In a recent entry on this blog I remarked that “apparently restaurant owners are easy preys for incompetent Web site designers” because they often end up with stupid Flash-only web sites. Well, it’s not just restaurant owners apparently, it’s also leading technology companies like IBM, Microsoft, HP, Cisco, etc. For proof, have a look at the web site for the CML effort: http://www.cml-project.org/
The only reason I would subject you to this awful site is that there is some news about CML. The group of companies involved in this effort has released a white paper to explain what CML is. You can spare yourself the Flash part by getting the white paper directly. If you want to spare yourself even more, you will also skip reading the white paper and just read my not-so-enthusiastic summary bellow. And you need neither Flash nor Acrobat to do so.
CML is going to solve parts or all of all problems related to IT management, and it will be based on SML. So much we can tell from the paper. Any other information that we glean there is contradicted somewhere else in the same paper.
No, really, isn’t CML a set of model elements expressed in SML? Yes, it says so:
“At its core, CML is a collection of models which are expressed as XML documents that describe IT entities and their relationships. As the basis for common modeling elements and semantics, the models describe information which can be exchanged between management tools and managed resources. This information is about IT systems and includes infrastructure (e.g., servers, application servers, Web services), logical entities (e.g., software license, incident reports, IT roles), and relationships (e.g., hosted, is hosted by, supplied by).”
But the next paragraph says:
“However, CML does not attempt to present a single model or single set of models which will ensure integration. Instead, CML provides a means for creating new models or extending, combining, or evolving existing models.”
We are also told that even though it provides a model of a server, it doesn’t supercede CIM. Oh, and it may not be just models either, it will also tell you about on-the-wire protocols:
“recommendations in the documentation, including the need to transmit models in an interoperable manner via an agreed upon set of wire protocols”
The “CML overview” picture leaves us with the same impression that it does everything and across the entire lifecycle at that. And I must admit that the nuance between “common modeling elements” and “shared modeling elements” escapes me, even after reading the definitions. I sounds like they are all reusable but some are more reusable than others…
If you are looking for a ray of hope, I found mine in the acknowledgement of the potential role of RDF/OWL which, combined with section 4.1.5, can be seen as hinting at an effort towards creating some simple ontologies for management integration. Which could be very useful if well-scoped.
If you are wondering about timelines, you’re out of luck. When referring to CML the white paper mixes present and future tenses, and also throws in some conditionals for good measure. Hard to guess how much is real at this point. And the next steps are? More scenarios and some guidelines. See you in 2009 (which BTW is consistent with my little theory about Oslo’s impact on CML).
All in all, this doesn’t mean nothing valuable will ever come out of CML. It just means that the group still hasn’t figured out what it wants to be when it grows up.
[UPDATED on 2008/01/24: Good news! They revamped the site to remove the stupid Flash interface. I hope my rant provided ammunitions to those inside the CML group who pushed for sanity. Also, they have put out a press release to announce, retroactively, the white paper. No surprise in the content of the press release and the associated vendor quotes. I wish that whoever wrote the quote for my ex-boss Mark Potts knew the difference between “compliment” and “complement” though.]