U.Va. Computer Science Professor Receives $1 Million Google Research Award to Study Energy Efficiency in Data Centers

February 10, 2010

February 10, 2010 — Sudhanva Gurumurthi, an assistant professor in the University of Virginia's Computer Science Department, is part of a research team that recently received a two-year, $1 million Google Focused Research Award to investigate creating more energy-efficient Internet data centers. The company may award an additional $500,000 for a third year after reviewing the group's progress.

Google Inc. announced on Feb. 2 the distribution of $5.7 million to 12 university projects addressing topics of interest to Google and the greater computer research community. The awards will cover research in four areas: machine learning; the use of mobile phones as data-collection devices for public health and environment monitoring; energy efficiency in computing; and privacy.

Gurumurthi's group received the largest of the awards in the area of energy efficiency in large-scale computing.

Internet-based companies would stand to benefit from more energy-efficient data centers because they currently use large amounts of electricity to store and manage increasing amounts of data.

Without improvements in energy efficiency, servers and data centers in the United States are expected to consume more than 100 billion kilowatt-hours, at a cost of $7.4 billion annually, by 2011, according to a 2007 report by the Environmental Protection Agency.

"Given how much energy data centers use, designing them to be energy-efficient is a globally conscientious pursuit," Gurumurthi said. "This will reduce these data centers' demand for electricity and their carbon footprint."

Gurumurthi and colleagues at Rutgers University, the University of California, Santa Barbara, and the University of Michigan will conduct research aimed at redesigning the architecture of servers in data centers to achieve improved energy efficiency.

Because servers use nearly an equal amount of energy when operating at peak levels as they do when idle, the researchers will be exploring how to redesign the machines so they consume energy proportional to the work they are doing.

At U.Va.'s School of Engineering and Applied Science, Gurumurthi will focus on developing energy-proportional memory and storage systems for such servers. One of the key aspects of developing such systems will be the use of solid-state memory technologies such as flash and phase-change memory.

Gurumurthi is a co-investigator on the team, which is led by Ricardo Bianchini, professor of computer science in the Rutgers School of Arts and Sciences, and Fred Chong, professor of computer science at University of California, Santa Barbara. Thomas Wenisch, assistant professor of computer science and engineering at the University of Michigan, is also a co-investigator on the team.

— By Zak Richards