Transparent Web caching with minimum response time

Distributed Web caching systems allow better load sharing and more fault tolerance in Web caching systems. Layer 5 switching-based transparent Web caching schemes intercept HTTP requests and redirect them according to their content. Employment of these schemes in distributed Web caching systems provides balanced server workload, reduced response time and improved cache sharing. However, none of the existing schemes attempt to minimize the HTTP request response time.; In this thesis, we propose a Minimum Response Time (MRT) Layer 5 switching-based Web caching scheme for distributed Web caching systems. MRT distinguishes non-cacheable requests from cacheable requests based on HTTP request header. It intelligently redirects cacheable requests to the cache server with the minimum HTTP request response time based on the information about cache server content, cache server workload, Web server workload and network latency. MRT extends ICP to support the communication between cache servers and Layer 5 switches. The MRT heuristic is a solution to optimize the performance of the distributed Web caching.; A number of simulation experiments are conducted under different HTTP request intensities, network latency factors, object expiration time values and number of cooperating cache servers. Simulation results show that MRT outperforms existing switching-based Web caching schemes, namely Content, Workload, RTT and LB_L5 in terms of HTTP request response time.