International Environmental Modelling and Software Society (iEMSs) 2010 International Congress on Environmental Modelling and Software Modelling for Environment’s Sake, Fifth Biennial Meeting, Ottawa, Canada David A. Swayne, Wanhong Yang, A. A. Voinov, A. Rizzoli, T. Filatova (Eds.) http://www.iemss.org/iemss2010/index.php?n=Main.Proceedings A Taguchi-Based Method for Assessing Data Center Sustainability Sidney L. Pendelberry Computing and Information Sciences Rochester Institute of Technology, Rochester, New York sidney.pendelberry@rit.edu Sophie Ying Chen Su Golisano Institute for Sustainability Rochester Institute of Technology, Rochester, New York sophie.chensu@rit.edu Michael Thurston Golisano Institute for Sustainability Rochester Institute of Technology, Rochester, New York michael.thurston@rit.edu Abstract: This paper presents a methodology for measuring sustainability performance of data center facilities. Treating data center operation as a continuous production process, a Taguchi framework is used to estimate the sustainability impact due to power utilization from the facility by calculating the loss to society as the facilities power utilization becomes less efficient than the Environmental Protection Agency’s guidelines. By tracking data center sustainability performance with continuous production process metrics, it is easy to identify and study performance characteristics, which if out of specification, have broader environmental impacts. This method evaluates the data center’s excess energy consumption as a performance quality control issue, and estimates the cost of poor performance. Average power usage effectiveness (PUE) is a standard data center performance metric. In this case study PUE is tracked continuously to provide real-time operations and controls feedback. Further, PUE is mapped using the Taguchi framework and the U.S. EPA data center benchmark to the societal cost of non-conformance with the performance benchmark performance deviation based on carbon emissions and associated carbon exchange markets. It is expected that the method used in this paper may provide useful information to policy makers, engineers, and decision makers regarding sustainable IT data centers. Keywords: Taguchi; data center; sustainability; PUE; carbon emission 1. INTRODUCTION In 2010, the power consumption of United States data centers was projected to be 95.5 billion kWh (ENERGY, 2007). With data center power consumption at 1.5% of the U.S. total and growing rapidly, “Greener” data centers offer a large opportunity to reduce U.S. power consumption and thus reduce U.S. carbon emissions. This paper asserts that a data center is considered to be “Green” in the United States when the measure of sustainability, Power Usage Effectiveness (PUE) meets or exceeds the specification of the United States Environmental Protection Agency¹s (EPA) best practice benchmark (ENERGY, 2007). PUE is defined as the ratio of the total power consumed by a data center to the power