Response to KPMG: Debunking the myth of SOA

Thursday, August 13, 2009 |

First off, let’s recognize that the KPMG article is an opinion piece that would have been much more helpful in 2003. There is at least one truth in the KPMG article - there is a need to cut through the SOA hype. However, it is a little ironic, that a large financial and risk consulting firm would make a perfect blanket statement such as “I’m yet to see many businesses extracting any real value from an SOA installation.” Perhaps the author, Jerry Iacouzzi, should have read CIO Magazine’s Winners of SOA Case Study Competition, a list of SOA-based projects and the real value they brought to organizations that completed them. Even if one hasn't had first-person experience with real value of SOA, that doesn't mean that SOA is all hype.

The article ends
by taking the debate up a level conceptually, What should actually be happening is that they [Business/IT] should be thinking more strategically about their IT architecture requirements or about the governance and control issues which arise from an SOA installation.” So if companies do that will their SOA investments suddenly produce value? Are there no other reasons to think about technology governance and control issues? What exactly are "IT architecture requirements"?

As silly as some of the article’s statements may be, what I really want to zero in on is the overall complaint
the author brings up. I think there is some validity there. But I have another interpretation about the source of the hype. It is not just that vendors and standards makers and large consulting firms have been pounding the SOA podium to get business. That’s natural and we should expect them to send that message. But, when I listen to the bevy of chants out there I often come away with the “Most services [SOA] are common across business units” or “SOA will break down business silos” messages. These messages suggest that SOA is a unifier of business operations, a re-arranger of business units into a monolithic company that is super-efficient. While such messages are true for some organizations it is not true for most. In fact, these messages are not sensitive to business operating models at all. They reek of IT adventurism.

There is a simple proof I can offer thanks to the work at MIT Sloan’s Jeanne W. Ross, Peter Weill and David Robertson. The work at MIT Sloan has identified four business operating models. Only one of them, called the unification model, behaves like the super-efficient uber monolithic organization that some in the SOA community seem to want to address. That same unification model lends itself to an enterprise view of infrastructure, service, and process simultaneously. But not all companies need such a grand unified vision to get value from SOA. In fact, most organizations don't need this grand unified vision to get value from SOA. Perhaps the lack of alignment between SOA hype and business operating model is what author, Jerry Iacouzzi, is reacting to in the first place, though that nuance is not discernible.

Dissecting the issues of SOA applicability run the gamut from departmental to business unit to enterprise. Which one are you? If your business units service different key customers and those customers do business with your diversified company for relatively unrelated product offerings, is there any need to optimize at the enterprise level beyond infrastructure? Optimizations for such a diversified company would yield most value at the business unit or even the departmental level. And such a ‘strategy’ would still produce benefits for the company, wouldn’t it?

So I think it’s important to realize that if we become frustrated with the ‘SOA hype’ that we should also take a deep look at our approach to SOA. Our frustration might just be another part of the ‘hype’ cycle.

0 comments: