developed to provide an alternative to building new data centers whenever NASA

requires additional data processing (NASA, 2010). Nebula’s IaaS provides scalable

computing and storage for NASA users’ scientific applications. Nebula enables

significant cost savings through better resource utilization, reduced energy consumption

and reduced labor costs associated with procuring infrastructure and creating new web

applications (NASA, 2010). Additionally, the Nebula cloud computing platform has

become the home of the Federal Government’s flagship website USAspending.gov

(Kundra, 2010). USAspending.gov 2.0 was completely reengineered to take advantage

of the cloud computing platform at Nebula. By tapping the capabilities of NASA’s

Nebula platform, unused capacity is available for use by other government agencies.

Beyond the NASA, the Department of Interior’s National Business Center (NBC)

provides payroll and personnel services for a number of government agencies (Wyld,

2010), as well. The NBC offers a prepackaged, integrated development environment

including a software development tool, applications and testing tools (NBC, 2010).

NBC’s private federal cloud gives federal users the advantage of using a pool of

networks, servers, storage capabilities, and desktop applications (SaaS) in a NIST-

certified secure dedicated federal environment. Federal organizations and agencies can

take advantage of end-to-end development and production pipelines on an as-needed

basis in a hosted environment (NBC, 2010).

These two examples of cloud computing service providers can be classified as

government sector clouds that provide for themselves and other government sector

consumers. By using available resources from other government agencies, the Federal

Government can increase the efficiency of its resources and reduces IT costs.

U. S. Government Sector Cloud Consumers.

GSA anticipates that cloud computing will become a major factor in reducing the

environmental impact of technology and help achieve important sustainability goals

(McClure, 2010), Besides the NASA, GSA and NBC, other examples of federal, state

and local government agencies using cloud-computing technology are listed in Table 1.

Table 1. Federal, State, and Local Agencies using Cloud Computing

(Kundra, State of Government Sector Cloud Computing, 2010)

FEDERAL AGENCIES

DoD (U.S. Army)

Army Experience Center

Department of the

Interior

Agency-wide E-mail

DoD (U.S. Army)

Enterprise E-mail

GSA

USA.gov

DoD (DISA)

Rapid Access Computing Environment

GSA

Agency-wide E-mail

DoD (DISA)

Forge.mil

NASA (Ames Research)

World-Wide Telescope

DoD (USAF) Personnel Services Delivery

Transformation NASA (Jet Propulsion

Laboratory) Be A Martian

DoE (Lawrence

Berkeley National

Labs)

Cloud Computing Migration

Social Security

Administration

Online Answers

Knowledgebase

Department of

Health and Human

Services

Supporting Electronic Health Records Recovery Accountability

and Transparency Board Recovery.gov Cloud

Computing Migration

STATE AND LOCAL AGENCIES

State of Colorado

(Office of

Information

Technology)

Launching an Enterprise Cloud

City of Canton (Georgia)

E-mail

State of Michigan

(Department of

Technology

Management and

Budget)

MiCloud

City of Carlsbad

(California)

Communication and

Collaboration Services

State of New Jersey

(NJ Transit

Authority)

Customer Relation Management

City of Los Angeles

(California)

E-mail and Office

Productivity

State of New

Mexico (Attorney

General’s Office)

E-mail & Office Productivity

City of Miami (Florida)

311 Services

Commonwealth of

Virginia (IT

Agency)

Application Development Platform

City of Orlando (Florida)

E-mail

State of Wisconsin

(Department of

Natural Resources)

Collaboration

Klamath County

(Oregon)

Office Productivity

State of Utah

(Department of

Technology

Services)

Cloud Computing Services

Prince George’s County

(Maryland)

School District E-mail

Foreign Government Public Sector Cloud Computing

Governments in other countries have adopted cloud computing technologies as

well. The government of the United Kingdom has created the “G-cloud,” a government-

wide cloud computing network, as a basis for funding a standardized environment for

running public services (Glick, 2009). European nations are also implementing IT

solutions around cloud computing services in health services, management of

government sector housing, transportation service networks, and education services

(Wyld, 2010). Japan’s Ministry of Internal Affairs and Communications plans to build a

massive cloud-computing infrastructure to support all the government’s IT systems

allowing the various ministries to integrate hardware and platforms to promote

standardization and consolidation of government’s IT resources (Rosenberg, 2009).

China’s efforts in cloud computing have been organized by local governments

and leaders. The government of Wuxi, in order to attract more firms to it’s local

economic development project, is working with IBM to build a cloud computing center to

provide on-demand computing services. (Wyld, 2010). The Vietnamese government and

universities are working with IBM to leverage the cloud computing model to help

establish a new department called Service Science Management and Engineering in

Hanoi (Nystedt, 2009). The government of Thailand is preparing to set up a private

cloud-computing platform in efforts to improve development and implementation of e-

commerce applications (Hicks, 2009).

Cloud computing is being implemented in both the commercial and government

sectors. Both sectors are taking advantage of using either using a public, private,

community or hybrid deployment clouds. While cloud computing is still in the beginning

stages in the government sector, commercial industry is more established in the use of

cloud computing. Unfortunately, the successes of cloud computing in the commercial

industry might not translate to success in the government sector. By examining current

IS theories, the perceived viability and its relationship to the willingness to implement a

new technology can be ascertained.

Current Theories

There are several theories that might be used to test the adoption, implementation

and success of information systems (IS) and information technologies (IT). Since the

seventies, research has contributed to develop a better of understanding of the causes for

low success rates of implementing costly information systems. The Technology

Acceptance Model proposes that perceived ease of use and perceived usefulness

influence other variables on technology acceptance (Davis, 1989). The DeLone and

McLean IS Success Model created a multidimensional measuring model with

interdependencies between different success categories (DeLone and McLean, 1992).

The Computer Self-Efficacy Theory demonstrated the utility of self-efficacy to

understand individual computing behavior (Compeau and Higgins, 1995). The Model of

PC Utilization confirms the importance of the expected consequences of using PC

Technology (Thompson et al., 1991). Task-Technology Fit states that IT will have a

positive impact if the IT capabilities match the task performed by the use (Goodhue,

1998). The Fit-Viability Model proposes that the fit and viability of the technology will

increase the performance (Liang, 2007), and the Fit-Appropriation Model, which argues

that IS performance is affected by fit and appropriation support (Dennis, et al., 2001).

