Everyone loves the thrill of discovery – finding something new buried in the forest of IT infrastructure, user devices and services. Seasoned service management experts will tell you that the data you discover can be a valuable resource to help your company address incidents, ensure security and manage IT costs. Unfortunately, more isn’t always better if your new discovery muddies the waters of insight instead of clarifying then. Discovery can be good, but proceed carefully and remember: just because you discovered it doesn’t mean you should add it to your CMDB.
You might say, “Okay, so it depends – some newly discovered data is good and some is bad. How do I know the difference?” There are 4 important questions you should ask before adding new data to your CMDB. These apply to data from both existing discovery tools and newly installed capabilities.
- How accurate is the new data? When you compare it to your current data, does it seem to fit into the overall big picture and depict what you expect to see? How well do you trust the accuracy of the source of the data? A simple contextual assessment with a commonsense sanity check will expose many large data accuracy problems. It won’t catch them all, but can help set expectations of data quality.
- Is the data current? Technology environments are constantly changing and evolving, so it is important to make sure that the data in your CMDB is both accurate and current. To assess data currency and check for obsolescence, you may be able to look at time-stamps on records, but you will more likely need to trace the discovered data to its source to understand when it was originally captured. If the data was entered into a source system manually, then it is likely older than if it was collected electronically (though this isn’t always the case). The key is to understand what point in time the data represents.
- Is the new data redundant and/or overlapping with your current data? Data redundancy is one of the biggest quality issues in CMDBs – particularly if a company is utilizing data streams from multiple collectors and discovery tools. Each collector records a specific point-of-view of the environment and it is very common for multiple collectors to see the same configuration item differently. If you suspect this is occurring, then you will want to compare the new data to the existing data in your CMDB and decide whether it is best to keep the original data, replace it with the newly discovered data, or combine the unique elements of both records. What you want to avoid is having multiple records for the same configuration item.
- Do you know the use of discovered data? This is perhaps the most important of the 4 questions, yet it is also the one that the fewest people take the time to answer. Every piece of data that you add to your CMDB creates a technical debt on your company to manage, update and eventually remove from your records. It may seem like a small incremental cost, but the costs of unutilized data increase quickly. Before adding new data to your CMDB, make sure you know its use. If you can’t define a use and purpose, then you likely do not need to add the data to your repository.
With each of these questions, it’s important not to discard data blindly and immediately if it doesn’t fit into your overall configuration picture. It is possible that you have discovered new data that indicates a change in your environment is occurring. These are highly valuable insights if you can find them, so it is well worth the effort to set the data aside for further analysis.
The best way to keep your operational data clean and provide clear insights to decision makers is to improve the quality (not just the quantity) of data you add to your CMDB. A typical company already has critical deficiencies in approximately 30% of its CMDB data. Newly discovered data can help to resolve these deficiencies or make them worse. The difference is the sophistication of data quality capabilities and how you use them to validate, verify and reconcile newly discovered data with your current data. By addressing the source of data quality issues as data is added to your CMDB, you will have less data volume to maintain and higher quality information to drive business and operational insights.