It takes time, but it’s lovely when you get borne out in an article, as Martin LaMonica discusses how Google manages to handle all those search queries. And it’s really simple – get cheap X86 machines, use an open source stripped-down kernel, fiddle with the filesystem to do simple block transfers with a simple triad mirror (master – dual slave) configuration. So low-cost, easy, and direct.
Of course, I recall the enterprise guys at the time laughing at using commodity cheap servers for “real enterprise applications”. So I guess Google isn’t a real enterprise application company, hmmm?
But there is a little nit in the ointment, so to speak. Power! Having lots of cheap servers and simple management reduces people overhead, but at the cost of much greater power consumption. As Urs Hoelzle, VP Operations and Engineering at Google says “”The physical cost of operations, excluding people, is directly proportional to power costs,” he said. “(Power) becomes a factor in running cheaper operations in a data center. It’s not just buying cheaper components but you also have to have an operating expense that makes sense.”