Latest from Building Automation

Modular Building Institute
Hpac 388 1 Pueinformationscreen200
Hpac 388 1 Pueinformationscreen200
Hpac 388 1 Pueinformationscreen200
Hpac 388 1 Pueinformationscreen200
Hpac 388 1 Pueinformationscreen200

PUE an Effective Metric for Data Centers

Nov. 1, 2011
Monitoring power usage effectiveness paints a powerful picture

As the importance of information technology has grown in the world of business, data centers have become the centerpieces at many companies. The demands placed on these facilities—and the HVAC systems that serve them—grow constantly and will likely continue to do so into the foreseeable future.

The Green Grid (www.thegreengrid.org) is an organization dedicated to helping the information technology industry improve energy efficiency in data centers. The non-profit, open industry consortium of end-users, policy-makers, technology providers, facility architects, and utility companies collaborates to improve the resource efficiency of data centers and business computing ecosystems.

The organization was formed in 2007, and on April 5, 2010, it announced a global consortium agreement on the guiding principles of data center energy efficiency metrics. Their goal is to create a set of globally accepted metrics for data center efficiency, and this effort requires a unity of communication.

Green Grid and its consortium organizations have established power usage effectiveness (PUE) as the industry’s preferred energy efficiency metric. The group will leverage guiding principles to help drive a universal understanding of energy efficiency metrics and generate dialogue to improve those metrics.

The fundamental first step is to prioritize energy saving opportunities by gaining an understanding of data center energy consumption. It also requires an analysis of demand-side systems compared to supply-side systems.

Demand-side includes processors, server power supplies, other server components, storage and communication. These components usually account for about half of total consumption. The other half encompasses supply-side systems such as the uniterruputible power supply, power distribution, cooling, lighting, and building switchgear. The supply-side equipment is not an independent consumer of power; its power consumption depends on the power demand.

Recognizing that all data centers are different and savings potential will vary by facility, data center managers still will appreciate recognizing this measurement.

Reductions in demand-side energy usage cascade through the supply side. A 2008 model1 showed that every watt of electricity that can be saved on the processor level (demand-side) creates 2.84 w of savings for the facility.

In 2009 HPAC Engineering published an article2 that was the collaborative effort of professionals from the Lawrence Berkeley National Laboratory. The article indicated calculating PUE requires knowing site energy usage, source energy usage, and total power usage.

Enter health insurer Medical Mutual of Ohio's Beachwood, Ohio data center. Medical Mutual is the oldest (1934) and largest health insurer in the State of Ohio, and it prides itself on providing security during a disaster to its clients. Since 1977, the company has provided customer service, data storage, billing help-desk services, call center staff, and Website services.

In 2009, Medical Mutual approved an integration of multiple HVAC systems within the Beachwood data center. At the time, one side of the facility was being controlled by a 14-year-old building-automation and control system from one manufacturer, while the other side was being controlled by a 2-year-old direct digital control system from another. The separate monitoring of two independent building control systems was becoming frustrating to energy-management specialist Ed Haywood, so Medical Mutual's senior management decided to eliminate the multiple front-end HMI terminals and perform a single integration that would be hosted on the company's ethernet. Enterprise HVAC Service & Control performed this integration successfully.

Afterwards, Medical Mutual's IT personnel saw the results of these two merged building control systems with dissimilar protocols, and asked Enterprise if the same system could be used to calculate and report hourly, daily, and monthly PUE.

Programmers for Enterprise performed this calculation by dividing the amount of power entering the data center by the power used to run computer infrastructure (uniterruptible power supply and data-center cooling equipment loads). This data is shown as a display of current PUE on a daily, weekly, monthly, and yearly basis.

Values are captured to logs every 15 minutes. Daily values reset at midnight every night, weekly values reset at midnight Sunday, monthly values reset at midnight on the last day of the month, and yearly values reset at midnight on New Year's Eve. The figures that accompany this article show a typical screen of data center calculated.

The overall goal of Mark Sweeney, Medical Mutual's manager of IT, is to have the data center earn the U.S. Environmental protection Agency's Energy Star status. He sees the constant awareness of PUE as a big step towards achieving this goal. He also knows that energy efficiency translates into lower premiums for Medical Mutual’s clients.

References
1. Judge, J., Pouchet, J., Ekbote, A., & Dixit, S. (2008, November) Reducing data center energy consumption. ASHRAE Journal, p. 14-26
2. Mathew, P., Greenberg, S., Ganguly, S., Sartor, D., & Tschudi, W. (2009, May) How does your data center measure up? HPAC Engineering, p. 16-21

Richard "Dick" Starr is president and CEO of Enterprise HVAC Service & Control, a design/build/maintenance contractor located in the Cleveland, Ohio area. He is a member of the national boards of both the Mechanical Service Contractors of America and the Mechanical Contractors Association of America. He can be reached at [email protected]. Enterprise HVAC Service & Control is a certified Tridium system integrator.