Improved P2P content discovery by exploiting user social patterns

One of the most challenging issues in Unstructured Peer-to-Peer (P2P) network is to efficiently locate information resources. In this paper we propose a novel concept in which peers exploit social relations in order to improve the search success rate. In social networks, people can directly contact acquaintances that have knowledge about the resources they are looking for. However, peers lack these capabilities in a P2P network. We can find different searching mechanism such as flooding, which drastically increases the communication overhead, or random walk (RW) that reduces the message overhead, but since it is a blind sequential search it may take a long time to route a query. In front of this we have proposed a two-hop algorithm that incorporates the social behaviors of peers and processes queries more efficiently [2]. However, there is a strict limitation that the recommended nodes must always have the query resources. In this paper, we propose a one-hop algorithm that uses social behavior patterns. In the proposed algorithm peers establish friendship relations and learn from past experiences to recommend suitable peers that will route queries in an efficient manner. The simulation results show that the proposed one-hop algorithm provides better average success rates compared to both the random walk algorithm and the two-hop algorithm by reducing the search to only one logical hop. In addition, our proposal minimizes the required network memory space by limiting the query record, useful friends and resources at each node in the one-hop algorithm.