CAJ | 학술논문

Peer-to-Peer (P2P) systems have attracted much attention in academic community and industry circles due to their promising applications in various domains. This paper presents the authors' research efforts on introducing complex query capabilities in a P2P environment consisting of numerous peers with large volume of data. An underlying hybrid P2P computing platform, named BestPeer is described first. The connection among peers within BestPeer is selfconfigurable through maintaining the nearest neighbor of peers, and the agent techniques employed in the system ensure its capability of providing sophisticated services. The designs of three P2P data management systems which are all based on BestPeer are described in detail. They provide support for information retrieval, query processing and Web services respectively. Advantages and limitations are discussed, while ongoing work is presented. Current systems can provide basic functions for keyword-based search, SQL-like query processing, and Web services querying and discovery. Some further topics on providing fully-fledged data management functionalities for P2P distributed computing systems with security guarantee are also discussed.