These theories have been used to both predict and facilitate the use of IS. Several of

these theories are reviewed in the following pages.

Technology Acceptance Model (TAM)

The Technology Acceptance Model (TAM) was written by Fred Davis in 1986 as

a Doctoral dissertation and published in 1989 in Management Information Systems,

Quarterly (MISQ). TAM posits that perceived ease of use and perceived usefulness are

of primary relevance in the acceptance of computers and IT (Figure 5). Davis defined

“perceived ease of use” as the degree to which an individual believes that it would be

effortless to use a particular system. He defined “perceived usefulness” as the degree to

which an individual believes that use of a particular system would improve job

performance. Thus, TAM posits that, taken together, these two beliefs lead to a

behavioral intention to use the target information system and that it is this intention that

leads to actual system use. He tested his model and survey scales in both a real-world

setting and a lab experiment and found good support for it.

Since then, TAM has been used as a foundation to study many different

information systems and technologies. For example, Lederer et al (2000) extended TAM

to examine the World Wide Web (WWW). They revealed that the ease of understanding

a web site by a user and the ease of finding the web site by a user predicted ease of use

while information quality predicted usefulness in a revised web site. In another study,

Viability of TAM in Multimedia Learning Environments: A Comparative Study (Saade,

Nebebe, and Tan, 2007) extended TAM to create a Multimedia Acceptance Model. Their

testing determined TAM is a solid theoretical model, which can extend to the

multimedia-learning environment. These examples show the utility of TAM in predicting

users acceptance of a variety of technologies.

As applied to cloud computing, the Technology Acceptance Model predicts that

cloud computing services would be accepted as long as a user believed it would be easy

to use and would improve their job performance. Some of the benefits of cloud

computing, such as flexibility, increased efficiency, cost-reductions and reliability, can

affect whole organization. Cloud computing is not directly aimed at improving any

particular user’s job performance, but a benefit is that it allows the organization to devote

more resources to running the business instead of running an in-house IT department.

The TAM model focuses on the end-user of a system, whereas cloud computing affects

the entire organization. Therefore, this individual-level model is not the appropriate

model to use to answer an organization-level research question.

Figure 5. Technology Acceptance Model

Source: Davis (1989), Venkatesh et al. (2003)

Task Technology Fit (TTF)

While TAM states the perceived usefulness and ease of use are relevant to

acceptance of computers and IT, Task Technology Fit (TTF) (Figure 6) can be described

as “the extent that technology functionality matches task requirements” (Goodhue, 1995).

TTF argues that the use of information technologies can produce different results

dependent on the configuration of the technology and the specific task that is used being

accomplished. Much like TAM, TTF has been extensively tested.

Figure 6. Task-Technology Fit

The Task Technology Fit Theory proposes that when the technology fits the task

and user abilities, then performance gains should be expected. Goodhue and Thompson

studied TTF and Individual Performance (1995) and their new model asserts that

information technology must be used and be a good fit with the tasks that it supports to

have a positive impact on individual performance. Zigurs and Buckland (1998) studied

TTF and Group Support Systems. Their research showed that an appropriate

task/technology fit should result in higher performing groups. They found that fit was

explicitly defined and linked to group performance. They also found that task complexity

is a fundamentally important aspect of task and is relevant in a Group Support System

environment.

Fit has been studied at many levels. Task-Technology Fit (Goodhue, 1995 and

Goodhue and Thompson, 1995) measure fit at the individual level. Zigurs and Buckland

(1998) measure TTF at the group level. TTF has also been measured at the user level in

e-tourism, user computer self-efficacy and in knowledge management (Turner et al.,

2006). Usoro et al. (2010) combined TTF and TAM to study e-Tourism. They chose

TTF since it was reasonable to expect that the consumer will favor e-commerce

applications that match their shopping tasks. Another example of examining TTF was by

Gebauer, Shaw and Gribbins (2010). They built on previous TTF research and presented

a model establishing the fit between managerial tasks, mobile information technology,

and the mobile use context, supporting that a good fit positively affects task performance.

Cloud computing provides benefits at the organization level, so expanding TTF to

examine if the cloud computing technology would fit with the organizations computing

tasks/needs is the next step in examining TTF.

Fit-Viability Model (FVM)

The Fit Viability Model (FVM) expands on the task-technology fit model.

Anthony Tijan, founder and executive vice president of a consulting firm, developed the

original Fit-Viability Model (FVM) in 2001. Developed from working on more than 100

consulting projects with a wide range of companies that were examining Internet

initiatives, Tijan replaced two criteria used in portfolio analysis with business viability

and business fit. Business viability captured qualitative data about a likely payoff or an

investment and fit measured the degree in which the investment matches a company’s

processes, capabilities, and culture. Taken together, these two constructs were used to

predict the eventual performance of a portfolio. With only slight modifications, this

model has been used to specifically address the adoption of a new technology.

Liang and Wei (2004) studied the adoption of mobile technology in business. At

the time, there were few studies on how organizations decide on adopting new (mobile)

technologies and which factors determine the success or failure of adopting this new

technology. In the revised FVM (Liang and Wei, 2004; Liang et al., 2007), the Fit-

Viability model integrates task-technology fit with the general belief of organizational

viability of information technology. According to their model, fit measured the extent to

which a feature of a technology matches the needs of the task. Viability measured the

extent to which the organizational infrastructure is prepared for adopting the technology

(Figure 7).

Figure 7. Fit-Viability Model

In FVM, the TTF framework is modified to use more objective assessment of the

match between the task and the technology without considering the abilities of the

individual. In their Adoption of Mobile Technology in Business (2007), only the nature of

the technology and the requirement of the task were considered for fit. For mobile

technology, mobility and reachability were two features for measuring fit. If the

requirements of task coincide with these qualities, its fit with mobile technology would

be high and higher performance would be expected.

In the FVM, viability was defined as the extent to which the organization is ready

for the technology and was posited to be influenced by economic feasibility, technical

infrastructure, and the social readiness of the organization (Liang et al, 2007). Economic

feasibility was measured by assessing the cost benefit of an IT project to determine

whether it can bring financial or intangible returns and determine whether the IT project

could bring a competitive advantage. In the Fit-Viability framework (Figure 8), a high

viability and high fit would result in the technology being a good target to implement.

