ECONOMIC AND TECHNICAL ASSESSMENT OF CLOUD

COMPUTING

Serdar Yarlikas

and Semih Bilgen

Informatics Institute, Department of Information Systems, Middle East Technical University, 06800, Ankara, Turkey

Department of Electrical and Electronics Engineering, Middle East Technical University, 06800, Ankara, Turkey

Keywords: Cloud computing, Economic variables, Technical variables.

Abstract: This position paper presents the current state of the research that aims to assess the economic and technical

benefits of cloud computing and to uncover best practices and lessons learned by service providers as well

as clients. In the first phase of the study, based on a review of the recent literature on cloud computing

principles, applications, achievements and challenges, variables that determine cloud performance and

benefits have been identified. Then, axioms and hypotheses on the interrelationships between the variables

have been formulated. Verification of the conceptual network of axioms and hypotheses entails exploratory

and validatory case studies. The paper elucidates the strengths and weaknesses, while pointing out the

opportunities and threats involved in cloud operation and adaptation, specifically in private organizations.

1 INTRODUCTION

We examine cloud computing practices from

technical and economic viewpoints. We construct a

conceptual framework of relationships between the

variables related to enablers and obstacles of

migration to cloud environments. Once the

hypotheses on those relationships are formulated

based on a survey of the relevant literature, they will

be further investigated through case studies.

2 THEORETICAL FRAMEWORK

Within the context of this study, cloud computing is

defined as a mode of Internet-based large-scale

distributed computing whereby shared software and

information resources are provided to users over

dynamically scalable and virtualized computing

facilities. In this mode of computing, applications

can be accessed over the Internet without prior

installation, as Software as a Service (SaaS).

Different web-based enterprise solutions such as

Customer Relations Management (CRM), Enterprise

Resource Planning (ERP), Human Resources

Management (HRM), Business Process Management

(BPM), and Expense Management (EM) will also be

investigated if they have been adapted to cloud

platforms, and as such, can be classified as SaaS

solutions. Within the literature that discusses the

critical deterrents for cloud computing, some authors

have concentrated on economic challenges, whereas

others have also considered technology related

issues (Armbrust, et al., 2010; Leavitt, 2009; Jiang

and Yang, 2010).

2.1 Technical Variables

The technical variables to be studied have been

identified as below. Unless specified otherwise,

possibly in arithmetical terms, these are generally to

be measured in a categorical scale such as (low,

medium, high). When categorized otherwise, an

ordering is defined on the categories so that the

hypotheses that refer to such ordering can be validly

presented as in Section 3.

Security refers to a broad set of policies,

technologies, and controls deployed to protect data,

applications, and the associated infrastructure of

cloud computing. Cloud security (CS) also includes

data protection, identity management, physical and

personnel security, application security and privacy

(Wang, 2009).

Network virtualization (NV) is a method of

combining the available resources in a network by

splitting up the available bandwidth into channels,

each of which is independent from the others, and

each of which can be assigned (or reassigned) to a

particular server or device in real time. Possible

167

Yarlikas S. and Bilgen S..

ECONOMIC AND TECHNICAL ASSESSMENT OF CLOUD COMPUTING.

DOI: 10.5220/0003960701670172

In Proceedings of the 2nd International Conference on Cloud Computing and Services Science (CLOSER-2012), pages 167-172

ISBN: 978-989-8565-05-1

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

levels of virtualization have been described, in

increasing levels, as para-virtualization, hardware-

assisted virtualization, live-migration, and pause-

resume. (Uhlig, et al., 2005; Clark, et al., 2005;

Travostino, et al., 2006; Padala, 2010).

Cloud deployment type (CDT) refers to the

ownership and operational categorization of the

computing infrastructure. Four major deployment

types are: public, private, community and hybrid

clouds (Chebrolu, 2011).

Upgradeability refers to the capability of

modifying network structure to improve the

performance in an easy and efficient way (Gutierrez,

Riaz, Pedersen, Madsen, 2009).

Manageability (MNG) is defined as the

collective processes of deployment, configuration,

optimization, and administration during the lifecycle

of cloud computing systems and services. Related

metrics are proposed as checklist of manageability

functions, number of steps to manage towards

desired state, time to manage, documentability,

elasticity of management, availability and continuity

of management and ease of use (Cook, et al., 2011).

Alternatively, manageability means the ability to

have visibility and control over services and usage,

enabled by cloud programmatic management

interfaces, cloud web management capabilities, self

service provisioning (Citrix, 2010).

Availability is the percentage of time the cloud

computing system is available for use via possibly

multiple cloud computing providers (Armbrust et al.,

2010; Helft, 2009; Beard, 2009).

