When I worked at the Jefferson County Police Department, we had a full fledged data center.  It had the giant air conditioner, backup batteries, raised floor —  The Works. But of course, it had a multi-million dollar mainframe surrounding it as well.  When we moved from one headquarters to another, the mainframe didn’t make the move. It actually would have cost more for the scrap dealer to come get it than it was worth. We left it there for the Corrections Department to deal with.

In the new headquarters, we had downsized entirely onto a Novell NetWare driven network. Beyond the saving cost of moving an antiquated IBM 4381 mainframe to the new headquarters, one of the cost saving advantages we had to the downsize was not having to have the climate controlled environment.

Our “data center” was reduced in size to four NetWare servers — tower PCs at that. They were located in an oversize office next to mine. There was no air conditioner. It had some basic server-class UPSes. Other than that, all it had were the servers, some monitors, the patch panels for the network, and a few routers.

With rack mounted servers, blade servers, rack mounted UPSes, and so forth today, it seems like the data centers of old are slowly returning. Although server-room closets still exist in small to medium businesses, often you see the glass houses of the mainframe-era IT department coming back in larger businesses. 

Does a “data center” have to be inside of a fishbowl? When does a data center become a data center?