Decision support tool for web cache management

Web caching is the subject of intense research and development since it seems to be very promising area. Web caching means storing copies of frequently used objects (documents) geographically close to users requesting them to reduce network load, servers load and user response times. Cache can be situated in different locations between user and servers with original content. It seems that the most significant improvement can be achieved by using the proxy server - a dedicated web server. Various parameters affect web cache performance: cache size, limitations on document sizes or documents removal policy. Several metrics are used to evaluate this performance: hit rate for example is the ratio of documents obtained by using the cache mechanism comparing to the total number of documents requested. Choosing adequate parameters for interesting traffic patterns to improve cache performance is not a trivial cache management problem. Prototype tool based on multicriteria model analysis and supporting such web cache management is presented. Simple case was examined where small number of sets of cache parameters (variants), miniature traffic representation and only two performance measures are considered.