Figure 8. Fit-Viability Framework

In the FVM (Liang et al., 2007), the technical infrastructure was defined as being

composed of the IT platform and the information service required for supporting the

application. Liang et al., (2007) concluded that the technical infrastructure of an

organization includes computing, information management, and the associated

communication platform. The IT infrastructure provides the basis that supports

technological operations and augments business development.

The final factor of viability includes the social readiness of the organization.

Liang et al., (2007), state that user satisfaction and system usage are two common criteria

for evaluating the success of IS implementation. The organizational factors used in the

FVM model include the process reengineering, employee acceptance, and top

management support. Liang and colleagues saw the influence of business processing

reengineering, user competence and top management support as fundamental influences

indicating organizational readiness.

Several theoretical and practical contributions resulted from their study. They

concluded that organizations should consider both system fit and viability when

considering adoption of a new technology, mobile technology in this case. Previous

research focused on either the fit or organizational factors, which were not complete by

themselves. The Liang study validated the FVM framework and showed its practical

applicability. Using the framework in Figure 8, organizations can determine whether an

information technology is fit and viable for the organization, or whether changes need to

be made to the organization or the technology.

One limitation of this theory was that it needed to be expanded to other

technologies or issues other than mobile technology (Liang et al., 2007). One of the

goals of this thesis was to test a new model where fit and viability affect organizational

willingness to implement cloud computing.

Viability-Willingness Model for Cloud Computing

The fit-viability model was adapted to help answer this thesis’ second research

question. It was specifically adapted to help determine whether a new technology (i.e.,

cloud computing) is a viable option for implementing in an organization and if the

organization is willing to implement the technology. This new Viability-Willingness

Model (VWM) posits that the cost of the technology, along with the organizational inertia

and the fit of the technology leads to a perception of viability, which in turns leads to a

perception of willingness to implement the technology. Cost has both a direct and an

indirect relationship with perceptions of willingness. Further definitions and related

hypotheses are detailed below (Figure 9).

Figure 9. Research Model: Viability-Willingness Model

Cost.

Gartner (2008) states the total cost of ownership (TCO) model includes the cost of

hardware installation, software optimization, warranty and any license or maintenance

agreements, IT labor and PC support costs, maintaining security assurance and software

upgrades. Gartner (2008) also includes cost analysis of mobility requirements; migration

costs and planned long-term expenses such as capital expenditures, and lifecycle

replacement. The TCO model should act as a framework to help understand major cost

categories associated with cost-ownership. In addition, cost/benefit analysis must include

tangible and intangible costs and benefits including network stability and bandwidth

constraints. Cloud computing promises to cut operational and capital costs and let IT

departments focus on strategic projects instead of keeping datacenters operational (Velte

et al., 2010).

The VWM proposes that if costs of a new technology are perceived to be more

than the current technologies, then the perceptions of viability of the new technology

would be expected to decrease. Cost, while related to the perceived viability, could also

be directly related to the perceived willingness to implement the technology. Perceptions

of only a limited cost savings from implementing a new technology would be expected to

adversely affect the organization’s willingness to adopt that technology. Therefore it is

hypothesized that:

H1: Cost is negatively related to the organization’s willingness to implement

cloud computing in the Department of Defense.

H2: Cost is negatively related to the perceived viability of cloud computing in the

Department of Defense.

Organizational Inertia.

Organizational inertial is the degree to which members of the organization have

been motivated to learn, use and accept new systems (Seddon, 2010). The unit of

analysis chosen for assessing organizational inertia is the organization, not the individual.

In older organizations, systems and behavior tend to become institutionalized, acting as a

source of organizational inertia, which limits the ability of organizations to adapt

(Hannan and Freeman, 1977). As organizations age and get better at replicating routine,

they also become more inert (Hannan and Freeman, 1984). Hannan and Freeman (1984)

also argue that an organizational change creates a liability of newness that exposes

organizations to a higher risk of failure.

Change in an older organization creates new roles and new relationships similar to

those of a new organization (Brown, 2002). During implementation of the technology

and follow-on upgrades, substantial energy is spent on change-management, training, and

support to overcome organizational inertia. Each project is different, so organizational

inertia is measured for each project (Seddon, 2010). Since cloud computing has different

service models, implementing cloud computing is considered one project for this thesis

instead of multiple projects encompassing different services. Organizational inertia is

important for determining whether the organization is likely to accept change or not.

Since inertia limits the ability of the organization to adapt, high organizational inertia

would reduce the viability of a new technology. Thus:

H3: Organizational inertia is negatively related to the viability of cloud

computing.

Fit.

Drawing from Zigurs and Buckland’s (1998) definition of fit, we define fit as the

match of the computing platform with the computing needs of the organization. Fit is

important to determine whether cloud computing technology is suitable for the

organization’s computing needs. If the IT staffing personnel do not agree that it is a good

fit for the organization, then viability is expected to decrease. Therefore:

H4: The fit between cloud computing and the organization’s computing needs are

positively related to viability of cloud computing.

Viability.

Viability is defined as the extent to which the organization is ready to implement

new technologies considering the economic feasibility, IT infrastructure, and readiness of

the organization (Liang et al., 2007). The viability construct has three influences that

determine whether a new technology, cloud computing in this case, is viable for an

organization. The total cost of ownership of using or operating the technology

determines the economic benefits. The inertia of the organization determines if the

organization quickly adapts to new ideas and technologies. Finally, the technology must

fit the needs of the organization. Those three forces influence whether a new technology

will be viable. Viability is important in determining if the technology is of practical use

in the organization and helps determine the willingness to adopt a new technology.

H5: Viability of cloud computing is positively related to the willingness to adopt

of cloud computing.

Willingness.

Willingness is defined as being disposed to, inclined toward or openness to

something. While there are not many studies in willingness to implement or adopt

information systems, there are several conceptual models that discuss willingness.

Werner (2004) proposes that a willingness to use telemedicine was affected by the

participants’ attitudes towards telemedicine, the relationship between the patient and

physician and the level of anxiety toward technology. Chambers et al. (2003) examined

the impact of certain personality types and inclination to use technology. Using the

Myers-Briggs Type Indicator and a questionnaire designed to use technology, they found

that intuitive/thinking types of personalities were more likely to use technology. Turner,

Thomas, and Reinsch (2004) followed communication scholars in predicting that the

perceived attributes of a new technology will significantly affect the willingness to try the

technology. They argued that task situations (types of medical care in their study) would

affect the relative importance of the perceived attributes.

