Decreasing System Load by Caching in Structured P2P Systems

Peer to peer systems are distributed systems, decreasing system load is important for improving the scalability of systems. Caching is always used to achieve load balance in structured p2p systems currently, but none of the current caching algorithms take system load into consideration. We proposes a caching model for decreasing system load in structured peer to peer systems, describes how peer uses a passive file requested statistical algorithm to evaluate the reducible query load caused by caching a file to a neighbor, and discusses how to compute the updating overhead under different updating algorithms. Each peer determine whether caching a file to a neighbor is worthwhile is based on the relations between the query load reduced and the updating overhead caused by the caching. Simulation results indicate the caching model can decrease system load effectively.