CS&E Profile: Jon Weissman

Distributed systems, metacomputing, cluster computing, scheduling and resource management, I/O, parallel processing, and operating systems.

Ph.D. 1995, Computer Science, University of Virginia

M.S. 1989, Computer Science, University of Virginia

B.S. 1984, Applied Mathematics, Carnegie-Mellon University

About

Professor Weissman's research interests are broad, including: distributed systems, grid computing, operating systems, scheduling and resource management, storage systems, network systems, parallel processing, health care and medical applications.

He's been honored with a promotion as an IEEE senior member and an ARL Success Story honor for his Virtual Data Grid Project. Weissman has also received the National Science Foundation's CAREER award, a Super Computing award, and a Teaching Medal of Excellence from the University of Virginia.

Weissman has authored or co-authored more than 60 refereed articles, published in journals and elsewhere. He has chaired and served as a committee member on many conferences, and currently serves on the editorial board for the Journal of Parallel and Distributed Computing.

Research

My research is in experimental distributed systems with an emphasis on software systems and middleware. My current focus is in three distinct distributed computing areas: Computational Grids, Scalable Network Services, and Smart Environments. Computational Grids are the seamless integration of geographically dispersed computing resources connected by emerging high-speed wide-area networks. My group is working on middleware to support adaptive scheduling and resource management for high performance parallel applications in clusters and Grids. In Scalable Network Services we are building an infrastructure (called the Service Grid) for the dynamic selection and replication of system services in wide-area networks. Our research concerns policy design and trade-offs for providing efficient and reliable delivery of services to clients. In Smart Environments we are building middleware to support smart, pervasive, personalized, ubiquitous computing in both wired and wireless LAN environments.

The problems that interest me lay on the boundary between applications and systems. I believe applications and systems need to be tightly coupled in order to meet end user objectives such as reliability, high performance, and predictable performance, in distributed systems. To achieve this goal, many of the software systems and middleware I develop are application-aware. In addition, many of our systems are guided by performance models constructed from application-specific information provided to them. My approach is experimental, often combining on-line experiments to demonstrate feasibility and applicability, with simulation studies to demonstrate generality. My ultimate objective is to build usable software systems that amplify the capabilities of real applications running in distributed networks (both wired and wireless).