Remember the good old days when applications ran on a single server, user access was controlled and there weren't...
too many devices, connections and access points to count? Remember how we once were able to draw a reasonable diagram of our operating environment? We used to be able to name and identify all the nodes and accurately approximate the source of a problem based on the symptoms.
Ahh, the days when spreadsheets, a couple of matrices, and some diligent detective work was (nearly) all that was needed to keep servers and services up, running and available. As author and humorist PJ O'Rourke observed (in a slightly different context), past perfect in all its manifestations is simply a verb tense, not revealed truth. Let's consider some of the realities of managing change in that 'simpler' time.
Remember the fun and frivolity involved in keeping track of the configuration of each server, what changes needed to be made and which had been made? Remember putting together and managing the update schedule and asking where exactly did we file our change records? For some reason, updates to the data in the log file, update book or spreadsheet never seemed to exactly line up with what was actually on the server.
Remember the days of the 'sneakernet?' Updates and patches allowed us to renew acquaintances and start new ones with each and every user as we walked the update disk around the enterprise. Keeping abreast of the application version actually installed or latest patch applied required a combination of the tracking abilities of a well-trained bloodhound and script interpretation and intuition that would reduce to tears a pharmacist adept at deciphering the cramped hand-writing of a hurried physician.
Today's change management environment
The walk-around was to be replaced by automated, remotely managed and implemented software distribution products that would make us much more efficient and save a lot of shoe leather. Unfortunately, the tools always seemed to need just a bit more tweaking and adjustment before they would do the 'perfect' update. Inconsistent formats, data translation and data reconciliation all had to be done manually, relying on an accumulation of experience and respectful questioning of more seasoned associates.
Today's environment is even more challenging to manage and handle. No, the tools aren't perfect, and we are a long way from having all of the record keeping, process definition and implementation in precisely the shape we want it, but at least we don't have to search our desks and filing cabinets for the current settings on the Web server. It's right there in the CMDB … now what's that server called?
ABOUT THE AUTHOR: Richard Ptak is an analyst with Ptak, Noel & Associates. He has over 30 years experience in systems product management.