Commentary

Roger Smith
 

Measuring Data Center Power Efficiency

Data centers consume massive amounts of energy. In a Forbes commentary in August, Kenneth Brill, executive director of the Uptime Institute (which monitors data center uptime), said the number of servers in data centers in the United States has grown from 5 million in 2000, to 10 million in 2005, to a projected 15 million in 2010.

Data centers consume massive amounts of energy. In a Forbes commentary in August, Kenneth Brill, executive director of the Uptime Institute (which monitors data center uptime), said the number of servers in data centers in the United States has grown from 5 million in 2000, to 10 million in 2005, to a projected 15 million in 2010."To avoid future energy shortages caused by increasing IT demands, 10 more power plants need to be built to the tune of $2 billion to $6 billion each and their cost is ultimately going to get passed on to IT through increased utility bills," Brill said.

Those kinds of numbers obviously will get the attention of government agencies. The U.S. Department of Energy and the U.S. Environmental Protection Agency recently published results of the two agencies' National Data Center Energy Efficiency Strategy Workshop, held July 8. The report, Energy Efficiency in Data Centers: Recommendations for Government-Industry Coordination, looks at different measurements currently being developed to help data centers calculate their power efficiency. Three leading metrics are PUE (power usage effectiveness), DCiE (data center infrastructure efficiency), and the Department of Energy's Data Center Energy Profiler, or DC Pro, which is a less mathematical and more "best practice" approach to measuring how efficiently a data center consumes power.


More Software Insights

White Papers

More >>

Reports

More >>

Webcasts

More >>

PUE is an emerging standard promoted by The Green Grid and others in the data center industry that allows data center managers to calculate how much power is driving the actual IT equipment versus non-IT elements such as cooling and lighting. The formula for PUE is Total Facility Power divided by IT Equipment Power: PUE = (Total Facility Power/IT Equipment Power). DCiE is its reciprocal, or PUE divided by 1. It is defined as: (1/PUE) or (IT Equipment Power/Total Facility Power) x 100%. According to Data Center Knowledge, the typical enterprise data center is estimated to have a PUE of 2.0 or higher, which indicates that for every watt of IT power, an additional watt is consumed to cool and distribute power to the IT equipment.

The Department of Energy's DC Pro is an online software tool designed to help industries "diagnose" how energy is being used by their data centers and how they might save energy and money. The beta version of DC Pro was tested by 610 users, and comments received from many of those users have been incorporated into Version 1.0 of the tool, which is now available for download. The tool asks an IT manager a six-page series of questions such as "Do you have a process for identifying abandoned/un-used servers and taking them offline?" or "What is the average age at which you replace your servers?" or "Are you using virtualization to consolidate your server workloads?"

Some of the DC Pro questions require hard number answers ("Lighting power density (watts per square foot?") or percentages, and for others simple "yes" or "no" answers will suffice. The software tool can help data centers quickly analyze their energy use and begin identifying opportunities for savings. It does this by detailing how energy is used by your data center, describing your plant's savings potential, and pointing out some specific actions you can take to realize these savings.

Uptime's Kenneth Brill is one of many people not enamored with the PUE and DCiE metrics. He says the PUE number has become a competitive tool for marketing manipulation and misinformation. According to Brill, data center operators reporting incredibly low PUE numbers may be reporting a spot measurement true only at an instant in time. "If averaged for a minimum of 24 hours, a week or, even better, a full year, the number would likely be much higher." Brill also points out that "benchmarking is useful only if you are comparing yourself with others with similar power and cooling equipment types, percentages of load utilization, tier design levels, and ZIP codes."

Brill makes a strong case for taking low PUE data center numbers with a grain of salt. Rather than getting bogged down in arguing power ratios, such as static versus dynamic or input versus output, IT managers should consider using a first-step tool like DC Pro that can identify opportunities for savings, improve their company's bottom line, and help reduce the environmental emissions associated with energy production and use.

Data center managers who want to further crunch energy numbers might consider cutting power consumption in their data centers in nonstandard ways, such as using direct current (DC) input power -- in lieu of alternating current (AC) -- in order to eliminate the need for wasteful AC-DC power conversion and cut down on waste heat and component failure. According to a data sheet from Fremont, Calif.-based Rackable Systems, distributing redundant DC power to each server -- and replacing the standard AC power supply with more reliable and efficient DC power supply -- will shift approximately 20% to 40% of the thermal load outside the servers or storage systems. As a result, system reliability is dramatically increased, and monthly power costs are reduced by as much as 30%.


Related Reading




Currently we allow the following HTML tags in comments:

Single tags

These tags can be used alone and don't need an ending tag.

<br> Defines a single line break

<hr> Defines a horizontal line

Matching tags

These require an ending tag - e.g. <i>italic text</i>

<a> Defines an anchor

<b> Defines bold text

<big> Defines big text

<blockquote> Defines a long quotation

<caption> Defines a table caption

<cite> Defines a citation

<code> Defines computer code text

<em> Defines emphasized text

<fieldset> Defines a border around elements in a form

<h1> This is heading 1

<h2> This is heading 2

<h3> This is heading 3

<h4> This is heading 4

<h5> This is heading 5

<h6> This is heading 6

<i> Defines italic text

<p> Defines a paragraph

<pre> Defines preformatted text

<q> Defines a short quotation

<samp> Defines sample computer code text

<small> Defines small text

<span> Defines a section in a document

<s> Defines strikethrough text

<strike> Defines strikethrough text

<strong> Defines strong text

<sub> Defines subscripted text

<sup> Defines superscripted text

<u> Defines underlined text

InformationWeek encourages readers to engage in spirited, healthy debate, including taking us to task. However, InformationWeek moderates all comments posted to our site, and reserves the right to modify or remove any content that it determines to be derogatory, offensive, inflammatory, vulgar, irrelevant/off-topic, racist or obvious marketing/SPAM. InformationWeek further reserves the right to disable the profile of any commenter participating in said activities.

Disqus Tips To upload an avatar photo, first complete your Disqus profile. | View the list of supported HTML tags you can use to style comments. | Please read our commenting policy.
T-Shirt Giveaway T-Shirt Giveaway: Each week we're selecting one great comment from our readers. The author of the comment will receive an InformaitonWeek Community t-shirt. So get posting!
Subscribe to RSS

Resource Links