Taken together, the two presentations form a roadmap for reducing energy use in data centers by employing commonsense when designing and building facilities. If you ever get a chance to hear either speaker live, do so: both are funny and engaging as well as informed.
I’ll start with Bhandarkar’s presentation because it really should have been first on the DatacenterDynamics program. He began with the premise that the data center is the server, and he pointed out how optimizing servers can make them the starting point for wringing a lot of waste out of data center operation. Bhandarkar also noted that server performance varies and that lower power processors often offer the best performance per watt, which leads to the conclusion that right-sizing server configuration can improve overall energy efficiency.
Bhandarkar’s background at Intel serves him well in executing this strategy because he went on to explain that Microsoft servers do not include some of the drives, ports, and fans that are standard on most server offerings. Certainly it’s the kind of insight that Ken Baker of HP had in mind when noted during his later presentation that being a server manufacturer gave HP had certain advantages in designing and operating data centers.
Noteboom’s presentation picked up right where Bhandarkar left off, as he told his audience of Yahoo!’s success in Lockport. Success, he explained, was achieved by smart siting policies, determination to innovate to control electrical and cooling costs, and commitment to efficiency.
More interesting, though, was how he described Yahoo!’s mind set. In doing so, he likened the ongoing maturation of the data center industry to the evolution of the auto industry. “Henry Ford,” he said, “popped the legacy bubble. So must you.” Noteboom is the direct descendant of an automaker. Not Henry Ford, but one of the 2000 plus manufacturers in business in 1908, who were soon to be displaced by Ford’s production line system of manufacturing. I believe the proliferation of container systems is just one phase of maturation and perhaps not even the final move to standardization as the industry continues to develop.
When I remarked about the success of these panels to Steve Worn, DatacenterDynamics CTO, he told me that the two concurrent sessions I had missed were just as interesting and had full attendance. These presentations, he said, had been put on the agenda later in the scheduling process as the result of interest in “mops, sops, and ops” expressed in research conducted by the organizers. This powerful insight led them to invite speakers who could address topics like operating procedures and policies. In these instances, attendees heard about the Uptime Institute’s new Operational Sustainability standard and from Mike Doolan of Johnson Controls about risk management. This research is key to DatacenterDynamics’s ability to constantly unearth new and interesting speakers.
I had the opportunity to talk to Bhandarkar, Noteboom, and Worn and am hoping to provide some space in an upcoming issue to bring their stories to all of Mission Critical’s readers. In the next issue, Julian Kudritzki of the Uptime Institute will be sharing his knowledge of the new Tier Standard.