Harsha dis 3 635

1520-9202/09/$26.00 © 2009 IEEE P u b l i s h e d b y t h e I E E E C o m p u t e r S o c i e t y computer.org/ITPro 57

Smart It

Cloud Computing: Will Commodity Services Benefit Users Long Term?

T he attention the IT com- munity has given cloud computing recently ri- vals that given by Web

2.0-loving people to Twitter. This magazine alone dedicated its March/April 2009 issue and sig- nificant other space throughout the year to the topic. In their intro- duction to the special issue, guest editors Jeffrey Voas and Jia Zhang asked, “Do you think [cloud com- puting] offers anything different? Is it just the same old IT packaged up in a new bottle? Or is it really new wine?”1 After a review of the basics as well as the challenges, we respond to that question.

The Basics In the March/April special issue, Geng Lin and his colleagues pre- sented cloud computing (or IT as a service [ITaaS]) as a three-layer service architecture:2

• top-layer applications delivered on demand in the software-as- a-service (SaaS) model;

• middleware providing applica- tion services and/or a platform- as-a-service (PaaS) runtime environment for cloud applica- tions; and

• a flexible infrastructure (hence,

infrastructure as a service [IaaS]) of distributed data center services connected via Internet- style networks.

The fundamental category of cloud services is IaaS—the build- ing blocks of an IT enterprise. Those who buy and use IaaS are predominantly the IT/system ad- ministrator types charged with obtaining general processing, storage, database management, and other core IT resources and applications. Services provided in this category include CPU pro- cessing on demand, virtual Web hosting, and storage on demand. The most notable vendors are Amazon’s EC2, GoGrid’s Cloud Servers, and Joyent.

PaaS is an emerging category of cloud services where develop- ers can design, build, and test ap- plications that run on the cloud provider’s infrastructure and then deliver those applications to end users from the provider’s servers. The early market leaders in this area include Google’s App En- gine, Microsoft’s Azure, Amazon Web services, and Force.com (by Salesforce.com).

SaaS is a cloud service con- sumed directly by the end user.

Instead of storing software on the user’s computer, providers deliver SaaS as a fully functional applica- tion over the Internet. Yahoo mail, Google Apps, Salesforce.com, WebEx, and Microsoft Office Live are all cloud service “products.”

This three-layer service archi- tecture is implemented on both a public basis (that is, via public clouds whose resources the enter- prise doesn’t own) and an enter- prise basis (that is, private clouds whose resources the enterprise centrally manages). In either im- plementation, the services provid- ed through the cloud are shared and remote with respect to end users.

The Challenges Cloud computing faces many of the same challenges as other information and network tech- nologies: performance, security, resiliency, interoperability, data migration, and transition from legacy systems.

Providers, users, and the federal government agree almost uni- versally that cybersecurity is the predominant risk. Given the seri- ous threats from varying sources (individuals, criminal groups, nation states) that can only be

H. Gilbert Miller and John Veiga Noblis©

a rg

u s4

5 6 | D

re am

st im

e. co

m

Authorized licensed use limited to: Cleveland State University. Downloaded on July 19,2022 at 17:52:11 UTC from IEEE Xplore. Restrictions apply.

58 IT Pro November/December 2009

Smart It

mitigated by a unified and coop- erative response from cloud ser- vice providers, network providers, developers, enterprise IT organi- zations, and users, such a ranking is appropriate.

Measuring the performance of cloud-provided services remains a challenge primarily due to the lack of measurement standards. Performance measures for cloud services lack the maturity of those used in telecommunications net- works, for example, where mea- sures such as availability, latency, and phase jitter are well-defined and understood by both custom- ers and network providers. In the world of cloud computing, perfor- mance definitions and standards vary wildly by provider—even those offering similar services.

Finally, as in any previous IT paradigm shift, users and provid- ers alike face the additional chal- lenge of migrating functions and data from legacy systems to the

cloud, as well as the interopera- bility of cloud services with legacy systems.

The Promise of New Wine But regardless of the specific challenge, cloud service providers might be better positioned to ex- ploit economies of scale to provide a level of security and service that might not be cost-effective for small- to mid-level enterprises.

Beyond its obvious evolution of technology, cloud computing rep- resents an evolution of how IT is purchased and provided. Cloud computing is an alternative IT sourcing strategy where shared, remote IT resources deliver scal- able computing capabilities over IP networks as an on-demand, metered, virtualized service. The result is a new level of commod- itization in the IT market—some- thing beyond just individual hardware and software as com- modities, but integrated platforms

and infrastructures delivered as services and commodities. And as cloud services become com- modities, users begin to benefit from an upwardly moving spiral in which

• technology advancements lower infrastructure costs and pro- mote operational efficiencies;

• improved efficiencies lower costs, which improves econo- mies of scale for providers and spurs customer demand; and

• strong demand attracts compe- tition, which lowers prices and spurs investment in new tech- nologies.

