Exploring the bounds of web latency reduction from caching and prefetching

TLDR

Prefetching and caching are widely used I/O techniques to reduce latency, and many researchers advocate their use for the Web. The authors derive theoretical bounds on the performance gains from these techniques and quantify them using Web proxy traces from Digital Equipment Corporation. Local proxy caching can reduce latency by up to 26 %, prefetching by up to 57 %, and a combined caching‑prefetching proxy by up to 60 %; the improvement depends on how far in advance objects are prefetched, but both techniques are limited by rapid web changes, with caching offering moderate assistance and prefetching more than twice as effective yet still constrained.

Abstract

Prefetching and caching are techniques commonly used in I/O systems to reduce latency. Many researchers have advocated the use of caching and prefetching to reduce latency in the Web. We derive several bounds on the performance improvements seen from these techniques, and then use traces of Web proxy activity taken at Digital Equipment Corporation to quantify these bounds. We found that for these traces, local proxy caching could reduce latency by at best 26%, prefetching could reduce latency by at best 57%, and a combined caching and prefetching proxy could provide at best a 60% latency reduction. Furthermore, we found that how far in advance a prefetching algorithmwas able to prefetch an object was a significant factor in its ability to reduce latency. We note that the latency reduction from caching is significantly limited by the rapid changes of objects in the Web. We conclude that for the workload studied caching offers moderate assistance in reducing latency. Prefetching can offer more than twice the improvement of caching but is still limited in its ability to reduce latency.