PAY-AS-YOU-GO COMPUTING. WHAT A concept! Such ideas that turn established norms on its head are usually the ones that define the next ‘leap’ forward. But just as usually, they could fall victim to their own hype and unrealistic expectations. With all the wide-eyed cheerleading going on about the value of utility computing, a reality-check certainly doesn’t hurt.

If you haven’t heard of the utility computing model as yet, let me be the first to invite you back to earth, and direct you to two excellent articles on this subject in the October 2003 issue of Silicon India. The one titled “Grid Computing Comes of Age” presents the way an Oracle executive envisions how this emerging concept will add value to an enterprise. The other one, titled “On-Demand is IN Demand” is a vendor-specific take on utility computing. Both of them succeed in looking into the future and highlighting what’s in store for the world of enterprise computing.

Judging by the amount of press that Utility Computing is getting, it has definitely ‘arrived.’ But there are still a lot of unanswered questions and unproven claims that need to be addressed. To many skeptics, it may seem like a solution in search of a problem. But in reality, the concept of on-demand computing is a combination of technology developments and evolving business practices. The Utility Model essentially solves business problems of allocation, availability, and optimization of IT support.

The Utility Model Explained
The Utility Model is quite simply utility pricing bundled with utility computing. In other words the capability of end users to consume only as much computing resource as they want is utility computing. This is only the technical part of the equation. The business side of this equation is the capacity of utility service providers to bill the end users based on service utilization. This constitutes utility pricing.

Many vendors who offer these services are born-again Application Service Providers (ASP) and Storage Service Providers (SSP). Perhaps this explains why the utility service model is very similar to the ASP model. The resemblance extends even further as these companies begin to dub themselves as Utility Computing Service Providers (UCSP) and Service Infrastructure Solution Providers (SISP).

What is compelling about the utility model is that IT functions have increasingly become more about services and less about infrastructure and hard assets. To that end, any business that needs access to these services may be better off tapping what is available out there than setting out to build and maintain their own service delivery capabilities.

Some Searching Questions
Analysts haven’t lost much time in heralding the advent of Utility Computing (see box story “Expert Opinion”). But before companies rush into adopting this model, what is important is for them to ask themselves a few tough questions. Some of the questions are:

• Does this present a significant opportunity to cut costs without sacrificing quality-of-service? A watertight service level agreement would go a long way in ensuring that this is indeed the case.
• Does this truly help the company to focus on its core-competency? If not, then you have to baby-sit the service-provider. If that turns out to be the case, you might as well in-source the service and keep the expertise at home.
• By adopting the utility model, do I surrender control of my IT architecture and application development strategy? If you do not retain control, utility computing could easily become a case of the tail wagging the dog.
• Does this help minimize significant capital investments in the future? Anticipation of future uptick in demand is mainly the reason why companies sink huge sums of money on excess infrastructure capacity. To the extent that this can be avoided, it helps to outsource from a utility service provider.
• How does this affect the security of the enterprise? If there is even a perceived threat to system security as a result of moving to the utility model, the experiment may not be worth the risk.
• Will the drive towards standardization (in an effort to fit the utility model) kill off innovation? If your foray into utility computing forces you to focus more on cost recovery than on value creation and innovation, the price you pay may be in the form of lost edge over competition.

Zoning In On the Details
The devil, as we all know, is in the details. The vendors have worked hard to spell-out the overarching concept and the gains to be had by adopting their brand of utility computing. The financial Goliaths have dipped their toes in to get a feel for it (See box story “Early Adopters”). It is now up to the IT industry to chart a clear course that focuses on implementation details and best practices.

Companies that have aggressively pursued server consolidation and standardization initiatives are but one step away from embracing the utility model. The key now is to produce hard evidence that the utility model consistently delivers on its promise. Any other outcome will only serve to underscore a cynic’s point of view – that all this hype about utility is just an exercise in futility.

V. Anantakrishnan is a senior technology manager at AXA Financial in New York. He has over a decade of experience managing and implementing large scale systems integrations projects in the telecommunication and financial industries. He holds a masters in international business from Columbia University, and a masters in IT management from Carnegie Mellon University. He can be reached at vak88@hotmail.com