INTERPRETING THE RESULTS: THE COMPUTING UTILITY MODEL

15 January 2002

by Rob Austin

One message that emerges clearly from the results [of a recent Cutter survey that asked respondents about their awareness and enthusiasm for the computing utility model] is that the computing utility industry is still in its early days. People recognize the theoretical economic benefits of the model, but they see it as a distant, rather than imminent, trend. The IT professionals who responded to the survey are people who work in the day-to-day details of the field and are attuned to issues of how such a model might operate. And they don't see it happening just yet.

One could argue that the core group of respondents to this survey, IT professionals, might be predisposed to skepticism about the computing utility model, which is a version of outsourcing. The economic benefits that might accrue from implementing the model are largely cost reductions, the type of benefits that CFOs like but that leave many technologists worried. IT professionals, who will have to manage the day-to-day challenges of this type of outsourcing, might suppose that their jobs will get harder before they get easier in transition to such a model. This supposition is probably entirely correct. The transition to computing utility models in many companies is likely to be somewhat traumatic for those on the front lines.

The psychological barrier that makes people hesitate to trust storage of sensitive data to outsiders is clearly present in the survey responses and could be interpreted as a general form of related concerns about security and instability in the service provider market. The bottom-line worry of IT professionals here is that vendors might not do as good a job in taking care of vital company data. Although there may be merit in this concern, it is equally clear that companies routinely trust their fates to outside firms in many other areas. Manufacturers have implemented just-in-time systems even though it makes them vulnerable to shutdown by supplier mismanagement, for example. Businesses have flocked to voice mail in large numbers, even though the potentially sensitive content of voice mail messages is stored by a service provider. History has shown that operational vulnerabilities can and will be managed if there are cost savings to be gained.

I have often heard IT professionals say things like "Outsourcing is the last thing you want to do if you care about computer security," or "Software components are the last thing you want to use if you care about software quality." These statements share a pair of compelling characteristics:

1. They are probably true in the short run.
   
   
2. They will probably not be true in the long run.
   
   

IT professionals can tell you, based on their own experience, how bad some service providers are at computer security and how poorly some popular software components run. But the forces driving outsourcing and component use are economic forces. Trying to oppose these forces is an exercise in futility.

It is a deep economic truth that throughout history, people who focus on an activity and develop a capability for doing it best end up doing that activity for everyone else. Most of us no longer grow our own food or make our own clothing. Most businesses no longer generate their own power. The IT world is not immune to this economic truth. Although service providers may do a poor job of security today, eventually most will achieve economies of scale across multiple customers that will allow them to justify greater investments in computer security than any customer could afford. Similarly, software components may be buggy and annoying now, but eventually refinement of those components will make them a far better choice than trying to write custom software to provide equivalent functionality. It will always be important for companies to retain control over core activities from which they gain competitive advantage. But few companies today gain competitive advantage from their office automation software. Eventually that will be provided in the lowest cost, most reliable manner. Hence, the computing utility model will be driven by this economic truth.

In the early days of this century, power professionals inside individual companies likely swore they would never give up their own power plants. The respondents to this survey raise excellent points and objections to the computing utility model. But there is an element of clinging to their own power plants in the responses.

Predicting timing is always difficult. It is difficult to say when it will happen. But time will, eventually, tell.

-- Rob Austin, Fellow, Cutter Business Technology Council

[For more on the Cutter survey on the computing utility model and Rob Austin's comments, see the October 2001 issue of Cutter Benchmark Review, available from Cutter Information Corp. at +1 800 492 1650 or +1 781 641 9876, fax +1 800 888 1816 or +1 781 648 1950, or e-mail service@cutter.com.]