I am interested in the performance analysis and design of computer systems, particularly distributed systems. I use analytical models to capture the important characteristics of a computer system, and then I prove theorems about these models that allow me to redesign the system to improve its performance. Here "performance" might denote response time, energy use, throughput, capacity, etc. Most of my research involves inventing new analytical techniques in the area of performance analysis, as well as new algorithms for resource allocation. Unlike many theoretical computer scientists, my analysis is based on stochastic (probabilistic) models of computer systems. There is no "adversary" sending us worst-case inputs. By contrast, there is a stream of requests, whose arrival times and service demands come from empirically fitted distributions. These distributions might be correlated, and they often exhibit heavy tailed service demands and high variability. I believe that many conventional wisdoms on which we base computer system designs are not well understood and sometimes false, leading to inferior designs. My research revisits very classic questions in system design. Here are a few examples of commonly-held beliefs that my research challenges: