[an error occurred while processing this directive]
[an error occurred while processing this directive]
|
|
|
Cary Williamson Department of Computer Science University of Calgary Calgary, Alberta |
|
This talk describes a probe-based sampling approach to study the behavioural properties of Web server scheduling strategies, such as Processor Sharing (PS) and Shortest Remaining Processing Time (SRPT). The approach is general purpose, in that it can be used to estimate the mean and variance of the job response time, for arbitrary arrival processes, scheduling policies, and service time distributions.
In our work, we apply the approach to trace-driven simulation of Web server scheduling to compare and contrast the PS and SRPT scheduling policies. We identify two types of unfairness, called endogenous and exogenous unfairness. We quantify each, focusing on the mean and variance of slowdown, conditioned on job size, for a range of system loads. Finally, we confirm recent theoretical results regarding the asymptotic convergence of scheduling policies with respect to slowdown, and illustrate typical performance results for a practical range of job sizes from an empirical Web server workload.
Dr. Carey Williamson joined the Department of Computer Science at the University of Calgary in July 2001 as an iCORE (Informatics Circle of Research Excellence) Professor specializing in "Broadband Wireless Networks, Protocols, Applications, and Performance". His educational background includes a B.Sc.(Honours) degree in Computer Science from the University of Saskatchewan in 1985, and a Ph.D. in Computer Science from Stanford University in 1991. Prior to joining the University of Calgary, he spent 10 years on faculty in the Department of Computer Science at the University of Saskatchewan, where he received a University Master Teacher award in October 2000. Dr. Williamson's research interests include Internet protocols, wireless networks, network traffic measurement, workload characterization, network simulation, and Web server performance.
This talk represents joint work with graduate student Mingwei Gong.
[an error occurred while processing this directive]