In applying the VWM to cloud computing, the perception of willingness is

defined as the openness to implement cloud computing in the organization. The

perceived viability of cloud computing and the cost of cloud computing are posited to

affect the willingness construct. Perceived willingness is how comfortable or open the

organization is to the new technology. Cost is directly and negatively related to

perceived willingness since the cost of implementing and using the technology could

affect whether the technology is implemented. In addition, viability is positively related

to the perceived willingness. If the technology is a viable option for the organization,

than there should be an increase in the willingness to implement the technology. Having

established the theoretical foundation for this thesis, chapter three discusses how the

subsequent research was conducted.

III. Methodology

The Research Strategy

This thesis examines the appropriateness of cloud computing as an IT model for

the DoD. To address the two research questions and test the hypotheses takes two

complimentary approaches. The first approach is to compare perceptions of cloud

computing between the commercial industry IT personnel and DoD IT personnel. The

second approach is to test the theory and relationships in the Viability-Willingness Model

for cloud computing in order to examine the relationship between viability and

willingness to implement a new technology. The results from these two efforts will

provide answers to the research questions.

A survey methodology was selected for both approaches. The survey method was

selected since its purpose is to describe attitudes, beliefs, and behaviors and is also a

common tool used for testing a certain theory or causal relationships. Surveys are

developed to describe the attitudes, beliefs and behaviors of a population (Patten, 2009).

Researchers select a sample of the population, study the sample, and then make an

inference to the population from the sample data. Surveys can also be quantitative

research. That is, results from the survey are easy to quantify which allows for statistical

analysis.

Survey Development

This survey was developed specifically to address the viability of cloud

computing and the willingness to adopt cloud computing. Throughout the research, it

was discovered that there were few studies that addressed cloud computing at the

organization level. The vast amount of surveys addressed the use or acceptance of a

technology that was already implemented. Since such approaches do not specifically

meet the needs of this research, a survey was exclusively designed for the VWM. The

only construct for which questions were previously validated was the organizational

inertia questions. The remainder of the questions was developed to address the constructs

in the model and to match a question in the IDC Enterprise panel survey. All questions

were tested in a pilot study to verify that they addressed the individual constructs. After

editing the questions, the survey was tested one final time before being distributed to the

sample.

Survey Structure

All questions that address hypotheses in the model are based on a five-point

Likert scale where 1 is represented as the strongly disagree or the most negative aspect of

the question and 5 represents strongly agree or the most positive aspect of the question.

The survey is broken down into multiple sections (Appendix B), each section addresses a

hypothesis of the model. The sections of the survey are: Cloud Computing, Economic

Impact, Organization Information, Reforming Federal Information Technology

Management, and Demographics.

The sections relate to hypotheses and have several specific questions that address

aspects of each hypothesis. The first part of the survey develops the respondents’

knowledge of cloud computing, how well cloud computing aligns with their

organization’s computing needs, and the issues surrounding cloud computing. The next

section, Economic Impact addressed economic aspects of implementing cloud

computing. The third section of the survey, Organization Information, examines the

effect of cloud computing in the organization. The fourth section, Reforming Federal

Information Technology Management, analyzes whether organizations are willing to

implement cloud computing using different types of service providers. The last section

records the demographics of the sample.

Survey Sample

A sample of convenience was used in this study since it was not possible to

contact all IT personnel across the Armed Services, the DoD, and the various DoD

civilians and contractors. The sample is from a list of IT personnel from the U.S. Army

Information Systems Managers Functional Area 53 list hosted by the U.S. Army Military

Academy (53Listserve).

Army Information Systems Managers were selected as the primary sample from

the DoD IT personnel population. Information Systems Managers are usually trained at

Fort Gordon, work in a variety of positions ranging from Automation Staff Officers, who

work on computers, networks, and manage information systems to positions in the DoD

and Army CIO/G6 that work in policy and procedures. Additionally, other DoD

information technology personnel were invited to participate in the survey.

Data Analysis

To determine the appropriate sample size needed to answer the research

questions, a power analysis was conducted according to the procedure outlined by Cohen

(1992). The power analysis revealed 64 survey responses were required to detect the

expected medium effect size statistical difference in the sample. Therefore, data

collection continued until 83 useable responses were obtained.

To compare the results on the question on the issues of cloud computing between

the two surveys, a Z-test is used. A Z-test compares the sample and population means to

determine if there is a significant difference between the samples when the sample is

large (n>30). The Z-value for the 95% confidence interval is +- 1.960. If the sample

returns a test statistic of less than 1.960 then we fail to reject the null hypothesis. If the

test statistic is more than +- 1.960, then we reject the null hypothesis since the samples

returned are significantly different.

Next, the appropriateness of the Viability-Willingness model was assessed using

SmartPLS 2.0. SmartPLS is a structural equation-modeling tool based on Partial Least

Squares (PLS). PLS has an advantage over traditional statistical techniques since it is

able to concurrently test the measurement and structural models without being covariance

based. Additionally, PLS is not constrained to data sets that meet homogeneity and

normality requirements (Chin et al., 2003). A significant advantage of PLS is that it can

handle smaller sample sizes relative to other structural techniques. These inherent

strengths make PLS a highly appropriate approach to analyzing the data set.

Using SmartPLS version 2.0 (Ringle, Wende & Will, 2005), the model was

evaluated to assess the measurement model and the structural paths between the

constructs. To obtain reliable results and t-values, 200 random samples of 83 were

generated using a bootstrap procedure. Finally, the hypotheses were evaluated by

assessing the sign and significance of the structural path coefficients using two-tailed t-

test statistics. Since SmartPLS does not calculate goodness-of-fit statistics. R2 values

were evaluated to assess the ability of various proposed relationships to predict a

significant degree of explanatory power in each construct and t-values were evaluated to

determine the strength of the various paths (Schuessler, 2009).

Instrument Validation

The validation process started by constructing a PLS model where the individual

survey questions were assessed to determine how well they measured their associated

construct. First, the internal consistency (reliability) statistics were examined. Reliability

tests determine if the set of variables are consistent with the intended item being

measured. All reliability measures meet the acceptable lower limits of .70, and one

exceeds the lower limit of .60 (see table 2) for a newly defined scale (Hair et al., 2006).

Next, the survey items are assessed for construct validity by performing a factor

analysis of each item in the survey and calculating the reliability of the resulting factors.

