In this paper we compare the performance of different caching architectures. We derive analytical models to evaluate hierarchical and distributed caching in terms of client's perceived latency, bandwidth usage, load on the caches, and disk space. Hierarchical caching has short connection and low bandwidth usage, however, it requires high performance caches at the top-levels of the hierarchy. On the other hand, distributed caching has short transmission times and does not require intermediate cache levels. With distributed caching the bandwidth usage in the network is higher, however, the traffic is better distributed, with more bandwidth in the less-congested lower network levels. Additionally, we study a hybrid scheme where a certain number of caches cooperate at every level of the caching hierarchy using distributed caching. We find that a "well configured" hybrid scheme can combine the advantages of both hierarchical and distributed caching, reducing the retrieval latency, the bandwidth usage, and the load in the caches. Depending on the hybrid caching architecture, the current load in the caches/network, and the document size, there is an optimal number of caches that should cooperate at each network level to minimize the overall retrieval latency. Based on analytical and experimental results, we propose small modifications of the existing cache-sharing schemes to better determine the degree of cooperation at every caching level.
Analysis of Web caching architectures: hierarchical and distributed caching
IEEE/ACM Transactions on Networking, Vol. 9, N°4, August 2001
© 2001 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
PERMALINK : https://www.eurecom.fr/publication/366