Peer Clustering and Firework Query Model in the Peer›to›Peer Network
Abstract
Clustering technique is used in database and information retrieval system for organizing data and improving retrieval efficiency. We surmise such functionality is valuable to a Peer-to-Peer (P2P) distributed environment. In this paper, we introduce the concept of peer clustering at the level of overlaying network topology, thus, data inside the P2P network are organized in a fashion similar to a Yellow Pages. Moreover, the usability of these systems depends on effective techniques to retrieve information, however, the current strategies used in existing P2P systems are inefficient. To avoid query messages flooding and saving resources in handling irrelevant queries, we propose a content-based query routing strategy, the Firework Query Model, to improve existing retrieval methods. In contrast to broadcasting the query message, our query message is routed intelligently according to its content. Once it reaches the target cluster, the query message is broadcasted to all peers inside the cluster much like an exploding firework. We design and implement a DIStributed COntent-based Visual Information Retrieval (DISCOVIR) system with content-based query functionality and improved query efficiency. We demonstrate its scalability and efficiency through simulation.
Authors: Cheuk Hang Ng, Ka Cheung Sia, Chi Hang Chan, Irwin King