According to Hair et al. (2006), item loadings of .5 or greater represent items of practical

significance. After removing any items that that fail reach .5 on any factor, it was

determined if the items for each construct loaded higher on their own construct than on

other constructs. All survey items loaded above the threshold (see table 3) and were kept

for the remaining analysis.

Finally, determining discriminant validity requires testing the average variance

shared between a construct and its measures (AVE) (Gefen et al., 2000). The average

variance shared between the constructs and their measures should be greater than the

benchmark of .5 (Fornell and Larcker, 1981) and greater than the corresponding

correlations between constructs themselves (see table 4). The matrix supports the

discriminant validity of our scales in that the elements in the matrix diagonal are higher

than .5 in all cases, and higher than the off-diagonal correlations between the elements in

their corresponding row and column. Therefore, the variables and the constructs pass the

tests for reliability and validity as noted in chapter 4.

Chapter Summary

This chapter outlined the methodology of this research demonstrating one way to

analyze survey samples. By adapting the FVM to a Viability-Willingness model, we are

attempting to determine the perceived viability and the perceived willingness to

implement cloud computing in the Department of Defense.

IV. Analysis and Discussion

Chapter Overview

This chapter presents the results from the survey and the tests of the model

proposed in Chapter 2. First, the results from the cloud computing survey of DoD IT

personnel are compared with the survey results from the International Data Corporation

Enterprise Panel, 2009 Survey on Cloud Services. Since government and commercial

industries have different customers, stakeholders, and goals driven by the nature of the

two industries, a comparison of the question regarding the issues associated with cloud

computing is used. The second part of this chapter presents the results of the analysis of

the Viability-Willingness Model using SmartPLS.

Comparative Analysis: Commercial and government sectors

International Data Corporation conducted a survey of its commercial enterprise

panel in 2009 on the top challenges and issues of cloud computing. Two-hundred sixty-

three IT executives/CIOs and their line-of-business colleagues completed the survey

about their companies’ use of, and views about, IT Cloud Services. As part of the survey,

the respondents were asked to rate the challenges/issues ascribed to the cloud/on-demand

model (Figure 10). The Y-axis is the concerns with cloud computing, while the X-axis is

the number of personnel that stated the concern was significant. The survey used a 5-

point Likert scale ranging from 1=Not Significant/Concerned to 5=Very

Significant/Concerned.

The top three issues for the commercial industry as noted from the IDC survey

were security, availability, and performance (Figure 10). These issues were also rated as

the most significant in 2008. IDC explains this as a call for better service level assurance.

The number four issue in the survey was a concern that cloud computing might cost more

than current computing models. This perception seems to contradict what proponents of

the cloud model tout, which is that implementing cloud computing leads to cost savings

(Kundra, 2010). The issue of bringing the data or services back in house was rated as the

fifth most important. That is, the survey respondents were concerned that if the cloud

computing model did not work, it would be more difficult to move the services back

under the organization’s control. The last two issues that the survey members stated were

an issue was the ability to integrate with in-house services and the ability to customize

the service to what is needed. These are important since they both convey the amount of

organizational control over their data and services.

The IDC Enterprise panel survey from 2009 shows that there are still major

concerns about cloud computing. While it was expected to find that security,

performance and availability were issues with a majority of the respondents; on-demand

costs and the ability to bring the services back in house were not. There is still

uncertainty about the overall cost of cloud computing and whether it will be a success as

a computing platform for organizations.

Figure 10. IDC Commercial Enterprise Panel 3Q09 Survey Results

For comparison to the IDC Commercial Enterprise Panel survey, the 2010-2011

cloud computing in the military survey done in this research sampled DoD IT personnel

(n=83) on the same issues of cloud computing. On the military IT survey, the

respondents rated the challenges/issues ascribed to the cloud/on-demand model (Figure

11). The Y-axis is the concerns with cloud computing, while the X-axis is the number of

personnel that stated the concern was significant. Both surveys used a 5-point Likert

scale ranging from 1=Not Significant to 5=Extremely Significant. Following the IDC’s

method, the results were taken from the respondents that selected 3, 4, or 5 and measured

in percentages. For an accurate comparison of commercial and government industries,

scales and questions were maintained for data integrity.

Figure 11. Military IT Personnel Survey Results

A comparison of the results of the question of rate the challenges/issues ascribed

to the cloud/on-demand model shows similarities and differences between the two

samples (Figure 12). The results are taken, and measured in percentage of the

respondents that selected 3, 4, or 5.

Figure 12. Comparison of Survey Results

Availability.

Availability is the first issue compared between the two surveys. Availability is

the state of being able to ensure users can use any information resource when ever and

whenever it is needed. There is a significant difference (Z = 3.464) between the numbers

of military IT personnel (98.8%) that rated Availability as a significant issue compared to

the number of IDC Enterprise personnel (82.9%). Availability depends on the

accessibility of the data, the system, the applications and the infrastructure used to access

that data. The loss of availability is critical if there is a serious incident, network failure,

or natural disaster. The difference between the two surveys for availability is significant.

For the military IT personnel, the availability of the data and services could be critical to

the completion of mission in a combat zone. For example, the Army mission is to

provide to combatant commanders the forces and capabilities necessary to execute the

National Security, National Defense, and National Military Strategies. The Army’s

Signal Corps supports this mission by providing and managing communications and

information systems support for the command and control of combined arms forces. The

lack of availability of the network, data, or information systems could affect the mission.

The commercial industry does not have the same mission as the Department of Defense.

Where loss of availability would impact the commercial organization in terms of time

and profit, it might not affect National Security, which would be a reason for the

significant difference in the number of personnel that annotated availability as a

significant concern.

Performance.

The next issue selected as significant was performance. Cloud computing

performance can either be client-oriented or cloud-oriented (Linthicum, 2010). Client-

oriented performance is where users constantly interact with the cloud provider where

there is latency with the constant back-end machine-to-machine communications that

occurs between the SaaS provider and the browser. Cloud-oriented performance is where

the processing occurs in the “cloud” and is compared to the performance of completing

the processing on-premise. There is a significant difference (Z = 2.553) between the

numbers of military IT personnel (95.2%) that rated Performance as a significant issue

compared to the number of IDC Enterprise personnel (83.3%). While it was not defined

as whether performance was client-oriented or cloud-oriented performance, more military

IT personnel thought performance was an issue. Client oriented performance issues may

not be as noticeable in terms of latency unless there is network saturation (Linthicum,

2010). Whereas cloud-oriented performance can be an advantage, performing large

amounts of processor intensive calculations or queries can take many hours on the local

network. The scalable nature of the cloud allows additional processors to be quickly

added resulting in calculation or queries taking minutes to complete. Commercial

industry has more experience with cloud computing and a lower number of the IDC

Enterprise panel survey respondents stated that performance is an issue. For DoD IT

personnel, latency on a bandwidth constricted tactical network is not acceptable and

therefore more survey respondents would rate this as an issue.

