Towards Adaptive Probabilistic Search in Unstructured P2P Systems

So far, query routing strategies of unstructured P2P system are described qualitatively or conducted expensively. In this paper, we propose an adaptive query routing method by using quantitative information in the form of probabilistic knowledge for the purpose of (1) maximizing the likelihood of locating desired resource, and (2) using feedback from previous user queries to update the probabilistic information for guiding future ones. To achieve the goal, two kinds of probabilistic information are considered: information about overlap between topics and coverage and completeness of each peer. A declarative formalism for specifying the two kinds of probabilistic information is described, and then the algorithms for using and maintaining such information are presented. Finally, a preliminary experiment is conducted to evaluate the efficiency and effectiveness of our proposed approach.

[1]  Alon Y. Halevy,et al.  Using Probabilistic Information in Data Integration , 1997, VLDB.

[2]  Dimitrios Tsoumakos,et al.  A Comparison of Peer-to-Peer Search Methods , 2003, WebDB.

[3]  Hector Garcia-Molina,et al.  Improving search in peer-to-peer networks , 2002, Proceedings 22nd International Conference on Distributed Computing Systems.

[4]  Christopher R. Palmer,et al.  Generating network topologies that obey power laws , 2000, Globecom '00 - IEEE. Global Telecommunications Conference. Conference Record (Cat. No.00CH37137).

[5]  Chris Buckley,et al.  Implementation of the SMART Information Retrieval System , 1985 .

[6]  Hector Garcia-Molina,et al.  Efficient search in peer to peer networks , 2004 .

[7]  Hector Garcia-Molina,et al.  Routing indices for peer-to-peer systems , 2002, Proceedings 22nd International Conference on Distributed Computing Systems.