Recently SearchDataCenter.com spoke to Khattar about his decision to implement containment, the resistance he faced from within his own company, and what data center design is pooh-poohed today but could become common five years from now.What prompted you to consider data center containment five years ago?
Mukesh Khattar: I found that the data center was starting to use a lot more energy, and we found that hot air was seeping back into the cold-air side at the top of the racks. In order to prevent that, we had to blow even more cold air so it pushed the hot air from getting into the cold air side. I wanted to reduce that because if I could reduce extra cooling, I would get efficiencies there. So I decided I wanted to put a physical barrier between the hot and cold air. What kind of barrier did you choose, and why?
M.K.: Hot-aisle containment wasn't very attractive because the temperature in the hot aisle could go into the 90s, the 100s, and it wouldn't be very hospitable for workers. The cold-air containment was not very energy efficient, because you had to supply colder air -- [in the] 50s and 60s -- because the hot air had to be room temperature for people working. So we decided to go with a rack containment solution where we put a duct from the top of the rack to the [ceiling] plenum and then returning back to the CRAC [computer room air-conditioning] units. How did people react when you said you wanted to do containment?
M.K.: There was a lot of resistance . … Nobody had done it in the industry. I had to through navigating within our own operations group, our consulting engineer, with our design construction group, with our suppliers and everybody else. It was a change in practices and they were putting a big investment in, and what if things didn't work out? Who would get the blame? So how did it work out?
M.K.: We were expecting a savings of 19 months payback period. We got less than five months payback in actual operation, and the rest is history. So five years ago, containment was a practically unknown, even frowned upon in data center design. What design today has a lot of resistance but will be common five years from now?
M.K.: I think the industry is probably beginning to see air-side economizers as the differentiator, but there is a lot of hesitation in implementation there. I think in a couple of years it will be taken for granted that you have to use that. But there is the issue of particulates and quality of air, but the real issue is humidity control. This is why ASHRAE widened the recommended humidity range, to make it easier to design facilities with an expanded range of humidity operation, which allows air-side economizers to be more effective.
Mark Fontecchio can be reached at firstname.lastname@example.org.