For most of the organizations, maintaining a healthy CMDB (Configuration Management DataBase) is a really hard job.
Is it really that difficult or are we doing something wrong?
Let me tell you something, it is not that difficult and yes, you are doing something wrong.
Organizations start projects for CMDB implementation and most of them result with a failure. You are not destined to fail, all you need is to plan it right and take time to analyze before you implement, not the other way around.
And do not forget that the configuration management process is the core of IT service management, so have your CMDB and configuration management process before you start with other processes like change management for the best result.
First of all, start with a good scope definition. You don’t want to fill your CMDB with configuration items that you are not really dealing with. Take time to understand what different units work on, where do they keep their data, whether scanning and auto-populating the CMDB is possible, and on what extent the scanning tools can populate your CMDB.
Now you know what you can do and what you should do.
Decide on the classes and carefully select the attributes for each class depending on your needs and the capability of your data sources. You now have your organization’s data dictionary with all the details.
The next step is to decide on the relations between classes: Applications run on servers, depends on network and so on.
You are now ready to implement it once you have everything ready on paper.
Then comes the related processes, your service management will be shaped around the CMDB. How you are going to track the change history, when you are going to update a CI, how you are going to use the dependencies in order to fix the problems, they are all the subjects of related processes but cannot be done if you don’t have a steady CMDB.
It is not that difficult to accomplish but it must be done in the right order. Do it right once and you will be ready to excel in the other processes as well.