Bio: Amin Vahdat joined the Jacobs School faculty in the 2003-04 academic year. He completed his undergraduate and graduate work at UC Berkeley, where he received his Ph.D. in Computer Science in 1998. After Berkeley, Vahdat worked briefly as a research associate at the University of Washington, before joining the faculty of Duke University, first as an Assistant Professor then as an Associate Professor in 2003. Vahdat is a recipient of the Alfred P. Sloan Fellowship (2003), an IBM Faculty Partnership Award (2002-03) and an NSF CAREER Award (2000). He has published extensively, with more than 50 publications appearing in leading journals and conferences.  

Research: Professor Amin Vahdat's research focuses on building scalable, high-performance, highly available Internet systems. As the Internet has evolved and grown exponentially, it has become increasingly dynamic and unpredictable, with networks constantly being added and removed and traffic always changing. Vahdat is investigating how systems can self-organize and dynamically react to the changing characteristics of the network to achieve peak performance and availability. He has written on flexibly maintaining data accuracy in the face of faulty and congestion-prone networks, and accurate emulation of wide-area networks. Vahdat has spent the past ten years working on infrastructure and abstractions needed to support wide-area applications that are geographically distributed, highly available, incrementally scalable, and dynamically reconfiguring. In his work on operating systems for mobile computing, Vahdat has investigated new approaches to designing computer architectures and software with power as a primary measure of performance. With his colleagues, he has designed a variant of Linux, called ECOSystem, that uses energy as the common currency for allocating all system resources to competing applications. Overall, the goal of Vahdat's work is to maximize performance and availability of next-generation and mission-critical applications--where revenue, lives, or sensitive information are on the line--with minimal manual configuration.