A load-balancing structured store model on P2P network

This paper proposed a structured p2p model called balance-peer, introduce a method to partition the data rang without any global information on the load to. Our Balance-peer model is used to realize a partition-by-keyword indexing service based on DHT with each node in the system responsible for part of the keywords of the inverted index list. In this paper we use the concept of data partition in parallel databases to realize the load balance in P2P systems. At last this paper has the performance test. The result convinces us that our storage balance strategy is valid.