I've recently been talking about an operations consolidation process with a large SMB that has more than a dozen medium size business units scattered throughout the country. They've been considering consolidating there computing operations into a single datacenter which they would build where the company got its start (and still has a presence), a small town in north central Pennsylvania.
They don't currently have a traditional datacenter; each unit's main office basically has a computer room with anywhere up to 50 or so servers in a few racks, and with a couple of local IT guys who are responsible not only for their computer room, but also for supporting local operations. The owner of the business believes that if he consolidates those servers into one datacenter he will get the efficiencies of scale such an operation can bring while freeing up his local IT guys to support end-users more effectively and to deploy new, forward looking applications and services.
The business owner had given me a call just to get another opinion; to make sure that he had crossed all his ‘t's" and dotted all his "i's", and at the time we spoke it seemed that he had done his due diligence and that his plans were solid and made sense. However, a recent news article I caught has made me rethink his issues. It seems that a small town in Oregon lost a datacenter project because of their inability to deliver enough power to meet the datacenter's needs without a 15-20 million dollar overhaul of the local power grid..
I admit that I totally overlooked that possibility; residing in the Northeast Corridor on the Eastern Seaboard when I have business concerns about power delivery it's about cost, not about the fundamental capacity of the system to deliver what I need. So as you look at rural, low-overhead locations for your future datacenter plans, here's one more item to add to your checklist: How much will it cost to get power to where I need it.