Scalability (SCL) refers to the ability of a system

to grow in one or more dimensions as more

resources are added to the system. These dimensions

include the number of concurrent users that can be

supported and the number of transactions that can be

processed in a given unit of time (WebLogic, 2011;

SearchCloudComputing, 2011; CCD, 2011).

Agility (AGL) refers to the ability to rapidly

change the cloud application to meet user

requirements. (Citrix, 2010).

Operational efficiency (OEF) is the capability of

a cloud vendor to deliver cloud services to its

customers in the most cost-effective manner possible

while still ensuring high quality of its service and

support (Webopedia, 2011).

Data locality (DL) refers to the phenomenon that

the collection of the data locations referenced in a

short period of time in a running computer often

consists of relatively well predictable clusters

(Denning, 2005).

Resilience (fault tolerance) (RSL) is the property

that enables a system to continue operating properly

in the event of the failure of some of its components

(SunGard, 2010).

Isolation failure refers to the failure of

mechanisms that separate data that belong to

different organizations sharing a cloud (Harris and

Alter, 2010).

Concurrency (CCRCY) is a property of systems

in which several computations are executing

simultaneously, and potentially interacting with each

other (Expert Group, 2010; Kovachev et al., 2011).

Vendor lock in degree (VLID) denotes the level

of the customer dependency on a vendor for

products and services (Armbrust et al., 2010; The

Open Group, 2010).

Cloud performance (CP) is determined by

various metrics referring to network performance,

application performance and data center

performance. These three performance components,

are, in turn, influenced by all software and hardware

infrastructures. Network performance is measured

mainly by throughput and latency metrics. (Li, et al.,

2011; Myerson, 2011).

Cloud infrastructure performance, on the other

hand, can be evaluated with such metrics as CPU

(Central Processing Unit) utilization and application

traffic (Henkel, et al., 2007

Distance (DIS) is the relative location of the

cloud customer to the cloud providers and the other

cloud customers. Distance impacts response times,

thus cloud application performance (Weinman,

2011).

2.2 Economical Variables

The economic variables to be studied have been

identified as follows:

CAPEX (Capital Expenditure) includes the cost

of entire data center infrastructure, including servers,

storage arrays, software licenses (when needed),

routers, and load-balancers. CAPEX means funds

used by a company to acquire or upgrade physical

assets such as equipment. (The Open Group, 2010).

Total cost of ownership (TCO) can be defined

arithmetically as (Neville, 2010; The Open Group,

2010):

TCO= (Network management expenses+ the

expenses of services)/ (number of customers of

service providers), or

TCO= (Network costs+ Router/switch

costs+personnel costs + training costs + hardware

costs + software costs + endpoint costs+ energy

costs+ facilities costs)/ (number of customers of

service providers).

Outage duration cost (ODC) is the product of

CLOSER2012-2ndInternationalConferenceonCloudComputingandServicesScience

168

the length of outage duration, the number of

customers served, and a fixed customer outage

duration cost (Kuntz, et al., 2001).

Total cost of outages= f (duration of outage, size

of data center, business disruption, lost revenues, IT

equipment failure, accidental/human error), where

Duration of outage= f (length of outage duration,

the number of customers served, fixed customer

outage duration cost). Outage duration cost refers to

the cost of losing access to the data, accrued when

the system fails to provide or perform its primary

function in a period of time. Outage duration cost

can also be calculated by multiplying outage

duration hours with amount of money lost per hour

(Newvem, 2011; Kuntz et al., 2001).

Switching cost refers to the effort and expense

involved in switching from one cloud provider to

another cloud provider in terms of cloud services

and products (Klemperer, 1987).

Cloud migration cost is the cost of carrying all

applications from the current computing

environment to cloud environment. It can be defined

mathematically as follows (The Open Group, 2010;

Blaisdell, 2011):

Cloud migration cost=Setup costs + Migration

costs + Ongoing costs

Setup costs+ migration costs= The costs of

moving application and business moving to the

cloud.

Ongoing costs= f (CPU time, GB of RAM,

terabyte of storage)

Data replication cost refers to the cost of sharing

data in order to ensure consistency between

redundant resources, such as hardware and software

components, to improve reliability, fault tolerance,

accessibility when the data is stored on multiple

storage devices (CNTC, 2003). Of the many factors

that have a bearing on data replication costs, support

for heterogeneous storage, licensing fees, and

network bandwidth requirements are the most

significant (CNTC , 2003).

