Web Cache Prefetching
An increasing amount of information is currently becoming available through World Wide Web servers. Document requests to popular Web servers arrive every few tenths of a millisecond at peak rate. Without caching, the World Wide Web will soon become a victim of its onw success. Experience suggests that Web servers become a inaccessible as soon as they become popular. As a result, popular Web servers appear slow, and users become frustrated with their service. An effective way to imporove Web performance is the sophisticated use of caching and prefetching.
In this project, I will first explore the notion of prefetching and caching a World Wide Web server's documents in its main memory to reduce the disk I/O so as to reduce the latency. Second, I explore the notion of the server pushing the documents to the client side to reduce the latency caused by the network traffic. Of course, the side effect of the incorporation of prefetching and pushing is the increased memory usage and increased netwrok traffic. Using the cooperative server and client developed here in the lab, and using the lab web server log file, I will try to find a good prefetching and pushing algorithm to reduce the server response time and server load at a low cost, i.e. low memory consumption and low traffic increase.
This research is being conducted primarily by Yi-Qiang Huang. For more information, email huangyq@pel.cs.byu.edu.
© 1996, Performance Evaluation Laboratory, Brigham Young University. All rights reserved. Reproduction of all or part of this work is permitted for educational or research use provided that this copyright notice is included in any copy. Send comments to webmaster@pel.cs.byu.edu. This document last modified on Wednesday, 01-Apr-98 14:30:22.