Abstract

Cooperating proxy caches are groups of HTTP proxy servers that share cached objects. This paper models the speedup in average user response time for proxy cooperation and derives expressions for the upper bound, a mesh organization, and a hierarchical organization. The expressions are parameterized using empirically determined hit rates and HTTP response times. Models account for overhead of different discovery mechanisms (ICP or metadata directory) and clarify the trade-off between discovery overhead and effectiveness. We find proxy cooperation to be only marginally viable when the sole criterion is average response time. By offering a choice of Web sites, however, proxy cooperation can potentially reduce the variability in response time, the number of pathologically long delays, and congestion near busy Web servers. Keywords--- Web, proxy, cache, hierarchy, mesh, cooperative proxy caching I. INTRODUCTION RADITIONALLY a cache is defined as a fast, temporary store for commonly used ...

Citations