Our research is in large-scale distributed systems and distributed computing. In particular, we are designing decentralized mechanisms for resource management in large distributed systems and are interested in understanding the associated tradeoffs between performance and scale. Our research is of experimental nature, grounded in deployed, real distributed systems (such as grids, peer-to-peer systems, and online social networks) whose behavioral patterns, we believe, can often be algorithmically exploited to improve performance.

Media highlights: Cheating in online gaming: link; Tango panopticon: link; Commercial clouds for scientific computing: link, link