Service reuse a myth?

If you can't reuse services, then what's the point of SOA? 

None?  Its value is certainly diminished.  Side note:  I like this article that explains SOA simply, and provides a nice neat list of some of the benefits at the bottom.

I have my fears, and I've heard a lot about this in the past few days.  Are people just getting scared?  Is there some real issue with services that will make the paradigm go the way of OO -- reuse in theory, but not in practice? 

Are the complexities of managing such an interconnected architecture with reuse just unreasonable?  Will each client really want it's own slightly special service that suits its needs and levels of performance?  Do people simply not feel comfortable giving up vertical control of business components?  Does it require an unreasonable amount of governance that most architecture groups simply don't have the ability to apply?  Is reuse only more achievable in particular circumstances?  Do we need to focus more on the A of SOA?

Are we simply entering the trough of disillusionment? 

Unfortunately, I have way more questions than answers.

Comments

[Mark Griffin]

Stephen,
These are all very good questions. I think we may be entering into the reality stage. The stage in which things have to actually get implemented as opposed to just talked about. That's always the fun stage.

Where's the fun? It's finding out that creating and managing loosely coupled, reusable, abstracted, autonomous, discoverable, stateless, composable services is not nearly as easy as a lot folks would lead you to believe. As with everything, some companies are going to do okay with it while others are going to have disasters. My thought is that IT shops that were fairly disciplined will have a much better chance than the cowboys or cowgirls.

We are finding that our EAI experience is very helpful. There is a significant amount of overlap. However we too are in the early stages so the verdict is still out.