scope:

Introduction

The Green Grid Association (TGG) works to improve the resource efficiency of information technology (IT) and data centers throughout theworld. In 2007,TGG published two self-help data center energy efficiency metrics for end users: power usage effectiveness (PUE™) and its inverse, data center infrastructure efficiency (DCiE™). Since the original publications, PUE has surpassed DCiE in industry adoption. In fact, PUE is currently viewed as the industry-preferred metric for measuring infrastructure energy efficiency in data centers, for either guiding new facility design or monitoring existing facility operations management.

In 2009,TGG, in collaboration with organizations around the globe, began driving toward a set of metrics and indices that can be formally adopted by all participating organizations to improve worldwide data center energy efficiency.As a result, global agreement has been reached that affirms PUE as the industry's preferred DCiE. The following are the two task forces and supporting organizations that have agreed on the PUE metric, measurement methodology, and reporting convention (TGG 2009a):

• Data Center Metrics Coordination Task Force (U.S. Regional Task Force)

• 7×24 Exchange

• ASHRAE TC 9.9

• TGG

• Silicon Valley Leadership Group

• Uptime Institute

• U.S. Department of Energy Save Energy Now and Federal Energy Management Programs

• U.S. Environmental Protection Agency's ENERGY STAR^® Program

• U.S. Green Building Council

• Global Harmonization of Data Center Efficiency Metrics Task Force (Global Task Force)

• European Commission: Joint Research Center Data Centers Code of Conduct

• TGG

• Green IT Promotion Council, Japan

• Ministry of Economy, Trade, and Industry, Japan

• U.S. Department of Energy Save Energy Now and Federal Energy Management Programs

• U.S. Environmental Protection Agency's ENERGY STAR Program

PUEis an excellent metric for understanding howwell a data center is delivering energy to its IT equipment. The metric is best applied when looking at trends in an individual facility over time and when measuring the effects of different design and operational decisions within a specific facility. For example, improvements to the efficiency of the mechanical or electrical infrastructure will result in a lower PUE value.

No two data centers have exactly the same operating characteristics, such as infrastructure or IT loads. Therefore, the comparison of two data centers based on public reports of their PUE results was not initially recommended because many attributes of data center design, engineering, implementation, and operations affect PUE. However, since the original publications, the industry has globally adopted the PUE metric and begun to compare PUEs among data centers, raising further questions about how to interpret individual results, how to compare different results for the same data center, and howto compare results across different data centers.As there are various ways to calculate PUE, stakeholders in the industry have expressed concerns around the consistency and repeatability of publicly-reported measurements. In response, TGG has published a set of rules and guidelines regarding the process that organizations should follow when making public claims regarding their data centers' PUEmeasurements and has provided a freePUEreporting tool(TGG2009c). Chapter 6 of this book outlines thePUEreporting requirements and also includes the definition of standard nomenclature that will enable individual claimants to communicate key information about their measurements.

Standardized PUE nomenclature, proper and transparent public reporting guidelines, and the availability of key information about reported results will enhance both the credibility and usefulness of the PUE metric. To further enable equitable comparison of PUE results among data centers, additional attributes such as age, geographic location, capacity loading, size of facility, infrastructure resiliency, and the like, should be taken into consideration.

Provided the measurement guidelines, reporting guidelines, and the additional data attributes (as mentioned above) are obtained, TGG andASHRAE TC 9.9 believe PUE can be used to compare DCiE.

The PUE metric is associated with the data center infrastructure. PUE is not a data center productivity metric, nor is it a stand-alone, comprehensive efficiency metric. PUE measures the relationship between the total facility energy consumed and the IT equipment energy consumed. When viewed in the proper context, PUE provides strong guidance for and useful insight into the design of efficient power and cooling infrastructure architectures, the deployment of equipment within those architectures, and the day-to-day operation of that equipment.

Changes in PUE are most meaningful when they are seen as the data center's response to changes in infrastructure equipment or infrastructure operations. Studies investigating the effect of PUE on changes in infrastructure equipment or operations should ensure that any changes occurring in the IT load over the study's period of time are properly accounted.

This book provides guidelines for determining the PUE of a dedicated data center and that of a mixed-use facility or specialized facility, along with results, scalability analyses, and case studies. It also outlines the approach to reuse waste heat. Issues such as availability of equipment and the productivity of the data center require different metrics and different analyses.

Data centers are complex systems. Frequently, changing certain aspects of the facility (e.g., commissioning a new class of server) can produce apparent changes in PUE results for another aspect of the facility. TGG and ASHRAE TC 9.9 advise data center owners and operators to take the greatest care around issues of availability. If an action or initiative taken to improve PUE has a negative effect on the availability of the IT equipment in the data center, owners and operators should reviewthe potential impact on the data center and determine whether or not to proceed any further with that action or initiative.

In addition, PUE does not provide any guidance or insight into the operation or productivity of IT equipment. TGG is currently investigating additional metrics and approaches to provide guidance and insight into these areas (TGG 2009b). It is also possible, even likely, that changes in the deployment or operation of IT equipment will affect PUE results. For example, organizations implementing virtualization in their data centers may reduce overall IT load but see an increase in PUE. In these instances, the fixed overhead for power distribution and cooling has not changed, but the reduction in IT load delivers a seemingly poorer PUE result. PUE metric users should not become overly concerned when changes in one area affect results in another, but rather consider the factors that contributed to the PUE increase as further opportunities for improvements.

Source energy represents the total amount of raw fuel that is required to operate the data center and includes transmission, delivery, and production losses. A source energy based PUE must be used when energy sources other than electricity are delivered to the data center. TGG and ASHRAE TC 9.9's recommended approach for measuring and calculating PUE by source energy will yield the same output as an allelectric data center calculating a site-based PUE. Please see Section 4.9 Source Energy for more information.

Numerous internal references have been incorporated into this book with the goal of maintaining an executive overview of each metric and key topic, while enabling end users and operators to obtain more detailed information, including implementation data, lessons learned, and best practices. These references also allows those in particular verticals (e.g., government or healthcare) and business areas (e.g., executive management, marketing, sales) to pull the level of information required in a quick and efficient manner. This is intended to enhance the use of this book and of PUE-related metrics.