Enterprise SOA

Ian Bailey

This article is about services in general. They may be implemented by systems or by people, or both.

Service-Oriented Architecture offers great potential for streamlining businesses. The idea goes that if you can wrap-up some kind of functionality as a well-defined service, others will pick it up and use it as and when they need it. If you get the granularity of the services right – not too small that you need hundreds of them to do the simplest things, and not so large that they only serve one purpose – it should be possible to chain them all together and so support complex business processes. This is called service orchestration, and there’s lots of material out there on the subject (usually from middleware vendors trying to flog their service buses).

Re-use of services has the potential to streamline an enterprise and allow it to be flexible in the face of changing markets and threats. Cool stuff, sounds like management-speak, but even I can see it’s sensible thing to want to do. The reality is a bit more complex though – especially for larger organisations. These tend to have repeated functionality across multiple departments and divisions. The received wisdom is to analyse the business and see which providers offer the best solution, wrap them up as a service and tell everyone to stop using their own capability and to use the new one. How can you be sure you’ve picked the best provider ? What are your metrics for selection ? I don’t think you can be sure of anything – especially if the goal is to change the business – the providers currently perceived as the best in class might go to pieces in a changed (SOA) environment.

‹ prev | 1 2 3 | next ›

del.icio.us