SeCond: a system for epidemic peer-to-peer content distribution

Abstract

We propose an efficient cooperative content distribution protocol in which the cooperation among participants is based on a peer-to-peer (P2P) paradigm. Our main contribution to P2P content distribution is the use of an epidemic communication approach. Since epidemic algorithms are easy to use, robust and adaptive to dynamic conditions, they have found several application areas in distributed systems such as failure detection, data aggregation and database replication. However, they have not been used in cooperative content distribution before. In addition to the use of epidemic algorithms for state exchange among peers, we propose some methods in order to increase utilization of system resources during distribution of the files. We demonstrate effectiveness and scalability of the protocol through our simulation model