The Green Grid › RecommendationsForMeasuringandReportingOverallDataCenterEfficiencyVersion2

Posted In:

Metrics and Measurements

17 May, 2011

Download

The purpose of this document is to provide recommendations on measuring and publishing values for PUE at data centers. These recommendations represent the collective work of a task force representing 7x24 Exchange, ASHRAE, The Green Grid, Silicon Valley Leadership Group, U.S. Department of Energy Save Energy Now Program, U.S. Environmental Protection Agency’s ENERGY STAR Program, United States Green Building Council, and Uptime Institute.

We should not lose touch with the primary objective versus the academics of measuring PUE

Measuring accurately the PUE by some approach whether 0, 1, 2 or 3 is good and scientifically sound but somehow we are losing touch with the original objective of why we are measuring PUE.in the first place. PUE by itself is meaningless if it is just used as a statistical boasting right versus another data centre. Whether it is based on which category the primary objective of all this is to identify which of the systems is not performing or operating efficiently as well as what cause that particular system to behave inefficiently. Using a Category 0 approach is as effective as using a category 3 measurement if the whole exercise enabled the owner to identify his areas of inefficiency and he do something about it. Is it practical to go down to a category 3 level of measuring PUE without going through the same exercise of measuring all the other systems such as CRAC/CRAHs to the same level of granularity or measuring how well the IT equipment such as servers and storages are utilsed? We should not make measuring PUE a complex and expensive process for the data center owner. Keep it simple. Trying to compare one's DC PUE which is what the category level seems want to achieve with the category levels. It is not realistic unless the two data centers are located side by side and are designed and operated exactly the same way.

Posted at 12:29 AM on February 20, 2012 by Robert M Pe

Re: Many issues with this that need to be corrected

Hello Steve, PUE 0 is not intended to be consistent with PUE 1,2, or 3, specific to power versus energy measurements. The Taskforce chose to create a demand measurement for those who are not prepared for an energy measurement, but we wanted to get as many people as possible thinking in the right direction around data center infrastructure energy efficiency. Acknowledging that many folks will begin with demand and move to the more accurate energy measurements is intentional to engage as many adopters as possible. Specific to applying the source energy factors, this is to drive consistency between data centers with different fuel supplies. Comparing a data center which is supplied by primary fuel versus a data center that is supplied by primary and secondary fuels would be inconsistent. The source factor approach makes the calculation consistent among data centers with varying fuel types. Specific to how renewable and CHP production is treated, the boundaries are drawn to include the infrastructure efficiencies on-site. This aligns with the main premise of PUE which is to measure the energy efficiency of the data center infrastructure, not how efficiently one obtains energy. For further dialogue around these topics please send a note to gdcmetrics@lists.thegreengrid.org. Thank you for your comments and we welcome you to engage further with The Green Grid. Best regards, Dan Azevedo

Posted at 10:19 AM on July 26, 2011 by kavi\admin

Many issues with this that need to be corrected

PUE0 uses demands, but 1, 2, and 3 use annual energy - not consistent. PUE 1, 2, and 3 use source energy estimates, which is totally inconsistent with ASHRAE 90.1, 189.1, 90.2, 100, IECC Chapter 5, and LEED. The source energy estimates are totally flawed and will lead to bad design desicions that will increase energy usage. The treatment of on-site renewables and on-site CHP is not logical and inconsistent - why is electric production from renewables treated differently from electric production from CHP (which could be powered by imported oil)? It does not make technical or analytical sense.

Posted at 01:58 PM on June 25, 2011 by Steve Rosenstock

Water consumption in the utility generation system?

It is interesting to note the inclusion of water in the facility energy input. However no offset is accomodated for water usage in the delivered energy from the utility. I raise the point because water usage can be high locally in a system that enables a low PUE but that very lowering of grid consumption can result in an overall reduction in water usage. See our White Paper for a UK application on www.arkcontinuity.co.uk

Metrics and Measurements

Report

Comment Policy

Comment Policy

Recommendations For Measuring and Reporting Overall Data Center Efficiency Version 2 - Measuring PUE for Data Centers (May 2011)

Share

We should not lose touch with the primary objective versus the academics of measuring PUE

Re: Many issues with this that need to be corrected

Many issues with this that need to be corrected

Water consumption in the utility generation system?

Comment On This Document

Related Content