Configuration Management: Life or Death for IT
Every decision we make in IT, indeed every decision we make in life, is based upon accurate information. There are no exceptions to this basic law of nature. Think about examples of decisions big and small, both in and out of the IT domain. None can be made with any confidence unless the right information exists.
This sums up the whole notion of configuration management. It is the process and the system of technologies that ensure the right information exists to make decisions, both automated and manual. For this reason, configuration management is the most important of all the ITIL processes. One can effectively argue that change management is equally important, as the two form the nucleus of every function we perform. If I had to choose one over the other, however, configuration gets the edge.
If information is the single most critical element of our world, then collecting, validating, and managing the data elements is the single most important function we need to perform. This is configuration management. There are many dimensions to configuration management, including familiarity with the process, infusing IT staff with this knowledge, and implementing the right technologies.
The traditional way we collected the data that comprises this information has been awful. If it existed at all, it was usually resident in people’s heads and if it existed in electronic form, it was often difficult to access and understand and it was almost never linked to other data to allow more sophisticated analysis and decision making. As our technologies are maturing, we are finally seeing the emergence of products that can automate the collection, reconciliation, assembly, verification, and management of this data. If these steps are automated, there is a far higher likelihood that the data will be accurate.
Accuracy is the thorniest, but most crucial aspect of configuration management. If the data is not accurate, it is useless. I learned this lesson in 1986 when I built my first configuration management database (CMDB). It was a great description of my environment, and I had a wonderful staff working hard to populate it. Still, it failed. It failed miserably because we had no means to keep it accurate. Decisions being made were wrong because the data we wanted to rely upon was wrong.
This points out one of O’Donnell’s laws of ITSM: Complete absence of data is better than incorrect data. When you have no data, there is no illusion about its viability. If you have data, you come to rely on it almost with blind faith. As we continue to automate complexity out of our day-to-day lives, the data driving these automations absolutely must be accurate. Automation essentially embodies blind faith into software and blind faith is only acceptable when there is overwhelming trust in the results. With inaccurate data, there can be no trust and therefore, no automation. Configuration management is the answer, but proven processes and the right innovative tools are needed. Because it is so fundamental to IT success, there is no room for sloppy execution in configuration management. Sloppiness will be dangerous, possibly fatal to the entire operation.
Allow me to use an analogy with another remarkably innovative field: genetics.
I was just listening to National Public Radio’s Fresh Air program. The host, Terry Gross was interviewing Craig Ventner, a pioneer in DNA sequencing. There are striking similarities between the CMDB and DNA. They are both the primary source of data that govern the entire behavior of their respective systems: IT and living beings.
The future of genetics promises to alter medicine, longevity, economics, politics, and every aspect of our lives more dramatically than any innovation ever before. What do we need to make this giant leap? Accurate data. Scientists need to map the genome so they can understand the characteristics of their new medications and other developments. Even the slightest error can be disastrous.
Genetic sequencing is proceeding in earnest because these scientists grasp the gravity of their innovations and they must ensure that all bases have been covered. The sequencing provides the data accurately because the sequences are not fabricated, emulated, or embellished in any way. If scientists produced a sequence of my DNA, the results are 100% me, not an assumption or estimation of me. Not a single chemical compound in any of the roughly three billion base pairs can be overlooked.
We can view genetic sequencing as the analog of configuration management in the IT domain. Each is the basic building block of the overall efforts of genetics and IT respectively. Neither can be effective if these basic building blocks are wrong. Actually, they can be downright devastating if they are wrong.
Just as molecular biologists must relentlessly pursue accuracy in their sequencing, we in IT must pursue accuracy in configuration management. Our professional lives depend on it.