Most popular movies spawn a sequel, and it looks like the widely accepted power usage effectiveness (PUE) has achieved a similar distinction. PUE has become the most popularly used data center metric lately, perhaps as in some cases, if only as marketing tool. Along these lines are the “Mirror, mirror on the wall, who is the greenest data center of them all?” type of pronouncements recently made by some high-profile social media organizations.

Robert J Meyers, of the U.S. Environmental Protection Agency (EPA) Energy Star for Data Centers program generally agreed. “Since February 2010, EPA and DOE have co-chaired a committee that includes representation from numerous industry groups. Together, these groups issued a set of Guiding Principles for energy-efficiency metrics in data centers, in February 2010. In July 2010, the Task Force issued a Version 1 PUE document, with expanded guidance of how to compute PUE in accordance with these guiding principles, in a standalone data center building. In May 2011, the Task Force released a Version 2 PUE document, which expands on Version 1 to offer guidance for certain mixed-use settings,” he said.

On May 17, the Data Center Efficiency Task Force issued a 14-page document titled, “Recommendations for Measuring and Reporting Version 2–Measuring PUE for Data Centers,” or PUE 2.

table 1

Mapping of PUE categories to TGG framework

The Green Grid originally created PUE (and DCIE) early in 2008, but PUE really gained momentum when the Data Center Efficiency Task Force adopted it in early 2010. The task force comprises the 7x24 Exchange, ASHRAE, 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, Uptime Institute, and, of course, The Green Grid.

Mark Monroe, executive director of The Green Grid, said, "When you look at the organizations involved, it's truly amazing the level of agreement and collaboration we have on the future of efficiency metrics," In addition he reflected, "The new definitions help to ensure the consistency, transparency, and clarity to the reporting of data center energy use and help to move the industry to the next level in terms of progressively better measurement capabilities."

The revised PUE is intended to clarify and reiterate the measurement requirements of the PUE metric and presumably prevent PUE envy. It reflects and reiterates the “Harmonizing Global Metrics for Data Center Energy Efficiency” definition of PUE, which was released by the taskforce in February of this year.

fig 1

PUE has been likened to the EPA's mileage ratings on automobiles.

While still called PUE, the metric now specifies energy usage (expressed in kilowatt-hours) not power (kilowatts) as the primary basis for the metric. This difference still seems to be one of the sources of confusion to those using and quoting PUE numbers for their data centers. Alexandra Sullivan, technical development manager, Energy Star for Commercial Buildings, of the U.S. EPA’s Energy Star program, said, “As we work with data centers in the  program, we continue to find many facilities that measure power rather than energy. EPA is continuing to collaborate with the industry to close this gap and make IT energy metering standard practice. The collective recommendations from this diverse task force seek to provide continuity in the way PUE is measured and reported. Consistency in PUE measurements will help improve energy efficiency by making it easier to track, report, and compare PUE values.”

The revised metric includes four measurement categories (0-3), of which three upper categories (1-3) specify annualized energy consumption as the basis for the calculations. Even PUE “category0,” which still allows power demand in (expressed in kilowatts), now requires that the calculation be based on the highest power used by the facility, not just the lowest power measurement taken on a cold night when the chillers are off. PUE Category0is a demand-based calculation representing the peak load during a 12-month measurement period. This change should eliminate some of the seemingly exaggerated PUE claims.

The new PUE document continues to follow the three guiding principles and the “Harmonizing Global Metrics” document issued in February by The Green Grid and the Task Force and which was also adopted by some international organizations. Moreover, the new document includes more details on how to properly measure and calculate PUE for data centers located in a mixed-use building, in particular building supplied chilled water or condenser water.

fig2

Visual explanation of how the new categories in the revised PUE document differ.

One of the significant changes in PUE Version 2, which may be easy to overlook, is a caveat regarding the requirement to reference the PUE category when proclaiming a data center’s efficiency: “When publishing PUE, the category must clearly be indicated using a subscript e.g., PUE0, PUE1, PUE2, PUE3. A PUE reported without the subscript is not considered to be in compliance with these recommendations.”

Very few (if any) of those who have previously made public proclamations of their PUE results indicated which PUE category methodology and protocols they used to derive their claimed results. Presumably, going forward they will hopefully include the category in future PUE announcements. It should be interesting to see the how many organizations will adhere to this requirement in any new crop of PUE announcements.

