Abstract
The Web has become a significant source of various types of data, which require large volumes of disk space and new indexing and retrieval methods. To reduce network load and improve user response delays, various traditional proxy-caching schemes have been proposed (Abonamah, Al-Rawi, & Minhaz, 2003; Armon & Levy, 2003; Chankhunthod, Danzig, Neerdaels, Schwartz, & Worrell, 1996; Chu, Rao, & Zhang, 2000; Fan, Cao, Almeida, & Broder, 2000; Francis, Jamin, Jin, Jin, Raz, Shavitt, & Zhang, 2001; Paul & Fei, 2001; Povey & Harrison, 1997; Squid Web Proxy Cache, 2004; Wang, Sen, Adler, & Towsley, 2002). A proxy is a server that sits between the client and the real server. It intercepts all queries sent to the real server to see if it can fulfill them itself. If not, it forwards the query to the real server. A cache is a disk space used to store the documents loaded from the server for future use. A proxy cache is a proxy having a cache. The characteristics of traditional caching techniques are threefold. First, they regard each cached object as having no dividable data, which must be recovered and stored in their entirety. As multimedia objects like videos are usually too large to be cached in their entirety, the traditional caching architectures cannot be efficient for this kind of object. Second, they do not take into account the data size to manage the space storage. Third, they do not consider in their caching-system design the timing constraints that need moving objects.