It's well documented that governments use financial incentives and disincentives (usually through tax policies) to drive public policy. Can this work within large SOA environments? IBM says yes -- and has its own experience with this form of governance to prove it.
Often, SOA programs may become victims of their own success. Business units could go crazy developing services that duplicate or conflict with one another. This was one of the major challenges IBM faced in the early days of its own SOA effort.
According to a new article by SearchSOA's Rich Seeley, IBM first launched its own SOA effort back in 2002, a time when most companies were just starting to figure out Web services. Things started rolling slowly for the first three years as IBM's SOA team experimented with with exploratory services and proofs of concept, said John Falka, chief architect for SOA governance in IBM's internal software group.
By 2005, however, things were hopping. The number of projects grew. However, Falka and his team saw that redundancy of Web service development was becoming an issue. Divisions within IBM were developing their own Web services that performed the same functions as Web services developed by other divisions.
To address this, the team began using something they called service-oriented modeling and architecture (SOMA) to identify Web services that could be reused in new applications. Thus began IBM's first comprehensive SOA governance effort, which employed financial tools to incent or disincent such development. "We provided funding for the single build-out of a service as opposed to the multiple build outs of a service and raised the operational cost benefits of not having redundant services that were generating additional costs," Falka explained.
Leave a comment