It should be noted that the PUE Version 2 document also states, “The U.S. DOE’s DC Pro tool will be incorporating PUE calculations that are consistent with PUE2” and “The U.S. EPA Energy Star Portfolio Manager tool will be incorporating PUE calculations that are consistent with PUE1.” These steps alone should go a long to ensuring that new PUE standard is as well accepted as the earlier version.

Paul Scheihing, DOE Industrial Technologies Program, Office of Energy Efficiency and Renewable Energy stated, “DOE and EPA worked together to unite key data center organizations to develop these PUE metric recommendations.”  Moreover, he added, “The DOE considers the PUE metrics as important and has thus integrated them into our DC Pro tool and Data Center Energy Practitioner training.”

fig3

UPS readings as displayed on an Infrastruxure energy-efficiency dashboard maybe used to determine a PUE1 value

In a related discussion, Sullivan stated that the Energy Star for Data Centers program now has ten data centers that have successfully met the criterion and have received the Energy Star award. The Energy Star Program also requires submission of PUE as annualized energy, not a power snapshot.

Michael K Patterson, senior power and thermal architect, Intel Architecture Group, who represented ASHRAE on the Task Force said, “ASHRAE’s participation, as well as the 9.9 Technical Committee, helped to shape the cooling component energy measurements and values of the PUE [version 2] document, especially in the area of externally supplied chilled or condenser water, such as those found in data centers located in mixed use buildings.”

Moreover, Patterson felt “the TC 9.9 Guidelines helped the data center industry understand how broader environmental conditions can significantly improve energy efficiency. It therefore allows the data center operators to better correlate the relationship between IT equipment reliability and a wider environmental envelope. This would allow data centers to operate their cooling systems more efficiently.”

fig4

Readings show here from the Eaton Powerware floor mounted PDU could part of a PUE2 report.

Even the Uptime Institute, which until recently has been fairly insular, provided feedback to The Green Grid as a member of the task force. According to Matt Stansberry, Uptime’s director of Content and Publications, "PUE has high adoption rates but can lead to inaccurate reporting, (and) misrepresentation. Further adoption of carbon and water usage reporting metrics (WUE, CUE) will be important to the data center industry."

Stansberry shared results of an Uptime survey released on May 12th that seem to help explain the why a substantial portion (if not a majority) of data centers operate at or near a PUE of 2.0, a finding first made by the EPA’s Energy Star program as it gathered preliminary data for its data center program, Stansberry added. He noted that much work remains to be done beyond revising the PUE standard. For instance:

• Fifty-five percent of data centers are still at or below 70°F 

• Only 12 percent are managing temperature as efficiently as possible (by measuring and controlling temperature at the air intake to the IT Equipment)

                                                                                                       • Forty percent are still using return air temperature (at the CRAC/CRAH)

The PUE metric may not be perfect and is still evolving. Not all organizations subscribe to it, and some simply misuse it. Nonetheless, it has caused a much greater awareness of energy usage and efficiency in the data center industry. The advent of PUE Version 2, as defined and now agreed to by the industry, as well as government entities of Task Force, will hopefully increase broad acceptance and use. Perhaps the new document will also reduce seeming exaggerated efficiency claims by some and produce a more accurate and realistic picture of the data center environment.

fig5

Rack-mounted PDUs, such as this product from PDI, can provide values for a PUE3.

Already The Green Grid and others are taking it on themselves to create yet another new framework for evaluating data centers. Stansberry concluded,  “Mark Acton, network director in EMEA, has been working with Harqs Singh of The Green Grid to provide an industry test-bed for the organization’s new Data Center Maturity Model (DCMM). The DCMM is a new framework for data center owners and operators to self-assess the environmental friendliness (maturity) of their data centers. Uptime Institute organized volunteers from the Network to help test this new industry benchmark. The volunteers will plot where their data centers fall on the Maturity Model and work with Uptime Institute and Green Grid to point out areas for improvement. Anonymous feedback from volunteers is planned and will be published as a joint Uptime-Green Grid industry document.”

The PUE metric may not be perfect and is still evolving. Not all organizations subscribe to it, and some simply misuse it. Nonetheless, it has caused a much greater awareness of energy usage and efficiency in the data center industry.  The advent of PUE Version 2 as defined and now agreed to by the industry, as well as government entities of Data Center Energy Efficiency Task Force, will hopefully increase broad scale acceptance and use. Perhaps it will also reduce seeming exaggerated efficiency claims by some and produce a more accurate and realistic picture of the data center infrastructure’s energy efficiency.