On the latency of BFS based interval cooperative Web caching

The steady growth in Web applications is attracting an increasing the number of Web clients. This results in an increase in Internet traffic with the obvious consequence of growing delay in obtaining a Web document, as everybody experiences frequently. Caching is one of the solutions proposed to reduce this latency: a proxy (also called Web cache) is situated in the path between the client and the original server that acts as a cache to serve the requests of the client retrieving the requested document from its memory without relying to the original server. Web caching alone is not a panacea: if a high number of requests are directed to a single proxy, it becomes a hot spot in the network. So, it is the use of set of proxies that cooperate in order to satisfy the clients requests is needed. In [A. Zabian et al.] we proposed a mechanism for the communication between a set of Web caches reducing the time needed to localize the requested document in the system. The proposal is based on BFS (breadth first search) algorithm. We present the results of simulation experiments performed to assess the average latency experienced by hypothetical users of our system.