Speaker: Andrew Myers, Carnegie Mellon University
Date: October 8, 1998
As a growing number of web sites introduce mirrors to increase throughput, the challenge for clients becomes choosing which mirror will offer the best performance when a document is to be retrieved. In this paper we present findings from measuring 9 clients scattered throughout the United States retrieving over 490,000 documents from 45 production web servers which mirror three different sites. We have found a number of interesting and unexpected trends in the data that will aid in the design of protocols for choosing among mirror servers. Though server performance varies widely, we have observed only small, time independent variations in the relative performance of mirror servers. Further, a change in an individual server's transfer time is not a strong indicator that its performance relative to other servers has changed. We have also found that clients wishing to achieve near-optimal performance may only need to consider a small number of servers, rather than all mirrors of a particular site. Finally, we noticed that the choice of document affects the choice of server in a significant fraction of the fetches performed.