Saturday, October 29, 2005
Enterprise perspectives on SOA
SOA isn't about ESB's or some other marketing ploy used to convince executives that they are behind the times if they don't listen to the message. Curious to find why no one ever talks about the notion of business architecture and how SOA should be aligned to it? Many vendors create technical reference implementations but none of them have even made the attempt at creating a business reference implementation.
IBM has the potential to show leadership in this regard as they have core knowledge about many industry verticals from their learnings on San Francisco. Within the insurance vertical, they have the IIA framework that is a great starting point for creating business oriented reference implementations.
Many insulting firms have been savage in incorporating the word SOA into their marketing material without developing a true understanding of the problem at hand. Enterprises are game to get a "strategy" without figuring out reasons others have failed. Simply, SOA done without organizational change is just plain dumb. The success or failure of SOA is directly correlated to the ratio of intelligence to stupidly. At least one blogger has acknoweledged that blame for SOA failures should be placed on stupid people.
SOA attempts to solve for interoperability but only focuses on technical interoperability. How about management insulting firms as part of "strategy" making recommendation on how to fix the organization chart within large enterprises? I can tell you that one guy who gets it is Sam Lowe, Chief Enterprise Architect of Cap Gemini who has a great entry on the real issues making or breaking enterprise SOAs.
I suspect that CapGemini is capable of creating real strategies around SOAs in that they not only address the organizational issues with their clients but also within their own walls. I suspect they don't have a lot of kindergartners doing "strategy" and backing up the bus. Their strategy seems to eschew having "analysts" who may at best have five years of experience in IT and instead prefer folks who have done real work and gained real experience provide consultative advice.
Speaking of "analysts", maybe the second fatal mistake realized by enterprises is asking the vast majority of them for guidance. In my humble opinion, the folks over at ZapThink and Burton Group seem to get it, but the rest should be simply ignored by corporate America. Of course, Burton Group is more politically correct than myself, but you should check out the three major impediments to SOA...
Another problem with SOA where the dots haven't yet been connected is the notion of identity. Craig Burton is starting a good conversation on the problem here. Collectively, we all need to connect the dots...