In the fall we posted information about the efficiency of Google data centers and promised to update this information every quarter. We've now collected data for the fourth quarter of 2008 and published them to our sustainable computing website. Specifically, we're keeping track of the efficiency of any Google-designed data center with an IT load of at least 5 MW and a time-in-operation of at least 6 months. In Q4 our average power and cooling overhead in these facilities was 16%, bringing the overhead for the trailing 12 months to 19% (down from 21% a quarter earlier). For comparison, a recent EPA report put the overhead of the average enterprise data center at 100% or higher. We're very happy to have further improved our efficiency, and a number of factors contributed to that result.
First, efficiency is affected by seasonal weather patterns — cooler weather is better than hot weather, and several of our facilities benefited from that in Q4. Also, we continually review our efficiency metrics so that we notice, for example, that one of our data centers is not performing consistently with others of similar size and locale. So we'll take a closer look at optimizing that facility. Are we using fans to cool spaces that don't need to be cooled? Is the thermostat at the right set-point? Can we reduce the time the chillers need to run while keeping the machines operational? So we apply lessons we've learned from better-performing data centers to other facilities, and several such improvements took place in Q4. For the nitty-gritty technical details, visit our data center efficiency page.
While we've made a lot of progress in data center efficiency, we're still learning. As we continue to explore ways to use the least amount of power to do the most amount of computing, we'll continue to share our data and best practices with you. In early March we will participate in the CeBIT conference where we plan to disclose more details on our sustainability efforts as part of this year's theme of "Green IT." Stay tuned.
Tuesday, January 27, 2009
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment