Building a scalable bipartite P2P overlay network

Summary form only given. In unstructured peer-to-peer (P2P) systems, the stochastic peer connection and peers' randomly joining and leaving a P2P network without any knowledge about underlying physical topology can cause serious topology mismatching between the P2P overlay network and the physical underlying network. Some existing techniques have been proposed to address topology mismatching problem without shrink search scope. However, these techniques involve considerable amount of overhead, and have other disadvantages, such as slow convergence speed and synchronization requirement. To address the limits of existing solutions, we propose a scalable bipartite overlay (SBO) among peers in Gnutella-like systems or among the super-peers in KaZaA-like systems. SBO employs an efficient strategy to reduce optimization overhead by intelligently distributing optimization tasks in different peers. Our evaluations show that the total traffic and response time of the queries can be significantly reduced by optimized SBO without shrinking the search scope.