Popularity-based content replication in peer-to-peer networks

Abstract

Pure peer-to-peer (P2P) networks is widely used, however they broadcast query packets, and cause excessive network traffic in particular. Addressing this issue, we propose a new strategy for content replication which prioritizes popular and attracting contents. Our strategy is based on replication adjustment, being cooperated with index caching, as well as LRU-based content replacement, and a more effective replica placement method than well-known probabilistic ones. We also present some experiment results to show that our strategy is better than other ones in regards to network traffic and storage consumption. © Springer-Verlag Berlin Heidelberg 2006.