SDC: A Distributed Clustering Protocol

Network clustering is an important technique used in many large-scale distributedsystems. Given good design and implementation, network clustering cansignificantly enhance the system's scalability and efficiency. However, it is verychallenging to design a good clustering protocol for networks that scale fast andchange continuously. In this paper, we propose a distributed network clusteringprotocol SDC targeting large-scale decentralized systems. In SDC, clusters aredynamically formed and adjusted based on SCM, a practical clustering accuracymeasure. Based on SCM, each node can join or leave a cluster such that theclustering accuracy of the whole network can be improved. One big advantage ofSDC is it can recover accurate clusters from node dynamics with very lowmessage overhead. Through extensive simulations, we conclude that SDC isable to discover good quality clusters very efficiently.