A Distributed, Searchable Cache for Tightly Coupled Networks

We describe the design and implementation of a distributed, searchable web proxy cache designed to reduce and eliminate web traffic for itermittently connected environments with low bandwidth and/or high latency. We characterize our system using a proxy cache trace. We have found that searching over the cache using queries from extracted the cache tends to yield higher results than searching using unrelated queries, that transcoding or eliminating images and other media altogether may be effective not only in reducing bandwidth if done by the ISP, but also in improving search results if done before the files are stored in the repository, and that an adaptive caching scheme may do better than LRU, but more experiments are needed to evaluate this.