Security.

While respondents from both surveys feel that security is a significant issue in

itself, there was no difference (Z = 0.537) between the numbers of DoD IT personnel

(90.4%) and the IDC Enterprise sample (87.5%). While security was not explicitly

defined in either survey, as stated in chapter 2, FISMA defines three security objectives

for information and information systems that serve as the basis for NIST’s analysis:

Confidentiality, Integrity, and Availability. In both surveys, the respondents thought

security was a significant issue. Security of the cloud service, whether protecting mission

essential data or services for the military or protecting the security of the customers’ data

and information for the commercial sector, is important for their organization. Losing the

security of data or information in the cloud for government or commercial industries can

have severe consequences. The security of the cloud may provide a roadblock on its

implementation.

Integrate with in-house IT.

The ability to integrate cloud computing offerings and services with in-house IT is

the next issue compared between the two surveys. There is not a significant difference (Z

= 1.009) between the numbers of military IT personnel (83.1%) that rated the ability to

integrate with in-house IT as a significant issue compared to the IDC Enterprise

personnel (76.8%). Organizations that use cloud computing want to maximize the

control of their business core systems, which can be in-house legacy systems and

integrate these systems across externally sourced cloud services (Gens, 2008).

Both government and commercial industry can feel this integration between in-

house and cloud computing systems is more of an issue. The military has a large number

of program-managed systems that perform specific functions. For example, the Army

uses a number or Army Battlefield Command Systems (ABCS) that provide specific

services for the commander and integration with these systems might prove to be

difficult. In the commercial industry, an organization’s inventory management systems

could be core business systems that is proprietary and could be difficult to integrate with

cloud services. In both cases, these systems could be more difficult to integrate with

other systems that use cloud services.

Ability to customize services.

Next, the number of personnel stating that the ability to customize software and

applications is a significant issue is compared between the two surveys. There was no

difference (Z = 0.223) between the numbers of military IT personnel (73.5%) that rated

the ability to customize services as a significant issue compared to the IDC Enterprise

personnel (76.0%). One of the advantages of cloud computing is the ability to quickly

customize software, platforms and cloud infrastructures. Organizations want to

customize “off-the-shelf” cloud services and tailor these services to the needs of their

businesses (Gens, 2008). It is plausible that the need for better fitting services drove a

large number of respondents in both surveys to report customization as an issue with

cloud computing.

On-demand costs.

On-demand cost is the next issue in the survey. On-demand costs are costs

incurred when using cloud computing in the organization. There is a significant

difference (Z = 2.878) between the numbers of military IT personnel (65.1%) that rated

On-demand costs as a significant issue compared to the IDC Enterprise sample (81.0%).

The on-demand cost issue may seem to contradict the reason for moving to a cloud

computing technology, given that cost savings are purported to be a benefit. One reason

behind this discrepancy is organizations fear that employees may use the service more

than what is budgeted. Commercial industry is conscious of what they spend and of their

profits, where the government sector does not necessarily see the bottom line, but they

work within budgets. The difference in the number of respondents stating on-demand

costs is an issue between the two samples is expected. The commercial industry has

more experience in implementing cloud computing, therefore they may see on-demand

costs exceed their initial expectations whereas the government sector is just beginning to

implement cloud services. Cloud computing can allow their organization to save money

on infrastructure (Del Nibletto, 2010), IT staffing and maintenance costs (Kundra, 2010).

If on-demand costs start to exceed the previous budget for the same type of services, then

the cloud computing service model would become an issue.

Bringing IT services back in-house.

Bringing IT services back in-house is the last issue compared between the two

surveys. There is a significant difference (Z = 4.202) between the numbers of military IT

personnel (56.6%) that rated bringing IT services back in-house as a significant issue

compared to the IDC Enterprise personnel (79.8%). Organizations wonder whether using

cloud services will lead to the same type of proprietary services that are dealt with today.

Proprietary services and software make it difficult to move services back in-house if they

are not satisfied with the cloud (Gens, 2009). Military IT personnel either do not think

they will move to cloud computing or only move limited services to the cloud and leave

core processes in-house. Therefore, there would be less military IT personnel that

thought this was an issue compared to their civilian counterparts.

Summary.

The comparative analysis of the IDC Enterprise panel 2009 IT Cloud Services

Survey and the 2010-2011 military IT personnel cloud computing in the military survey

had interesting results. More military IT personnel cited availability, performance and in-

house integration as significant issues than did personnel on the IDC Enterprise panel.

There was also a significant difference between the number of IDC Enterprise panel

personnel citing on-demand costs and bringing the IT services back in-house as concerns

than the number of personnel in the military IT survey. Cost may be expected to be more

important to commercial industry than the government sector, but cloud computing costs

are supposed to be lower than current computing methods and technologies. Conversely,

there was no difference in the number of respondents that rated security as a significant

issue, nor was there a difference in the number of respondents who saw the ability to

customize software as a major issue. Interestingly, stating that on-demand costs are a

significant issue goes against the reasons for implementing cloud computing. Bringing

services back in-house could be an issue if the cloud computing model does not work.

Since more than 79% of the commercial industry survey respondents rate this as an issue,

it could mean there is a lack of confidence in the long term viability of the cloud

computing model.

The survey comparison looked at the issues and concerns with cloud computing.

These issues are part of the viability of cloud computing. The needs of the organization

are considered when determining if the new information technology, or cloud computing,

is a viable technology. As noted above, there are ideas in the military about cloud

computing, which directly affect the perceived viability and perceived willingness to

implement the technology.

The remainder of the data in the military IT cloud computing survey was used to

examine the perceived viability and its relationship with the willingness of IT personnel

to implement cloud computing in the military. The following section examines the

relationship between the perceived system viability and perceived willingness to

implement cloud computing.

Discriminant Validity and Reliability

The Viability-Willingness Model for cloud computing looks at the relationship

between the perceived viability and perceived willingness. In the VWM, perceived

viability is directly determined by cost, organizational inertia and fit. Viability and cost

determine organizational willingness to implement. To test the model and its constructs,

the validity and reliability must first be tested.

