Keeping track of every version of each piece of hardware and software in your production infrastructure is key to ensuring that everything is working properly. So when something goes wrong and you must rebuild a system fast and accurately, such controls become a crucial part of recovery.
But the discipline required to build and maintain accurate records about production configuration can be daunting.
A modern enterprise IT production environment contains hundreds or thousands of major itemsservers, routers, switches, storage deviceseach of which contains dozens or hundreds of significant configuration items. Keeping even minimal information about each item results in a huge database. Most of these items can be discovered using automatic scanning tools, but even the best tools can't always populate every configuration item fully and accurately.
So you also need links to the asset databaseusually buried somewhere in financewhich records everything you bought to go into the production infrastructure, and you'll probably need some manual inspection and data entry with appropriate validation and verification to fill in the gaps.
You'll also want to track what parts of the production workload run on which equipment, so if something breaks you know which workloads and customers will be affected.
Getting to the end of all this is extremely taxing, but you are only at the beginning of the real challenge, because next and forever after you must track every change.
If you have 100,000 configuration items in the database and you refresh your production technology every four years, you will have to update 25,000 items every year or about 70 items a day on average. (Averages aren't very useful here because the change process is "bursty," with distinct peaks during major refreshes.) That's a lot of work just to stay up to date, even with automatic discovery and registration of new items as they are deployed.
Next page: Opportunity for Error