Achieving Application Performance on the Computational Grid

EECS Joint Colloquium Distinguished Lecture Series

pic of Fran Berman

Prof. Francine Berman
Department of Computer Science and Engineering
University of California, San Diego

Wednesday, March 8, 2000
Hewlett Packard Auditorium, 306 Soda Hall
4:00-5:00 p.m.

Abstract:

"Computational Grids", comprised of ensembles of distributed computation sites, remote instruments, data archives, networks, and other resources, are becoming an increasingly prevalent and critical platform for high- performance computing. However application performance on a Grid is often difficult to achieve due to resource heterogeneity and variability in the deliverable performance of shared Grid components.

A fundamental paradigm for achieving application performance on the Grid is adaptivity. The goal of the AppleS (Application-Level Scheduler) project is to develop a methodology and software for applying adaptivity to the scheduling and execution of Grid applications. AppLeS applications demonstrate that adaptivity is an effective and powerful paradigm for achieving application performance in the most difficult of environments -- where multiple users share resources, resources exhibit distinct and dynamic performance characteristics, and computation, communication and data must be coordinated over local and wide-area networks. In this talk, we describe the methodology, prototype software, and new research developed as part of the AppLeS project.

---------

Francine Berman is Professor of Computer Science and Engineering at U. C. San Diego, Senior Fellow at the San Diego Supercomputer Center, Fellow of the ACM, and founder of the Parallel Computation Laboratory at UCSD. Her research interests over the last two decades have focused on parallel and distributed computation, and in particular the areas of programming environments, tools, and models that support high-performance computing.

Dr. Berman's current research focuses on the development of performance-oriented software environments for networked heterogeneous distributed resources, also known as Computational Grids or Metasystems. Dr. Berman currently co-leads the AppLeS project with Rich Wolski , focused on the development of dynamic application scheduling methods and performance technology for multi-user distributed environments. The AppLeS group collaborates widely and is developing leading edge application scheduling methodology and software for a large number of distributed high-performance and computational science applications.

Dr. Berman has served on numerous editorial boards, program and conference committees in the areas of parallel computing and Grid Computing. She is Chair of the Applications and Tools Requirements Working Group of the Grid Forum and is involved in the Metasystems Thrust of the National Partnership for Advanced Computational Infrastructure (NPACI) where she provides leadership in application scheduling issues.

231cory@EECS.Berkeley.EDU