I recently pulled a white paper off of the CSC website that highlighted a discussion with Michael Capellas. The topic was Cloud Computing and the event was a Masterclass, “Masterminding The Future of Cloud Computing” held in October, 2010.
I found it interesting that the three events cited by Capellas that have enabled Cloud Computing were “leased lines to IP”, “Intel’s X-86 processing architecture”, and the “explosion in wireless access.” He does a nice job of explaining their contribution. These really relate to a set of standard services and are technology enablers that allow for more dynamic capacity and access.
While necessary, these technology enablers are not sufficient to provide the benefits of the cloud as presented by Capellas. The barriers to adoption are more than just legacy applications and the amount of “local” information stored on the desktop.
Private and public “clouds” have both similar and different hurdles to overcome prior to widespread adoption. Three examples are provided below.
Private clouds, conceptually, will never be as cost effective as public clouds due to the level of scaling, sharing of costs among multiple customers, etc. Realizing the full benefit of a private cloud requires standardization as in a public cloud but also an enterprise view of demand cycles and capacity requirements. Understanding enterprise demand will minimize the additional capacity reserved for peak loads.
Whether you are talking about private or public clouds, other hurdles include the need for consumption-based pricing of computing resources. This is not a trivial matter and no standards currently exist. Depending on the service provider, computing resources are defined and priced in different “units”. Comparability of these units is difficult at best since they are comprised of different components and range in structure from simple (CPU Minute, for example) to complex (consisting of combinations of factors such as memory, processor, and storage).
A third hurdle involves the skills (technical as well as management) required to realize the benefits of cloud-based computing. How many IT organizations are really capable of implementing the technology and processes required to support a scalable, demand-driven computing environment? Translating business requirements into a set of standard services, identifying the cost of business policies that inhibit the establishment of standard services, and understanding enterprise demand are all challenges that, if not addressed effectively, will severely limit the benefits of cloud computing.
The fact is, “Tier 1” IT service providers are struggling with these challenges and more. While pockets of services have been demonstrated (most notably, SAP services, for example), the ability to provide widespread benefit to the enterprise in the form of delivering general IT services has yet to be realized.
In summary, standard services, utility computing, cloud computing, etc. make sense, conceptually. Focusing on the technical challenges is interesting and relatively straightforward. Embarking on this journey, be prepared to spend most of your energies on the management and skills required to realize the benefits.
