Abstract
To meet the ever growing demands of Internet services and cloud computing, the data center industry is experiencing rapid expansion. Data centers consume billions of KWh electricity every year, and each can take hundreds of million dollars to build. Most data centers are conservatively provisioned and operated to achieve high reliability, resulting in wasted resources and high cost. In this talk, we take a holistic view of data centers and discuss opportunities and our experiences. We have designed and deployed sensors to monitor heat distribution in server rooms and software-based services to estimate server power consumption. We build models that bridge the cyber dynamics of computing and physical dynamics of the facility. The findings are used to advance the way equipment are provisioned, loads are distributed, and systems are operated. Finally, I will discuss a few challenges for scaling into the future, with alternative energy sources and alternative data center architecture.
Biography
Dr. Jie Liu is a Principal Researcher at Microsoft Research, Redmond, WA, and the manager of its Sensing and Energy Research Group. He is an ACM Distinguished Scientist. His research interests root in understanding and managing the physical properties of computing. Examples include timing, location, energy, and the awareness of, and impact on, the physical world. He has published broadly in areas like sensor networks, embedded systems, ubiquitous computing, and data center energy management. Dr. Liu is an Associate editor of ACM Trans. on Sensor Networks, the current steering committee chair of ACM SenSys, and the general chair of CPS Week 2015. He chaired a number of top tier conferences and served on review boards and panels. Among other recognitions, he received Best Paper Awards at MobiSys’14, SenSys’12, and RTAS’10, the Leon Chua Award from UC Berkeley in 2001, Technology Advance Award from (Xerox) PARC in 2003, and a Gold Star Award from Microsoft in 2008. Dr. Liu received his Ph.D. degree in Electrical Engineering and Computer Sciences from UC Berkeley in 2001, and his Master and Bachelor degrees from Automation, Tsinghua University, Beijing, China.