Had the opportunity to participate in a Q&A session with other Microsoft customers regarding our approach to Service Oriented Architecture. Overall, the session was geared to be very high-level (CIO, CTO) and it definitely was...almost nose-bleed level and I certainly didn't do much of anything to bring it closer to Earth!
If we are one of the furthest along in SOA adoption, then it might say something about how well SOA is understood, or how much value it provides. There were many slides by both SOA Institute and Microsoft that brought back sharp, painful memories of the COM era....implementation indifference, location transparency, loose coupling...it was like a 10 year class reunion on buzzwords!
All that aside, there are many elements of SOA (or COM, or -[insert here]-) that are worth considering; we've still got a good way to in maturing how we approach our goals. For instance, writing well defined services is not enough. If these services (and related assets ) aren't discoverable, via some common registry, then no one will use it and we still have a continuous drain on the organizations ROI.
Once these services are discoverable there needs to be governance on the service life cycle. Who is using it? How are they using it? Should there be a usury fee for the service? Is the service meeting its stated SLA for a customer? What do we do if its not? These are all types of scenarios that we need to be evaluating our need for.
We are off to a great start, but we need to start thinking about more of a long term investment vs. the quick fix that our initial approach gained us.