3 AXIOMS AND HYPOTHESES

In this section, relations between variables that are

not specific to cloud systems or that are self-evident

are listed as axioms, to be followed by cloud-

specific relations referred as hypotheses to be

verified by case studies.

3.1 Axioms

A1: In case of failure of mechanisms that separate

data of organizations sharing a cloud, security

decreases (Harris and Alter, 2010).

A2: Vendor lock-in degree causes switching cost

increase. (Armbrust et al., 2010; The Open Group,

2010).

A3: If availability increases, outage duration cost

decreases because the system can perform its

primary function and the effectiveness of cloud

applications also increases (The Open Group, 2010;

Wilson, 2008).

A4: Increasing agility reduces cloud application

migration cost (Khalidi, 2011).

A5: Increasing resilience increases availability

(SunGard, 2010).

A6: If scalability increases, cloud performance

increases (CCD, 2011).

A7: If availability increases, scalability increases

(Armbrust et al., 2010; Helft, 2009; Beard, 2009;

CCD, 2011; Varia, 2010).

: Agility increases upgradeability (Kundra,

2011).

3.2 Hypotheses

Hereafter, a (-) or (+) between two variables denote

that an increase in the first variable causes a

decrease or increase, respectively, in the second one.

H1a: Public cloud deployment (PCD) (-)

vendor lock-in degree

The degree of vendor lock-in depends on the

type of cloud deployment (Chebrolu, 2011), i.e.

private, community, public or hybrid cloud

(Chebrolu, 2011). Public cloud deployment causes

lower vendor lock-in degree because cloud vendor

provides applications available to general public in

this deployment type and so cloud customer

dependency on cloud vendor decreases.

H1b: Private cloud deployment (PRCD) (+)

vendor lock-in degree

Private cloud deployment increases vendor lock-

in degree (Chebrolu, 2011).

H2: Network virtualization (-) security

In the cloud environment, all resources are not

virtualized. If a user can access the data of other

users, this is considered “problematic virtualization”

hence, network virtualization causes reduction of

security (Armbrust et al., 2010).

H3: Level of virtualization (LOV) (+)

manageability

Virtualization generally enables each application

to be encapsulated such that they can be configured,

deployed, started, migrated, suspended, resumed,

stopped, and thus provides better manageability

(Giordanelli and Mastroianni, 2010).

ECONOMICANDTECHNICALASSESSMENTOFCLOUDCOMPUTING

169

H4: Manageability (-) outage duration cost

In case of manageability problems, the system

can not work until the fault is found, and so outage

duration cost increases (The Open Group, 2010).

H5: Data locality (+) operational efficiency

Data locality minimizes the amount of data

movement which increases efficiency (Giordanelli

and Mastroianni, 2010).

H6: Data locality (+) cloud performance

Data locality improves end-application

performance which increases cloud performance

(Giordanelli and Mastroianni, 2010).

H7: Data locality (+) scalability

(Giordanelli and Mastroianni, 2010).

H8: Resilience (fault tolerance) (+) cloud

performance

(SunGard, 2010).

H9: Scalability (+) operational efficiency

(

CCD, 2011).

H10: Scalability (+) agility

Scalability, especially when dynamic, increases

agility (Citrix, 2010).

H11: Distance (-) cloud performance

(Weinman, 2011).

H12: Concurrency (-) security

(Expert Group, 2010; Kovachev et al., 2011).

(Anderson, 2008).

The conceptual network of the hypotheses is

presented in Figure 1:

4 CONCLUSIONS

Based on the literature review, strengths achieved

via cloud operation and adaptation have been

identified as higher resource utilization, efficient

resource management, increased agility, reduction in

capital expenditure and outage duration cost,

increased availability, decrease on data replication

cost, ease of access, information sharing, improved

security. On the other hand, weaknesses such as

authentication, ineffective and problematic network

virtualization, vendor-lock in, increase on switching

cost may also result. Threats are identified as

isolation failure, data protection, loss of governance,

difficulty in providing innovative processes, whereas

opportunities such as virtualization and scaling

structures of cloud and resilience can arise.

Obstacles of cloud computing are scaling

oriented issues, data security related issues, data

lock-in and possible loss of system and applications

availability.

Enablers of cloud computing have been

identified as virtualization, network services, high

speed networking and applications and open source

structure.

At the current stage of this research, 8 axioms

and 12 hypotheses have been formulated

constituting a conceptual network of relationships.

Verification of the hypotheses will be based on

exploratory and validatory case studies. Matured

hypotheses will then be studied with cloud service

providers as well as clients in validatory case

studies.



