2 thoughts on “SOA in action inside IBM, Part 1: SOA case studies”

  1. The article simply shows that it’s good to re-write old software from time to time. It does not show any actual value of SOA. A simple multithreading and new hardware would probably give the same results at 20% of costs. Both examples in the article sound close to “The customer had an old system written in QBasic and running on Pentium 286, then we came and installed 10 mainframe IBM computers and a lot of new software from IBM, and now the application runs 50 times faster. To make it really cool we also added some Web Services.”

    Best practices used and lessons learned: With enogh thrust even pigs will fly.


  2. For most organizations it just isn’t an option to simply “re-write old software from time to time”. It’s too expensive. Grabbing the problem by the root and rewriting the applications as loosely coupled services is a nice approach given, as you mention, that you have the hardware and infrastructure to host and manage the services. Also – simply rewriting existing application as standalone would mean that in a couple of years you would be back at square one with an application that is unmanagable and needs a new rewrite.

    I’m not necessarily advocating SOA for new applications but I’m saying that it is looking more and more like a viable solution in the long run.


Comments are closed.