This Blog
Syndication
Recent Posts
Tags
Calendar
Archives
Iterating on IT Service
February 2008 - Posts
-
CMDBf Specification: Implementing now makes sense
In December, I blogged happily that "DMTF Accepts New CMDBf Specification." Since then, the CMDB Federation Work Group of the Distributed Management Task Force (DMTF) has met four times and the CMDBf consortium, which wrote the spec before turning it over to the DMTF, is winding down. We are still working out the preliminaries for the work group--electing co-chairs, getting our charter into final form--administrivia that will go on for a few more meetings.
There are a lot of new faces around the table and some of the same folks who made up the original consortium team. More than twenty companies have signed up to participate in the working group--proof that CMDB federation has the industry's attention.
I had an opportunity to talk with CA customers about CMDB federation a couple weeks ago. There is some real excitement around the fact that CA is implementing the CMDBf web services in our 11.2 release of CA CMDB. And some skepticism about the possibility of vendor-neutral CMDB federation. I heard more than one person say that they will believe it when they see it. I don't think they will have to wait long.
I've been asked several times, "Why risk implementing the consortium specification? Why not wait until the DMTF adds its stamp of approval?" The answer involves several considerations. CA, like all the other consortium members, has already put a lot of work into the specification. We have built test mock-ups and prototypes just as if the CMDBf services were our own services that we were developing internally. Having put that work into the specification, we have a similar level of confidence in Version 1 as we would have in an internally developed Version 1. Given that level of confidence, why not put the spec into a product and offer it to beta sites? The specification is free and public for anyone to implement, so why not us?
The value of the specification rises with every implementation. The more vendors we can entice to join us in implementing the spec, the more value the specification accrues. We hope that every member of the consortium is thinking the same way, and we encourage non-members to look at the spec carefully and try an implementation. The open source implementation by the Eclipse COSMOS group should be a big help.
We are taking a risk by implementing early, but not as great a risk as you might guess. Flaws, if there are any, are more likely to be uncovered in the field, than around the work group table. Putting the spec to use is the quickest way to ferret out any subtleties we may have missed. That is the point of the beta test process and why every development manager loves his beta sites.
Yes, there is always a possibility that the DMTF will find a flaw in the specification and change it substantially. Then we would have to redesign and rebuild and find a way to migrate sites that have already implemented the product with the new spec. That is a risk we take with internally developed services as well as standard services.
With little to lose and much to gain, implementing now makes sense. I hope the IT community-at-large agrees.
Print
Email
Rate