Discriminant validity and reliability of the constructs were tested using Conbach’s

Alpha, as recommended by Hair et al. (2006). Reliability tests determine if the set of

variables are consistent with the intended item being measured. The internal consistency

(reliability) statistics using PLS results were: .691 for Fit; .845 for Inertia; .785 for

Viability; and .904 for Willingness (Table 2). The reliability thresholds for Inertia,

Viability and Willingness meet the acceptable lower limits of .70, while Fit exceeds the

lower limit of .60 for a newly defined scale (Hair et al., 2006).

Table 2. Reliability Statistics

Next, the convergent validity was determined by examining whether all items

loaded highly on their respective construct on PLS. A common rule of thumb is a

loading greater than .70 (Yoo and Alavi, 2001). In the Viability-Willingness Model, all

items loaded on their respective constructs from .783 to .899 (Table 3). All values lower

than .40 were removed for easier interpretation.

Table 3. Convergent Validity

In the cost construct, COST2 and COST3 did not load on the Cost construct and

were removed from the model. Additionally FIT3, FIT4, FIT5, FIT6, WILL5, WILL6,

WILL7, and WILL8 did not load on adequately on their construct and were therefore

removed from the model.

Determining discriminant validity requires testing the average variance shared

between a construct and its measures (AVE) (Gefen et al., 2000). Table 4 presents the

AVE matrices and construct correlations for the model. The matrix supports the

discriminant validity of our scales in that the elements in the matrix diagonal (AVE

results are bold) are higher in all cases than the off-diagonal elements in their

corresponding row and column. Each construct AVE should be higher than the

benchmark of .60 as recommended by Fornell and Larcker (1981). In addition, the AVE

must be larger than its correlation with other construct, and each time should load more

highly on its assigned construct than other constructs (Gefen, 2000).

Table 4. Discriminant Validity

Results

After assessing discriminant reliability and validity, the model is then processed

using SmartPLS. The results are depicted in Figure 13. All hypothesized relationships

except for one were supported. Path significance was estimated using a bootstrapping

procedure with 200 resamples, which tends to provide reasonable standard error estimates

(Mathieson, et al., 2001; Ravichandran and Rai, 2000). There was not a significant direct

affect of Cost on Willingness (-0.132 n.s.), which means that cost was not directly related

to Willingness (H1). There was a significant direct effect of Cost on Viability (-0.188*),

which supported H2: Cost is negatively related to Viability. Hypothesis 3 theorized that

Inertia is negatively related to and has a direct effect on Viability. PLS showed that

Inertia had a significant affect Viability (0-0.332*), supporting H3. H4 hypothesized that

Fit would be positively related to Viability. H4 was supported in that Fit had a

significant direct affect on Viability (0.601*). Lastly, H5 stated that Viability was

positively related to Willingness. Viability had a significant direct affect on Willingness

(0.457*) and therefore supported H5.

Figure 13. Viability-Willingness Model Structural Results

The Viability-Willingness Model explained 48.5% of the variance for viability

perceptions and 26.2% of the variance for the perceived willingness to implement cloud

computing. The model measures whether cost, inertia, and fit are significant predictors

of viability. Viability of cloud computing is partially formed by cost, organizational

inertia and the fit of the technology with the needs of the organization. Willingness is

only partially predicted by the viability of cloud computing. Performance, availability

and security might be significant factors in predicting viability since over 90% of military

IT personnel respondents cited those issues as significant.

The other, unmeasured, perceived attributes of cloud computing could also have a

direct affect on the perceived willingness to implement it. While viability contributed to

26% of the variance explained, the attitude towards cloud computing might be more of a

factor than whether the technology is viable for the organization. All questions on the

issues of cloud computing were selected by more than 50% of the respondents in the

military IT cloud computing survey as being significant. The perception that cloud

computing has multiple significant issues reduces the willingness to implement the

technology. Therefore, just because cloud computing is viable for an organization does

not mean there is going to be a willingness towards implementation; nonetheless it is a

major factor.

Chapter Summary

The collected data from the 2010-2011 Cloud Computing in the Military survey

suggests that cost, inertia, and fit were significant in predicting viability; additional

modifications to the survey instrument would increase the explained variance at the cost

of the parsimony of the instrument. This modification would also increase variance

explained in willingness, although adjusting the survey to include personality and attitude

might also significantly increase the perceived willingness.

V. Conclusions and Recommendations

Chapter Overview

This chapter reviews the two research questions and presents the conclusion of

this study. Sections discussing research limitations, follow-on research, and

recommendations for future research follow this section. In the general conclusion, the

benefits of this research are highlighted.

Research questions and Conclusions

This research was an exploratory examination of the opinions military IT

personnel of cloud computing in the in the Department of Defense and the Army. Its

research was motivated by desire to understand two overall research questions:

(1) Is cloud computing perceived as a viable technology in the DoD/Army?

(2) Is there a perceived willingness to implement cloud computing in the

DoD/Army?

The research provided an understanding whether military IT personnel perceive

cloud computing as a viable technology for their organization as well as the connection

between the perceived viability of new technologies and the willingness of IT personnel

to implement them.

A Difference Between Government and Commercial Industry.

A comparative analysis between IDC Enterprise Panel personnel and military IT

personnel demonstrated several differences between commercial and government sectors.

Availability, performance, and integration with in-house IT were identified by a higher

percentage of military IT personnel as significant issues than the IDC Enterprise panel.

Security and the ability to customize had the same significance between the two groups,

while the IDC Enterprise panel rated on-demand costs and bringing the IT services back

in house as more of an issue than did military IT personnel. The military IT personnel

appear more worried about availability, performance and integrating with in-house IT

systems and less concerned with cloud computing costs and the ability to bring services

back in-house at a later date.

Perceived Viability of Cloud Computing.

The Viability-Willingness Model was an exploratory model designed to

understand the interaction of the perceived viability of cloud computing and the

willingness to use it. While the model measured 48.5% of the variance in the viability of

cloud computing, there are other factors discovered in the survey that might also predict

the viability of cloud computing. The VWM demonstrated that cost, inertia and fit play a

role in the determining the viability of cloud computing.

Perceived Willingness to Implement Cloud Computing.

