The CMDB Imperative

Glenn O'Donnel and Carlos Casanova have written perhaps the first worthy analysis of the success and future of the CMDB, with a title befitting a good spy drama.

ITIL has been around for a surprising amount of time. It was developed in the 1990s for applying what was then considered to be best practice to the management of IT infrastructure in the British government and civil service (where else do spy dramas begin?). At the time, there was little management of IT systems anywhere, unless you worked on a mainframe or with Unix. In the latter case, management was more a case of incantations, sacrifices and wizard-like legerdemain.

It was only around the turn of the millennium that ITIL came more into the public consciousness, as a commercial organization was established to promote it. ITIL's branding of IT Service Management wrapped a business face around technical practices, and so far its branding has been much more successful than that of "System Administration", as promoted by organizations like the System Administrator's Guild (SAGE) and the League of Professional System Administrators (LOPSA).

A lot of unquestioning adoration of ITIL has been written in books and on the web, since ITILv2, and I have been consistently astonished at how easily academics and industry researchers have thrown themselves into the teachings of ITIL without so much as a question as a question about how one might define "best practice", or in what context. One of the key concepts that ITIL promoted was that of the "CMDB" or Configuration Management Database Today, the CMDB has been elevated to a status of "must have accessory" in many larger organizations, especially those that have auditing requirements.

Imagine my surprise then to read a book by industry analysts that is both more insightful and more balanced in its view of the state of the industry than most papers I have to read. O'Donnel and Casanova (surely good spy names) lay out what is good and bad about the concept of the CMDB (and ITIL), in the CMDB Imperative, and sketch out how it is not the be-all and end-all of IT management -- but really more of a concept in progress than an off-the-shelf application.

A couple of years ago, with the rewrite of Cfengine 3, I got interested in the issue of Knowledge Management, seeing it as the future of the IT management problem. Knowledge Management entered more prominently in version 3 of ITIL, through the DIKW lifecycle (from Data to Information to Knowledge and then hopefully Wisdom). I believe this to be a worthy focus, indeed the challenge of the coming decade. Most of today's technologies do little to address this challenge, though we are dabbling in this at Cfengine.

From the ashes of this Phoenix, two other terms that are rising in importance are CMS (Configuration Management System), which is a specialized active component that handles system configuration (fitting nicely with Cfengine), and SKMS (Service Knowledge Management System) which is growing into an actual Knowledge Management tool. Yes, most IT analysts really do talk in these secret codes. Hmmm.

Licensed to sell

I'm not sure whether people read books anymore, but I do hope so. I recommend this book to anyone wanting to understand the challenges of IT management in today's IT climate. It is refreshing to find a book that takes a realistic and considered view, rather than toeing the party line of an industry `movement'. It's not necessary to either love ITIL or leave it. It has sound advice in some areas, while it exceeds its own authority in others. As a crystal ball view of the future of the CMDB, well, stay tuned. I suspect that Cfengine's agents might give the CMDB Imperative a run for its money. In the mean time, this message will hopefully self-destruct within 5 years.