Learning model for efficient query routing in P2P information retrieval systems

Peer-to-peer (P2P) systems have recently attracted a lot of attention since they allows the implementation of large distributed storage space of information. Peer-to-Peer systems (P2P) represent the major technology of access upon various resources on Internet. However, the main problem in P2P systems is how to efficiently and effectively locate resources to answer a specific query (Query routing). In this respect, this paper proposes a semantic model in which a query can be routed to appropriate peers instead of broadcasting or using random selection. The main objective of this model is to achieve better results in a non-supervised task through the incorporation of usage data obtained from past search queries. This model allows us to discover the preference of users when downloading documents from some peers. The terms of use in past queries can provide a better choice of features queries. Hence, for each peer, our model learns from past queries to discover correlations between queries terms and peers. We implemented the proposed model and compared its routing search cost and retrieval effectiveness in terms of both recall and messages traffic with a broadcasting scheme (without learning). To test it, we defined a layer of routing on the PeerSim simulator. Experimental results show that our model performs better than other non-semantic query routing models with respect to accuracy. Through extensive experiments, we show that our proposed model improves the retrieval effectiveness as well as the search cost.