If correct, it would seem that the dire predictions of data centers being the scourge of the earth, energy-wise, may not be as high as the often-cited original 2007 EPA report to Congress, which projected as much as a 100 percent increase over the 2006 to 2011 time period.
The 24-page Koomey report was based on a series of prior reports by Professor Koomey (Koomey 2007a, 2007b, 2008b), as well as other well-documented sources. However, the underlying basis of the calculations seemed heavily focused on the number of 'volume servers' shipped, as well higher-performance servers and other IT equipment (storage and networking gear), from 2010 information supplied by IDC.
Key Excerpts from the 2011 Koomey report:
Electricity used by data centers worldwide increased by about 56 percent from 2005 to 2010 instead of doubling (as it did from 2000 to 2005), while in the U.S. it increased by about 36 percent instead of doubling.
Electricity used in U.S. data centers in 2010 was significantly lower than predicted by the EPA's 2007 report to Congress on data centers. That result reflected this study's reduced electricity growth rates compared to earlier estimates which were driven mainly by a lower server installed base than was earlier predicted rather than the efficiency improvements anticipated in the report to Congress.
The report's conclusions seemed reasonable in showing that energy growth was less than expected and the methodology looked sound at first blush. Moreover, it even references the EPA's Energy Star Data Center survey in 2009 (1.93 PUE) and the Uptime Institute Data Center Industry Survey in 2011 (PUE average of 1.83) and the positive impact of the improving PUE numbers (from an assumed 2.0 used in prior reports).
However, for all the well-compiled and analyzed information the report seems to have utilized, it ignores another view of the data center industry energy usage ... the actual data center itself.
While I have no detailed quantifiable evidence to question the results, I feel that other factors in the data center industry, such as the massive and ongoing growth of mega-size data centers, makes me believe that there is a larger increase in total energy usage than the report indicates. I would therefore suggest that there are several alternate (or at least additional) data sets that could also be used to more effectively analyze energy usage and trends in the data center:
The Bottom Line
While I recognize that what I suggest is not a perfect data set, nor is it all readily available, it seems that these additional figures could provide a more complete basis for analyzing the actual power being used by data centers, rather than the volume server shipments that are the primary underpinnings of the report.