SDC: a distributed clustering protocol for peer-to-peer networks

Authors:
Yan Li;Li Lao;Jun-Hong Cui
Affiliations:
Computer Science & Engineering Dept., University of Connecticut, CT;Computer Science Dept., University of California, Los Angeles, CA;Computer Science & Engineering Dept., University of Connecticut, CT
Venue:
NETWORKING'06 Proceedings of the 5th international IFIP-TC6 conference on Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; Mobile and Wireless Communications Systems
Year:
2006

Citing 4
Cited 0

An Efficient Peer-to-Peer File Sharing Exploiting Hierarchy and Asymmetry

SAINT '03 Proceedings of the 2003 Symposium on Applications and the Internet
Performance criteria for graph clustering and Markov cluster experiments

Performance criteria for graph clustering and Markov cluster experiments
A new cluster algorithm for graphs

A new cluster algorithm for graphs
A Distributed Approach to Node Clustering in Decentralized Peer-to-Peer Networks

IEEE Transactions on Parallel and Distributed Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Network clustering can facilitate data discovery and peer-lookup in peer-to-peer systems. In this paper, we design a distributed network clustering protocol, called SCM-based Distributed Clustering (SDC), for peer-to-peer networks. In this protocol, clustering is dynamically adjusted based on Scaled Coverage Measure (SCM), a practical clustering accuracy measure. By exchanging messages with neighbors, peers can dynamically join or leave a cluster so that the clustering accuracy of the whole network is improved. SDC is a fully distributed protocol which requires only neighbor information, and it can handle node dynamics locally with very small message overhead while keeping good quality of clustering. Through extensive simulations, we demonstrate that SDC can discover good quality clusters very efficiently.