Architecture and Implementation of Cluster-based Peer-to-Peer Topology and Its Application in Search

Abstract

Peer to peer networking has received great attention as it enables rich distributed resource sharing. In the past, many structured searching algorithms, such as Chord, CAN, etc., have been proposed to provide efficient resource searching. These algorithms, however, assume a global scale resource sharing and searching which incurs large overhead as the number of peer node increases. In this paper, we propose a stable hierarchical Chord (SH-Chord) searching based on a power-based cluster-based peer-to-peer topology (CBT). In CBT, peer nodes in close proximity are placed in a cluster. SH-chord sets up a Chord ring in each CBT cluster and uses a power-based load balance scheme to insert virtual nodes so that more keys are stored at nodes with higher power index. Key search across SH-chord rings uses a Chord server for inter-ring search routing. Simulation results show that with our hierarchical approach, key search efficiency is improved and Chord ring maintenance overhead is significantly reduced. Also, our power-based load balance approach can effectively reduce the key loss rate and improve peer-to-peer network stability.