James Hamilton on Improving PUE

James Hamilton of Amazon has a worthwhile blog post looking at some of the controversies surrounding Power Usage Effectiveness (PUE), which has emerged as the leading metric for data center energy efficiency. The growing number of facilities announcing low PUE numbers has prompted scrutiny about how PUE is being measured, along with pockets of cynicism about facilities and systems being designed to “game” the standard.

The PUE metric, which has been popularized by The Green Grid, compares a facility’s total power usage to the amount of power used by the IT equipment, revealing how much is lost in distribution and conversion. An average PUE of 2.0 indicates that the IT equipment uses about 50 percent of the power to the building. A number of data center operators, including Google, have announced PUE readings of 1.2 or below for their data centers.

Hamilton identifies some of the areas of contention around PUE, and then proposes a new standard called tPUE (Total Power Usage Efficiency) that refines PUE with standard processes for measuring total facility power and “productive” IT equipment power.

Hamilton’s tPUE proposes that total facility power be measured at medium voltage before all UPS and power conditioning takes place, while “Productive IT Equipment Power” measures server power, while excluding PSU, VRD and other conversion losses as well as fan or cooling power consumption.

Defining IT Equipment
Hamilton says this approach would address variability in what components are measured as part of the IT load, a distinction that shifts in containers and other design scenarios. “What we want to measure is the power delivered to individual components,” Hamilton writes. “We want to measure the power delivered to memory, CPU, etc. Our goal is to get power after the last conversion.”

Will this proposed revision to PUE gain traction? James Hamilton is well respected in the industry from his work with Microsoft and Amazon and has been an important voice in the growing discussion of data center best practices. His blog is widely read, and the tPUE proposal is bound to be seriously reviewed and discussed in the industry.

Green Grid & Clarifying PUE
Where’s The Green Grid on this? Hiding its jewels behind a paywall, unfortunately. In March The Green Grid sought to address concerns about PUE reporting by issuing a white paper with “a set of rules and guidelines and a required process that organizations should follow when making public claims as to PUE or DCiE measurements.”

Here’s the problem: The white paper is restricted to members of The Green Grid, which costs $5,000 a year to join.

Fortunately, Green Grid board member John Tuccillo provided an overview of the group’s approach at the Google Data Center Summit in April. He outlined three levels of PUE measurement, based on the particulars of where and when measurements are taken (see chart and video of Tuccillio’s talk). The Basic, Intermediate and Advanced PUE ratings are defined by where the IT equipment power and total facility power measurements are taken, and how often the data is collected.

Get Daily Email News from DCK!
Subscribe now and get our special report, "The World's Most Unique Data Centers."

Enter your email to receive messages about offerings by Penton, its brands, affiliates and/or third-party partners, consistent with Penton's Privacy Policy.

About the Author

Rich Miller is the founder and editor at large of Data Center Knowledge, and has been reporting on the data center sector since 2000. He has tracked the growing impact of high-density computing on the power and cooling of data centers, and the resulting push for improved energy efficiency in these facilities.

Add Your Comments

  • (will not be published)


  1. mark macauley

    I have had issues with PUE, metered power, and any other measurement as I equate the development of the yardsticks with which we are measured with legislation - those who write it don't 'get' it from an operations perspective. My personal belief is that you pick whatever measurement you want to use ind IMPROVE it. Your facilities, your yardstick, your numbers. You either get better and more efficient or you don't.