Todd Biske: Outside the Box

As reported by many bloggers/analysts/columnists, including this article by Tony Baer, BEA announced Project Genesis. Tony, in his article, stated that “there is considerable speculation that it will build on the [newly announced AquaLogic Enterprise] registry/repository.” I have a couple comments on the announcement that I wanted to discuss here.

First, we have yet another API in the metadata registry/repository space. Back in February, IBM created a whole bunch of brouhaha when they stated that they felt a new standard was needed for registry/repository communication, concluding that UDDI wouldn’t cut it. Now, we have the Metadata Interoperability Framework. Add this to UDDI, the WSRR protocol, HP/Mercury/Systinet Governance Interoperability Framework, and I’m sure this won’t be the end. You could certainly throw in LDAP for good measure, as well. Interestingly, no one has criticized BEA that I’ve seen for this move, but then again, they didn’t publicly come out and state that UDDI won’t cut it, either.

Second, given all of these APIs for metadata communication, it certainly does raise the question of whether metadata is really at the center of the SOA technology universe? Clearly, if all of the major platform vendors are now having to come up with custom protocols for communication with their registry/repository, it certainly means that there’s a lot going on with it. If you think about it, a metadata repository is a bridge between the design/development time world and the run-time world. Likewise, it also plays a key role in the run-time world and the world of operational analytics. Most people would also agree that most SOA efforts strive to separate non-functional concerns from functional concerns and allow for a policy-based approach to allow changes to be configured rather than coded.

We’ve certainly come a long way from the early days of the registry/repository where the focus was solely on design time discovery of services. Organizations that are solely focused on design time discovery of services still wrestle with whether they even need a registry/repository prior to reaching some critical mass of services. Vendors like BEA, IBM, and WebMethods are all certainly now pushing the Registry/Repository for much more than just this. I’ll admit, I have some concerns that we’re getting too far out ahead of the need of many enterprises. I’m a big proponent of a policy-driven infrastructure for SOA, but I’ll also admit that if you don’t have a sound understanding of your services and their consumers first, it’s not going to reap the benefits that are possible. Only time will tell whether we’re over-engineering based on what-if scenarios or if this level of sophistication driven by metadata is applicable to the masses.

4 Responses to “Is Metadata the center of the SOA technology universe?”