A distributed ranking strategy in peer-to-peer based information retrieval systems

The state of the art of P2P-based information retrieval is still at its infant stage and confronted with many challenges. One of the most urgent problems is how to rank the results retrieved from different peers. In this paper, we propose a fully distributed strategy to address this challenge. First, we identify the underlying diathesis resulting in the problem. Then we propose a fully distributed strategy to rank and merge results retrieved from different sites, including fully distributed statistics management and distributed ranking processing. Finally, we conduct an extensive experimental study to evaluate the strategy and the results verify the effectiveness of our solution.