The Adapter framework optimization, or, in fact, the minimization of all protocols/endpoint types aiming the realization of the Official Endpoint SOA pattern presented previously, depends on many factors (we have indicated them in the Conclusion section), but decisive is only one. As we already mentioned, this factor is not technical, unfortunately. At the end, the main question is the level of ownership we as architects have on refactored application endpoints. When the approach is invasive (we have to inject unified triggers and establish event handling and ESR schemas), we totally depend on the level of cooperation with the application owner/vendor. If this approach is not approved, then noninvasive methods shall be considered and some of them, as we explained earlier, could be based on disaster recovery data replication and here we could face even bigger resistance from DBAs/Operations.
Who can blame them? The approaches we discussed in the first part...