Thunking for work
www.flickr.com
This is a Flickr badge showing public photos from buraddo_bon. Make your own badge here.
« What is Thunking? | Main | ITIL; A good open... »
Wednesday May 30, 2007
CMDB: Six degrees of Separation From Kevin Bacon

A key role of a configuration management database (CMDB) is to store and report information about configuration items. A configuration item can be something as macro or micro as you like. It can be "the street address of your datacenter" or "the dip switch setting on a NIC". It stores all sorts of useful information on each item, like versions, known errors, history of incidents/upgrades, you name it and that information "could" be stored in a CMDB.

The Kevin Bacon game is an old favorite based on concepts of the shrinking world mostly attributed to the author Frigyes Karinthy. The game version is simple: pick any 2 people (actor, director, grip, stuntman) in Hollywood (or the world depending on the version) and there are no more that six links required to relate those people together based on there link to Kevin Bacon.

The underlying concept of the Kevin Bacon game is the utopian vision of the Configuration Management DataBase (CMDB).

A CMDB stores relationship information between each configuration item and other important bits of information (e.g. incidents, service levels, etc..). The way in which this information could be used to improve IT services delivery are limitless. Software Licensing metering, patch analysis, maintenance impact analysis, TCO calculations, capacity management, root cause analysis, incident management to name just a few.

One word of caution. The only thing more mind boggling than the limitless uses of this information, is the technical challenge of making it happen (Demystifying The CMDB, Andrew Conry-Murray). A detailed CMDB would need to store a relationship diagram that would quickly look like a three dimensional bowl of spaghetti. The ability to manage this data in an automated fashion (hence the need for a tool) becomes critical. You must recognize that collecting and managing this information needs a range of processes, including discovery, reconciliation, federation, synchronization, access control, auditing and reporting.

"ITIL cannot be implemented"

I was sitting at a conference in Asia a couple of years ago and a senior member of itSMF got up and made this statement. I can here the ITIL community screaming from here :), but it really resonated with me. Don't get me wrong, I am an ITIL supporter, ITIL Management Certified and active supporter. I had been presenting a continuous improvement approach to ITIL adoption in the conference, so the concept was completely aligned.

My simplistic view of ITIL is it is a collection of good ideas connected together as best practice reference framework. You would be hard pressed to find someone who has executed all of the processes in the framework, and even difficult to find someone who has completely implemented one. A true single CMDB does not exist (you can look here for some products that claim to do some of it), how could anyone have implemented it? Anyway, the rule of ITIL adoption is to work out what benefit (business, operational, etc..) you want to get and then adopt the parts of ITIL that help make that happen.

The classic case study for the failure to have this focus on benefit is best described by efforts of the 2007 F1 GP Factory team. Over the winter break, they went away to develop the FY07 car. The project goal was to improve every aspect of the previous years car. The best practice framework for air flow, engine power mapping, suspension are pretty well known, and these guys strived for the very best execution they could achieve. When they got to the track, they had spent $3 million USD on a car that was 1.5 seconds slower than the previous years car. As a race commentator said; "They spent greater than 3 million dollars on something they could have achieved with a 10 buck lump of lead".

Lesson learned: Always focus on the desired result !!!

The the critical point of adopting Configuration Management is identifying the right detail of information (configuration items) that will achieve the value you are trying to achieve. Don't start trying to catalog every line of a configuration file or field in a database, if your goals are achieved by basic asset management.

Posted at 06:32PM May 30, 2007 by buraddo in Delivering IT Services  |  Comments[2]

Comments:

[Trackback] A key role of a configuration management database (CMDB) is to store and report information about configuration items. A configuration item can be something as macro or micro as you like. It can be "the street address of your datacenter" or "the dip sw...

Posted by Thunking for work on May 31, 2007 at 05:05 PM PDT #

Hi Brad,

Saw you posting in the ITSkeptic about the user interface for CMDB dependecy mapping. Thought you might appreciate some actual live experience.

If you go to my web site you'll see that my company develops infrastructure documentation tools and there are white papers, presentations etc. on real aspects of CMDBs including customers.

To answer your question about we provide 3 or 4 different ways to create CIs and map them, though often we "import" them from existing service desks.

1. Manual entry where you select parent/child CIs and then select the type of relationship

2. Import the above from a CSV file or spreadsheet

3. Import from a Visio diagram which has CI groups, relationship types already set.

4. Automatically add when another databases or source has a critical item such as a server created.

This is the service mapping part of ITIL CM, which still is hardly understood. (We don't use the term CMDB as it confuses - we use service mapping as a better term)

Its now got to the stage that I deliver a course every month on how to map services to systems - mainly delivered to ITIL qualified people!

If you're interested in the techniques and how to apply them - drop me a note.

Dave

Posted by David Cuthbertson on June 20, 2008 at 09:53 AM PDT #

Post a Comment:
  • HTML Syntax: NOT allowed