Sometimes it is hard to understand how so many professionals in an industry latch on to one statistic on which they base all their theories, predictions, and forecasts without ever revisiting the basis of the original statistic or even tracking their performance to see if they are improving or not.

A simple statistic published back in 2006 about how much electrical energy was consumed by data centers launched the industry on an epic effort to improve energy efficiency in these facilities. As an industry, we developed and vociferously debated multiple metrics, created a draft LEED standard specifically for data centers, modified ASHRAE Standards, and produced an untold number of solutions for energy efficiency in data centers. We attended conference after conference to discuss our corporate social responsibilities (CSR) for energy consumption and carbon footprints.

Government at every level has actively engaged the industry and challenged us to develop more and better further performance standards and to promote discussion about the right solutions.

Yet for all these positive efforts, article after article, and speaker after speaker continue to refer to Jonathan G. Koomey, Ph.D.,’s 2007 paper, Estimating Total Power Consumption by Servers in the U.S. and the World. In this paper, Koomey, a staff scientist at Lawrence Berkeley National Laboratory and consulting professor, Stanford University, reported that data centers constituted 1.2 percent of the U.S. electricity sales. As one might expect, as this figure continues to be repeated it has grown and now is referenced as high as 2 percent.

The data in this report were compiled in 2005, and so in 2010 we have no new updated information. Industry analysts, vendors hawking their merchandise, and politicians looking to eliminate global warming all refer back to this old estimate.

What have we accomplished in five years? We deployed virtualization, cloud computing, air-side economizers, high-efficiency UPSs, in-row cooling, and much more. Where are these successes quantified? Which of the five projected curves did we hitch a ride on? (See figure 2.)

With the explosion in computing power, do data center power purchases now constitute 5, 8, or 10 percent of total electrical sales, or have we maintained or actually reduced the total electrical use of data centers? What would be the percent had we not taken these extraordinary measures?

We live in a world where instantaneous and accurate answers are expected. Do a search for more current data on the web and what do you get? Listings for Cisco, Cybercon, data center designers, The Green Data Project, data collection, and more. There is no shortage of white papers and vendors putting their best foot forward.

Perhaps I just do not know the right “key search words” or perhaps I need a specialized search engine, but I did not find any updated estimates of where the industry is going with respect to energy consumption. Further, if these data did exist, one would expect to it be fully exploited in the press and the vendor community alike.

Enlarge this picture Figure 2. Comparison of projected electricity use, all scenarios, 2007 to 2011.

So why don’t these data exist? Why isn’t some unbiased industry group following up on the original report with regular updates? Why can’t an industry that prides itself in producing instantaneous information provide even the most basic knowledge about its own power consumption? Is it some big secret?

At least part of the answer lies in the industry’s continued unwillingness to share information. Take the U.S. Environmental Protection Agency’s (EPA) data collection efforts, for example. Despite its extraordinary efforts over the past four years, the EPA could get information from only 100 data centers on which to base its programs. That’s 100 data centers out of tens of thousands nationwide. The data center industry should be embarrassed by this woefully inadequate participation.

A web search for colocation and hosting firms will turn up over 1,000 companies with multiple facilities. Why aren’t all these commercial entities, each of which measure every microwatt consumed in their operations contributing to the EPA’s database or some similar effort?

What about the thousands of enterprise data centers? What about their CSR when it comes to helping the nation accurately trend its energy consumption and its carbon footprint?

What about the health-care industry, where virtually everything is now regulated and reported? Everything, that is, except their data center power consumption.

OK, so I have asked a lot of questions and have no answers, which means we need a call to action.

Koomey’s 2007 report, based on 2005 data, was a wake up call, but most of us seems to have reached over and hit the snooze button. So now I call upon all of you—users, operators, vendors, consultants, and the like—to organize and start sharing/trending consumption data. Why can’t we create a real-time index like the NASDAQ or the Dow Jones Industrial Average?

Who among us is willing to contribute to an industry index? Yes folks, I am talking about real-time reporting of data center power consumption. Can we as an industry step up to the plate or will it take an act of Congress to force us to do what we should be doing anyway?

Let’s hear from you?

Special Note

In December 28, 2009, New York City Mayor Michael Bloomberg signed a series of energy bills into law that require NYC owners to conduct energy audits to assess efficiency and annually document energy consumption. The benchmarking bill requires owners to track and disclose energy consumption annually so they can see how efficiently their buildings function and enable prospective buyers and tenants to better assess the value of the building. Further, a sub-metering component of the bill recognizes that tenants, not building owners, are the consumers of electricity.

We will no doubt be seeing more of these regulations that include data centers located within buildings.