Perceived willingness to implement cloud computing can be more of an attitude

towards cloud computing. The VWM examined viability and cost as measures that

would have a direct affect on the perceived willingness to implement. The variance

explained was 26.2%, which indicates other measures would also have a direct affect on

the perceived willingness. Other measures that could affect the willingness to implement

cloud computing include personality traits (Chambers et al., 2003), attitudes towards the

technology (Werner, 2004), and actual task situations that would affect the relative

importance of the perceived attributes of the technology (Turner, Thomas, and Reinsch,

2004). Successful implementation of a new technology (cloud computing) could partially

depend on the willingness to make the changes necessary to make the technology work.

Limitations

There are several limitations to this research. The sample size (n=83) was large

enough to perform the needed analysis. However, a larger sample might allow for

additional statistically robust analysis, which may lead to an increased accuracy of the

inferences about the population, and refined detection of effects in the model. A broader

sample will expand the available tools used for analysis to include covariance-based

analysis. Expanding the sample demographic to include additional IT personnel more

evenly across the Department of Defense and the Federal government would allow for

additional analysis. By increasing the sample frame to include every member of the DoD

IT community from which the sample is taken increases the randomness of the survey

and reduces any biases in the sample.

The survey instrument had limitations. As an exploratory study, some items did

not initially load properly and therefore were not included in the analysis. This might

have occurred due to a misunderstanding of words in the question or the question itself.

It can also happen if the questions of the survey measured something other than the intent

of the construct. One set of questions was previously validated and two others loaded

properly on their respective constructs. Using previously validated questions or survey

instrument can lead to an increase of rigor, an increase of data quality and prevent

inaccurate conclusions. The wrong instrument of questions could increase the error in the

model and analysis.

Recommendations for Future Research

This research provided a starting point on determining the perceived viability and

perceived willingness to implement cloud computing in the military. Follow on research

should include increasing the sample frame and sample size to further validate this study.

Other areas for research should include perception differences among armed services,

different IT skill sets, and duty positions. Furthermore, additional research into the top

issues of cloud computing would be valuable in determining why those issues are thought

to be so significant. Additional research is also needed to clarify the relationship between

viability and willingness. This research highlighted that a technology can be a viable in

the organization, yet the organization still won’t implement it. Additional research

should be conducted on why an organization is not willing to implement a technology it

sees as viable. Further analysis is also needed on the difference between government and

commercial industry and the possible implications of implementing a new information

technology.

The VWM can be refined through additional research to increase the variance

determined in viability and willingness. A redesign of the cost construct by adding more

dimensions would help to see if a multidimensional construct plays a greater role in the

viability and willingness to implement a new technology. There may be specific parts of

the cost of implementation that affect viability and willingness differently. The

willingness construct should be examined for the influence of personality characteristics,

attitudes toward the new technology and possibly positive and negative affect. A more

in-depth study examining the perceptions of the different types of deployment cloud

models would provide the framework that organizations could use in examining the

viability and willingness to implement the latest technology trends.

Conclusion

Overall, findings suggest that cloud computing may be a viable option for the

Department of Defense, but willingness to implement cloud computing is not the same as

viability. This research indicates that simply because a technology has the potential to

improve an organization, it still may not be implemented. This thesis defined criteria for

measuring the viability and willingness to implement cloud computing. Viability was

affected by cost, organization inertia, and the fit of the technology with the organization.

Willingness was partially determined by viability perceptions, which supports previous

research that organizational attitude toward the technology also affects the willingness to

implement a new technology.

Today, the federal government is concentrated on reducing the budget deficit,

while completing the same missions and functions at a reduced cost. Whether it is

freezing pay wages, reducing work force, reengineering business processes or improving

efficiencies, technology can be used as an enabler for working smarter and more

efficiently. There are several areas discussed in the thesis that demonstrated the ability to

use cloud computing to support non-core business functions. Different commercial

industries use cloud computing differently. The DoD can learn from other government

agencies and commercial industries in ways to use cloud computing to reduce costs,

either through reduced hardware purchases, reorganization of personnel and jobs, or by

shifting non-mission essential processes to a cloud provider which would reduce

personnel needed to perform those functions. This research showed that just because

cloud computing is a viable alternative to the current desktop computing platform, the

organization might have a perception that cloud computing would not work. This

perception could hinder how quickly cloud computing is accepted in the DoD.

Appendix A: SRB Information

Appendix B: Survey Instrument

Appendix C: Survey Data

Appendix D: SmartPLS Report

Cross Loadings

cost fit inertia viability willingness

COST1 1.000000 -0.264952

-0.175177 -0.289511 -0.264672

FIT1 -0.236876 0.852295 0.185243 0.481584 0.320945

FIT2 -0.227624 0.894746 0.085833 0.564070 0.368675

INERTIA1r -0.109283 0.109694 0.827469 -0.201456 -0.058347

INERTIA2r -0.218048 0.099498 0.858527 -0.148874 -0.004516

INERTIA3r -0.110787 0.160601 0.811124 -0.196712 -0.049526

INERTIA4r -0.175709 0.115351 0.783988 -0.085278 0.101325

VIABILITY1 -0.203887 0.374463 -0.260587 0.804765 0.319953

VIABILITY2 -0.178400 0.452789 -0.320048 0.871319 0.528174

VIABILITY3 -0.342609 0.662744 0.049104 0.829887 0.371739

WILL1 -0.243832 0.270462 0.022472 0.431683 0.867694

WILL2 -0.239663 0.453971 -0.002378 0.432329 0.876659

WILL3 -0.204982 0.291138 -0.020901 0.454964 0.899804

WILL4 -0.245830 0.383365 -0.091208 0.428251 0.882597

AVE

cost 1.000000

fit 0.763489

inertia 0.673584

viability 0.698519

willingness 0.777512

Outer Loadings

cost fit inertia viability willingness

COST1 1.000000

FIT1 0.852295

FIT2 0.894746

INERTIA1r 0.827469

INERTIA2r 0.858527

INERTIA3r 0.811124

INERTIA4r 0.783988

VIABILITY1 0.804765

VIABILITY2 0.871319

VIABILITY3 0.829887

WILL1 0.867694

WILL2 0.876659

WILL3 0.899804

WILL4 0.882597

Path Coefficients

cost fit inertia viability willingness

cost -0.188419 -0.132310

fit 0.601123

inertia -0.332106

viability 0.457193

willingness

I Can, but I Won’t- An Exploratory Study on People and New Information Technologies in the Military.

Students also viewed