Kevin O’Hear is a senior sales engineer at Blazent
Recently, our son complained of pain in his leg, so we took him to the pediatrician. She isolated the source of the pain but couldn’t make a further diagnosis. We were referred to an orthopedic surgeon for X-rays, but when those proved inconclusive, we were sent for MRI and bloodwork. The MRI review led us to a children’s orthopedic specialist, who was finally able to diagnose a slight hairline fracture. With rest the best prescription, we could finally relax. On the way home my son wondered why so many doctors were involved and why it took so long. My wife tried to explain the concept of specialists to him, but you could see that the sequence of events, timeline to diagnosis and multiple players didn’t make sense to him. I’m sure he was wondering why he couldn’t take all the tests once and then have the team come together to do a quick, accurate diagnosis.
CMDB Data Accuracy
The same concepts apply when populating a CMDB. There are many systems of record that can provide pieces of information with regards to the IT estate, but no single system of record is designed to populate everything a complete CMDB needs. The reason is that today’s CMDB addresses many more attributes of the IT estate than CMDBs of just a few years ago. In the past, only true configuration attributes were tracked. These included things like asset information (asset tag, asset name, serial number) and configuration information about that asset (Operating System and version, IP address, maybe user information, etc.)
Today’s CMDB’s are more like a data warehouse for IT–a single place for traditional Asset and Inventory information, but also Operational Lifecycle data, Governance, Risk, and Compliance data, Usage information. And much, much more.
Traditional CMDB Discovery Tools Insufficient
For these traditional first-generation CMDBs, traditional discovery tools were sufficient. They primarily had two architectures: agent-based or agent-less. As the name implies, the agent-based solutions require some type of software agent installed on each machine that it is trying to monitor and inventory. This allows for detailed asset attribute collection, but is limited to the inventory that it is installed on. The agent-less approach requires software to be installed on a centralized server (i.e. within the datacenter) and then, taking just one example, goes out and scans based on a range of IP addresses. The benefit of this approach is that only a single instance of software needs to be deployed and can dynamically recognize and inventory new assets as they came on line, but only within the scope of what the centralized server has access to.
Blazent Approach to CMDB Data Accuracy
As the above limitations show, traditional discovery tools don’t collect enough information to provide a complete view of the assets within today’s IT estate. Today’s CMDB requires complete, current, and correct IT data to ensure continuous accuracy. Blazent has pioneered the multi-data source approach to enhance your IT asset data. Procurement systems can provide valuable information pertaining to inventory and asset data, but they do not have insight into operational views of that asset once it is deployed. SCCM is a common source for CMDB data, but only provides data pertaining to the Windows footprint. Virtual Center can provide data regarding the virtual machines, but lacks insight into the physical world. Configuration identifiable attributes must be collected from many different systems of record, then aggregated, normalized, and cleansed. Conflicting attribute values must be resolved and missing attributes must be populated. Blazent’s Master Data Management approach provides the business rules and algorithms to achieve this. It provides a framework to look across many different, unrelated data sources and apply user defined logic to achieve a single version of the truth to populate the CMDB.
Just as in the case where it took multiple sources to reach an accurate diagnosis, or a court case, where more witnesses can strengthen ones case, the same can be said that multiple sources provide for a more accurate population of your CMDB.