PRCD

NV

H4(‐)

H8(+)

H11

(

‐

)



MNG

H1a(‐)

LOV

H5(+)

H7(+)

OEF

AGL

H10(+)

SCL

DIS

H9(+)

H6(+)

PCD VLID

ODC

CPRSL

H1b(+)

H2(‐) H12(‐)

H3(+)

CCRCY

Figure 1: Conceptual Network of Hypotheses.

CLOSER2012-2ndInternationalConferenceonCloudComputingandServicesScience

170

REFERENCES

Anderson, R. (2008). Security Engineering: A Guide to

building dependable distributed systems (2nd ed.).

Wiley.

Armbrust, M., Fox, A., Griffith, R., Joseph, A. D, Katz,

R., Konwinski, A., Lee, G., Patterson, D., Rabkin, A.,

Stoica, I., Zaharia, M. (2010). A view of cloud

computing. Communications of the ACM, 53(4), 50-

58.

Beard, H. (2009). Cloud Computing Best Practices for

managing and measuring processes for on-demand

computing (ITIL). London, UK.

Blaisdell, R. (2011). How to determine how much your

cloud costs. Retrieved December 28, 2011, from, http:

//www.rickscloud.com/how-to-determine-how-much-

your-cloud-costs/

Chebrolu, S. B. (2011). Assessing the Relationships

among Cloud Adoption, Strategic Alignment and IT

Effectiveness. Journal of Information Technology

Management, 22(2), 13-29.

Citrix (2010). Is your load balancer cloud ready.

Retrieved December 28, 2011, from, http://images01.

insight.com/media/pdf/0311Whitepaper2NetScalerIsY

ourLoadBalancerCloudReady.pdf

Clark, C., Fraser, K., Hand, S., Hansen, J. G., Jul, E.,

Limpach, C., Pratt, I. and Warfield, A. (2005). Live

migration of virtual machines. Proceedings of the 2nd

ACM/USENIX Symposium on Networked Systems

Design and Implementation (NSDI), 273–286.

Cloud Computing Development (CCD, 2011). The Secret

to Doing Cloud Scalability Right. Retrieved December

28, 2011, from, http://www.cloudcomputingdevelop

ment.net/the-secret-to-doing-cloud-scalability-right/

Computer Network Technology Corporation (CNTC,

2003). Data Replication Considerations. Retrieved

December 28, 2011, from http://www.parallelprose.

com/parallelprose/white_papers/CNT%20Replication..

pdf

Cook, N., Milojicic, D., Talwar, V. (2011). Cloud

Management. Retrieved December 28, 2011, from

http://www.hpl.hp.com/techreports/2011/HPL-2011-

238.pdf

Denning, P. J. (2005). The Locality Principle.

Communications of the ACM, 48(7), 19–24.

Expert Group (2010). The Future of Cloud Computing.

Opportunities for European Cloud Computing Beyond

2010. Retrieved August 21, 2011, from http://cordis.

europa.eu/fp7/ict/ssai/docs/cloud-report-final.pdf

Giordanelli, R., Mastroianni, C. (2010). The Cloud

Computing Paradigm: Characteristics, Opportunities

and Research Issues. Retrieved August 12, 2011, from

http://grid.deis.unical.it/papers/pdf/RT-01-2010-Cloud

Survey.pdf

Gutierrez, J. M., Riaz, T, Pedersen, J. M. and Madsen, O.

B. (2009). Upgradeability and predictability analysis

for mesh topologies in optical distribution networks.

Proceedings of the Sixth international conference on

Wireless and Optical Communications Networks, 369-

373.

Harris, J. G., Alter, A. E. (2010). Six Questions Every

Executive Should Ask about Cloud Computing.

Retrieved August 13, 2011, from http://www.

accenture.com/SiteCollectionDocuments/PDF/Accent

ure_6_Questions_Executives_Should_Ask_about_Clo

ud_Computing.pdf

Helft, M. (2009). Google confirms problems with reaching

its services. Retrieved July 24, 2011, from http://bits.

blogs.nytimes.com/2009/05/14/google-confirms-pro

blems-accessing-its-services/

Henkel, J., Parameswaran, C., Cheung, N. (2007).

“Application Specific Embedded Processors” in

“Designing Embedded Processors”. Springer, 3-23.

Jiang, J., Yang, G. (2010). “Examining Cloud Computing

from the Perspective of Grid and Computer-Supported

Cooperative Work” in “Cloud Computing Principles,

Systems and Applications”. London: Springer, 63-76.

Khalidi, Y. A. (2011). Building a Cloud Computing

Platform for New Possibilities. IEEE Computer

Society, 44(3), 29-34.

