Content Copyright © 2006 Bloor. All Rights Reserved.
I have recently been taking an interest in configuration management databases (CMDBs). The way that I think about them is as master data management (MDM) for the IT department, though there are a number of differences, at least in the way that they are implemented.
Perhaps the biggest difference between CMDB and MDM is that the latter are typically implemented as siloed solutions that deal with just a single aspect of master data. Thus there are separate customer (CDI), product (PIM) and supplier (GSM) solutions. By contrast, a CMDB should encompass servers, routers, applications, operation systems, business processes, people and whatever other assets are available within the IT environment.
The second major difference is that MDM solutions tend to be either registry or hub solutions, whereby the former involves storing pointers to the data you are mastering and the latter involves holding a golden copy of the data centrally. By contrast, CMDB solutions need to take a hybrid approach with some data held centrally but using a federated approach for other assets.
This leads me to another difference. MDM vendors tend to put a lot of emphasis on data quality whereas CMDB suppliers do not. However, there does not seem to me to be any intrinsic reason why the quality of data should be any better in the IT department than elsewhere in the organisation. However, this is not to say that this is totally ignored: BMC, for example, has reconciliation and de-duping capabilities in the new release of CMDB but it is not full-blown data quality.
Finally, the other main difference is that CMDB products such as BMC’s have auto-discovery mechanisms (even of people through its Identity Discovery capability) that can automatically detect the various elements in play. Moreover, BMC can auto-discover not just assets but also the relationships that exist between assets, which is a significant advantage compared to other CMDB vendors. Leading on from that you can do things like impact analysis and see where particular assets are used.
Returning to the comparison with MDM, I think there are some lessons here that the MDM vendors could learn from.
Apart from its similarity to MDM the other reason why I am getting interested in CMDBs is that they are coming out from the umbrella of systems management and being offered as stand-alone products. Of course, this has always been the case for some of the smaller vendors but BMC has just announced, and will shortly be shipping, CMDB 2.0 (which also includes various new viewing, browsing and analytic capabilities), a part of whose release is CMDB Enterprise Manager, which is the stand-alone version of this product. This should prove particularly attractive to outsourcers, service providers and anyone else who has a diversity of environments to manage and control. It should also prove of interest purely as an asset management system.