It's finally been done: a serious certification for data centers based on energy efficiency.
But this certification is not going to be easy to obtain. Only those facilities above the 75th percentile on the EPA's benchmark scale will qualify, so it will definitely be meaningful and should encourage more management interest in energy conservation.
This is a high water mark in the effort to rein in runaway data center energy use, and to recognize those who do it. It may help catch the attention of executives who haven't faced these issues, and should provide a real goal for those who take energy cost and the environment seriously. Even if your facility won't qualify, it's at least an incentive to try.
Challenges facing Energy Star for data centers
This is not going to be without detractors. The rating will be based on power usage effectiveness (PUE), which can be controversial itself, but there have been modifications that make PUE better for this purpose.
First, you can't use the "snapshot" measurements that have enabled some data centers to publicize flattering PUE numbers. It will take a year's worth of data to apply, and it has to be the right data, taken in the right way.
Second, measurements must be based on energy, not power, which means a time base is required. You can't just read kilowatts off the UPS. You need kilowatt hours, or kWh. And all measurements must then be converted to "source energy," which means "source BTU." This compensates for different fuel types by recognizing the actual energy needed to produce the power that is used.
So what will make PUE controversial? PUE is a great way to measure improvements in your own facility but is not valid as a comparison among different data centers. However, the EPA has provided a well thought-out improvement to PUE. As configured for the Energy Star rating, it's far better than the nothing we have had until now to formally recognize energy-efficiency achievements.
How was the EPA's Energy Star Label for Data Centers developed? The EPA gathered extensive data on actual, measured performance from 120 data centers that volunteered to provide their information in support of this effort. No small achievement in itself! Data centers were of different sizes, represented a range of industries and were located across the country.
The EPA concluded from its evaluation that "only the annual IT energy consumption was found to be statistically significant in explaining variation in energy use; data showed that facilities with higher IT energy consumption have lower PUE values, on average."
One can only presume the following:
- Facilities that volunteered to provide their data were already measuring their performance and were managing their loads to at least some extent.
- Larger facilities, drawing more "IT energy," have both the incentive and the means to actively attack their cooling loads and to achieve economies of scale for their higher IT loads, resulting in lower PUEs.
Unfortunately, the raw data from which this reference model was developed has not been released, and probably won't be. Data centers that provided their data likely did so with an agreement of anonymity.
The one part of this program I find dubious is that there is no "climate correction factor" for different geographic locations. The EPA concludes that "it is the internal loads -- not outdoor conditions -- that dominate the data center's energy consumption." Making this more confusing is the EPA statement that energy-saving improvements, such as economizers, may yield a higher Energy Star rating, but that "the scale is not specifically designed to award points for such technologies." To this writer, this is obfuscation.
Consider identical 2.0 MW data centers in Minneapolis and Dallas. Minneapolis will be able to use free cooling effectively nearly year-round. Dallas may be able to get some savings but will clearly have fewer hours and days when free cooling can be used to its advantage. It's hard to see, therefore, how the Minneapolis data center won't use less total energy for the same IT load than the one in Dallas, thereby giving it a lower PUE and ranking it in a higher percentile on the EPA scale.
So while the EPA has avoided recommending specific technologies, it seems obvious that certain solutions, like free cooling, will de facto yield better numbers in cooler climates. Only time will tell whether this will skew the scale so that the majority of Energy Star data centers turn out to be in Northern climes.
But regardless of possible defects in the program, this is a giant step forward, with Energy Star a worthy label to hang on your building. The EPA should be commended for following this through and giving us another way to quantify how we're doing in this important quest to save energy.
And the government trumped the U.S. Green Buildings Council (USGBC) on this one! The USGBC, ostensibly the "keeper of the keys" for green design, has dragged its feet in considering data centers worthy of a different standard for its coveted LEED certification. Despite industry pressures and the startling numbers in the infamous 2008 EPA report on Data Center Energy Consumption, only recently has the USGBC considered that energy savings might be an important factor in awarding LEED points for data centers. The USGBC finally formed a committee, but hasn't announced anything yet.
As of June 7, 2010, the EPA has!
ABOUT THE AUTHOR: Robert McFarlane has spent more than 30 years in communications consulting, with experience in every segment of the industry. McFarlane was a pioneer in the field of building cabling design and a leading expert on trading floor and data center design. He is currently president of the Interport Financial Division of New York-based Shen Milsom & Wilke Inc. and a data center power expert.
What did you think of this feature? Write to SearchDataCenter.com's Matt Stansberry about your data center concerns at firstname.lastname@example.org.