The idea is the same: build a room that can house all of the computing power necessary for a client. As the processing capability of each square inch of server exponentially grows, so does the amount of power needed and heat generated. The methods of providing reliable power and cooling continually evolve to meet these changing needs – and that is where the fun begins.
Just a decade ago the solution was simple: Put enough computer room air conditioning (CRAC) units in the room, add a few more for redundancy, and you’re done. However, using air to cool rooms is becoming increasingly difficult, if not impossible. Water cooled solutions are now common place and the systems and equipment are becoming more complex.
The ultimate achievement of most data centers is 99.999% reliability. This means the data center will be off-line 0.001% of the time, or 5 minutes a year. In reality, the answer is that the facility manager/owner/customer expects 100% uptime. Nobody wants downtime. Designing systems that respond to equipment failure without impacting the data center equipment is what we strive to accomplish. The end result being peace of mind knowing the support systems will keep your data available.