US Data Centers Wary of Sharing Energy Data With Feds 101
1sockchuck writes "The EPA has been seeking at least 100 data center operators willing to share data about their energy usage to help the government develop an Energy Star program for data centers. Thus far, only 54 data centers have signed up, which suggests that few data center operators are eager to tell the government exactly how much energy they are using. The EPA issued a report to Congress last year on data center power usage, and is already developing an Energy Star program to rate servers. Can a program designed to rank the energy efficiency of appliances and computer monitors be a useful tool in addressing the enormous energy consumption of data centers?"
Re:Few? (Score:5, Informative)
They are seeking "at least 100 data centers" to participate.
It DOES NOT SAY that they asked 100. They have put out a general request for datacenters of 1000 square feet or more.
So, out of all the data centers in the US that are over 1000 square feet, only 54 have agreed to participate. That is NOT 54/100 or 54%, it is a drop in the bucket, and it does seem to indicate a reluctance to participate. Although it could just as easily be apathy.
Re:Few? (Score:4, Informative)
Re:Might not have anything to share (Score:4, Informative)
Re:Few? (Score:4, Informative)
However, even RTFA-ing twice didn't give you a complete grasp. The article never mentions querying any data centers, just that they want them. They may have called/emailed/mailed to some, or they may have just put it up on the web and hoped.
Additionally, it says that the participants will begin collecting 12 consecutive months of data at the start of the program, "and submit the data by June 1, 2009." That's next year.
Re:I can see the headlines now.. (Score:3, Informative)
Plus, data center managers are generally busy enough to avoid yet another stack of forms to fill.
Finally, some DC managers barely know where the power is going. They see the power bill, they seem some automated power meter reading at key points such as power distribution units, but they cannot really tell you how much a machine, or even a rack, is consuming.
The problem with manufacturers' plates is that the power rating on the plates is ridiculous overkill. A rack-moounted server rated for 600 W on the plate might draw 500 W at start up, when all the fans and hard drive motors are going to max power consumption (electric motors consume more power when starting up than in normal operations). The rest of the time, the server might consume less that 400 W. So you need to drag the server to a lab where its power consumption will be tested under various conditions.
Baring this, the DC people practice the fine art of "de-rating", in other words, they make an educated guess about how much the manuf's plate is exagerated. This can be hard to explain to bureaucrats looking for precise answers for their surveys.
Re:Might not have anything to share (Score:3, Informative)
Re:Transportation energy use is the key (Score:5, Informative)