The voice telecommunica- tions market realized this com- modity cycle from the late 1980s through the early 2000s.3 Over that 20-year period, the advent of all digital networks, the increas- ing capacity of digital switches and transmission facilities, and the convergence of voice and data services in an Internet protocol infrastructure enabled enterpris- es to increasingly use shared pub- lic networks. As Figure 1 shows, prices continued to decrease over the 20-year period due to this cycle. During this time, telecom- munications offerings became broader in number and sophis- tication, as well as increasingly ubiquitous.

This cycle helps put the chal- lenges we described earlier into perspective: cybersecurity remains a significant challenge, which left unsolved will prevent the demand volume needed for the commod- ity cycle to engage. IT managers will rightfully continue to exercise considerable caution when mak- ing changes that might affect their security profile. Again, solving the cloud computing security chal- lenge will require the coordinated and cooperative participation and execution by the entire provider,

Average best commercial prices

Federal government switched voice prices

1989 1992 1995 1998 2001 2004 2007

30¢

25¢

20¢

15¢

10¢

5¢

0¢

Figure 1. Previous case study. Voice telecommunications prices continued to decrease over a 20-year period due to a commodity cycle of technology advancements, efficiencies, and competition.

Authorized licensed use limited to: Cleveland State University. Downloaded on July 19,2022 at 17:52:11 UTC from IEEE Xplore. Restrictions apply.

computer.org/ITPro 59

developer, and user community. However, we’ll need to add an

additional challenge to engage the commodity cycle. The cloud com- puting community—both provid- ers and users—must develop a cost per delivered service measure that allows users to compare commod- ity services. In the voice telecom- munications case study,3 cents per minute of switch voice service was the rallying metric. Understood by users and carriers as a simple cost measure, the underlying per- formance characteristics of a min- ute of switch voice service evolved to ensure an apples-to-apples comparison between commodity services as provided by different providers. Such a measure will not only enable comparisons between cloud computing providers, but will also enable it between public and private clouds.

C loud computing’s potential lies not only in the evolu- tion of technology but also

in a sourcing strategy that uses resources shared by increasingly larger volumes of users. Further potential lies in the commoditi- zation of cloud computing ser- vices—that is, a market in which price primarily differentiates cloud computing services. Al- though the entire IT community will need to address significant challenges, the result will be de- creased prices, increased num- bers of services, and improved performance not only in the near term but also the longer term. The result should be a decade or more in which cloud computing providers and users all benefit from future technology advance- ments, cloud efficiencies, and market competition.

In response to the question Voas and Zhang posed—is cloud computing just the same old IT packaged in a new bottle, or is

it really new wine?—we respond that not only is it new wine, but one of a vintage that will improve with maturity.

References 1. J. Voas and J. Zhang, “Cloud Com-

puting: New Wine or Just a New Bottle?” IT Professional, vol. 11, no. 2, 2009, pp. 15–17.

2. G. Lin et al., “Cloud Computing: IT as a Service,” IT Professional, vol. 11, no. 2, 2009, pp. 10–13.

3. H.G. Miller, H.D. Levine, and S.N. Bates, “Welcome to Conver- gence: Surviving the Next Platform Change,” IT Professional, vol. 7, no. 3, 2005, pp. 18–25.

H. Gilbert Miller is a member of IT Professional’s advisory board and cor-

porate vice president and chief technology officer at Noblis. He has a PhD in en- gineering and public policy from Carn- egie Mellon University. Contact him at [email protected].

John Veiga is a principal engineer at Noblis, where he supports the federal government in its procurement of ad- vanced telecommunications and cloud computing services. Veiga has a BS in industrial engineering from Bradley University and an MBA from George Washington University. Contact him at [email protected].

Selected CS articles and

columns are available for free at

http://ComputingNow.computer.org

From the analytical engine to the supercomputer, from Pascal to von Neumann—the IEEE Annals of the History of Computing covers the breadth of computer history. The quarterly publication is an active center for the collection and dissemination of information on historical projects and organizations, oral history activities, and international conferences.

www.computer.org/annals

Authorized licensed use limited to: Cleveland State University. Downloaded on July 19,2022 at 17:52:11 UTC from IEEE Xplore. Restrictions apply.