Benchmarking Web servers is an active research area. Several benchmarks for Web servers have been developed, including WebStone [11] and SPECWeb. There are also studies on the overload behavior of the benchmarks and improvement of the benchmarks [1]. However none of the benchmarks can be easily used to measure proxy performances. We borrowed several ideas from these benchmarks when designing WPB. However, WPB incorporates unique characteristics that make it appropriate to analyze the performance of proxy servers.
There has also been many studies measuring the performance of Squid proxies in real use [9,10]. Our results confirm most of the findings in those studies. Different from those studies, WPB can be used as a tool to pin point the performance problems of a variety of proxy products.
Finally, there have been many studies on characteristics of proxy traces [3,8,5,6,7]. Our research borrows results from those studies in designing WPB.