Klemperer, P. D. (1987). Markets with Consumer

Switching Costs. Quarterly Journal of Economics,

102, 375-394.

Kovachev, D., Cao, Y., Klamma, R. (2011). Mobile Cloud

Computing: A Comparison of Application Models.

Retrieved August 23, 2011, from http://arxiv.org/ftp/

arxiv/papers/1107/1107.4940.pdf

Kundra, V. (2011). Federal Cloud Computing Strategy.

Retrieved August 18, 2011, from http://www.cio.gov

/documents/federal-cloud-computing-strategy.pdf

Kuntz, P. A, Christie, R. D, Venkata, S. S. (2001). A

Reliability Centered Optimal Visual Inspection Model

for Distribution Feeders. IEEE Transactions on Power

Delivery, 16(4), 718-723.

Leavitt, N. (2009). Is Cloud Computing Really Ready for

Prime Time? IEEE Computing Magazine, 42(1), 15-

20.

Li, A., Yang, X., Kandula, S., Zhang, M. (2011).

Comparing Public-Cloud Providers. IEEE Computer

Society, 15(2), 50-53.

Myerson, J. M. (2011). Craft a cloud performance metrics

policy. Retrieved December 28, 2011, from, http://

www.ibm.com/developerworks/cloud/library/cl-cloud

performmetrics/

Neville, J. (2010). INTUG Cloud Computing Position

Paper. Retrieved December 28, 2011, from, http://

intug.org/wp-content/uploads/2009/03/Position-Paper-

Cloud-Computing.pdf

Newvem (2011). Availability: Story of the Inevitable

Outage of the Cloud. Retrieved December 28, 2011,

from, http://blog.newvem.com/main/2011/08/availa

bility-story-of-the-inevitable-outage-of-the-cloud.html

Padala, P. (2010). Understanding Live Migration of

Virtual Machines. Retrieved December 1, 2011, from

http://ppadala.net/blog/2010/06/understanding-live-

migration-of-virtual-machines/

SearchCloudComputing (2011). Cloud Computing from A

to Z. Retrieved December 28, 2011, from, http://search

cloudcomputing.techtarget.com/feature/Cloud-compu

ting-from-A-to-Z

ECONOMICANDTECHNICALASSESSMENTOFCLOUDCOMPUTING

171

SunGard Availability Services UK Ltd. (2010). Resilience

is the Key to Success. Retrieved August 24, 2011,

from http://www.sungard.co.uk/Documents/Cloud-

Computing-White-Paper-SunGard.pdf

The Open Group. (2010). Building Return on Investment

from Cloud Computing: Cloud Computing Key

Performance Indicators and Metrics. Retrieved March

19, 2011, from http://www.opengroup.org/cloud/white

papers/ccroi/kpis.htm

Travostino, F., Daspit, P., Gommans, L., Jog, C., Laat, C.,

Mambretti, J., Monga, I., van Oudenaarde, B.,

Raghunath, S., and Wang, P. Y. (2006). Seamless live

migration of virtual machines over the man/wan.

Future Generation Computer Systems, 22(8), 901–

907.

Uhlig, R., Neiger, G., Rodgers, D., Santoni, A. L.,

Martins, F. C. M., Anderson, A. V., Bennett, S. M.,

Kagi, A., Leung, F. H., Smith, L. (2005). Intel

Virtualization Technology. IEEE Computer Society,

38(5), 48-56.

Varia, J. (2010). Amazon Web Services- Architecting for

the Cloud: Best Practices. Retrieved August 15, 2011,

from http://media.amazonwebservices.com/AWS_

Cloud_Best_Practices.pdf

Wang, C. (2009). Cloud Security Front and Center.

Retrieved March 1, 2011, from http://blogs.forrester.

com/security_and_risk/2009/11/cloud-security-front-

and-center.html

WebLogic (2011). WebLogic Server Performance Tuning

Guide. Retrieved December 28, 2011, from, http://

docs.oracle.com/cd/E13222_01/wls/docs51/admindocs

/tuning.html

Webopedia (2011). Operational Efficiency. Retrieved

December 28, 2011, from http://www.webopedia.

com/TERM/O/operational_efficiency.html

Weinman, J. (2011). Axiomatic Cloud Theory. Retrieved

December 28, 2011, from, http://www.joeweinman.

com/Resources/Joe_Weinman_Axiomatic_Cloud_The

ory.pdf

Wilson, S. (2008). AppEngine outage. Retrieved March

12, 2011, from http://www.cio-weblog.com/50226711/

appengine_outage.php

CLOSER2012-2ndInternationalConferenceonCloudComputingandServicesScience

172