Continually Answering Constraint k - NN Queries in Unstructured P2P Systems
Continually Answering Constraint $\pmb k$-{\it\bfseries NN} Queries in Unstructured P2P Systems

We consider the problem of efficiently computing distributed geographical k-NN queries in an unstructured peer-to-peer (P2P) system, in which each peer is managed by an individual organization and can only communicate with its logical neighboring peers. Such queries are based on local filter query statistics, and require as less communication cost as possible, which makes it more difficult than the existing distributed k-NN queries. Especially, we hope to reduce candidate peers and degrade communication cost. In this paper, we propose an efficient pruning technique to minimize the number of candidate peers to be processed to answer the k-NN queries. Our approach is especially suitable for continuous k-NN queries when updating peers, including changing ranges of peers, dynamically leaving or joining peers, and updating data in a peer. In addition, simulation results show that the proposed approach outperforms the existing Minimum Bounding Rectangle (MBR)-based query approaches, especially for continuous queries. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users. Supported by the Program for New Century Excellent Talents in Universities (Grant No. NCET-06-0290), the National Natural Science Foundation of China (Grant Nos. 60503036, and 60773221), the National High-Tech Development 863 Program of China (Grant No. 2006AA09Z139), and the Fok Ying Tong Education Foundation Award (Grant No. 104027).