CATS Seminar
Capital Area Theory Seminar
http://www.umiacs.umd.edu/users/liberato/cats.html
New Algorithms and Metrics for Scheduling Job Streams
Michael A. Bender
SUNY at Stony Brook
Wednesday, March 3, 1999, 2 p.m.
AVW 2120
Abstract
There are two important objectives when designing a scheduler for a webserver, disk, or routing switch: maximizing throughput and minimizing response time. We study the relationship between these two objectives for the problem of scheduling a stream of jobs on a single processor. We introduce a new metric, maxstretch, that naturally balances throughput and response time. Then we describe algorithms to optimize this metric and show simulation results.
This is joint work with Soumen Chakrabarti and Muthu.
Biography: Michael Bender received his B.A. in Applied Mathematics from Harvard University in 1992 and obtained a D.E.A. in Computer Science from the Ecole Normale Superieure de Lyon, France in 1993. He completed a Ph.D. on Scheduling Algorithms and Asynchronous Parallel Computation from Harvard University in 1998. Michael Bender is currently on the faculty of SUNY at Stony Brook.
Contact: Vincenzo Liberatore (liberato@umiacs.umd.edu)
The Capital Area Theory Symposia is an NSF sponsored series of symposia in theoretical computer science bringing computer scientists from around the world to the Capital area. The Symposia are given at the University of Maryland in cooperation with the Computer Science Department and UMIACS. NSF support under grant CCR-9732907 is gratefully acknowledged