LSB-Chord: Load balancing in DHT based P2P systems under churn

Peer-to-Peer (P2P) systems receive growing acceptance, especially DHT based P2P systems. DHT based P2P systems have an inherent workload uneven distribution problem. Skewed workload distribution will influence the availability and the performance of the system seriously. In this paper, an algorithm based on local state for load balancing in DHT based P2P systems under churn, LSB-Chord (local state based chord), is proposed. Each node places the duplication of its data objects to its successors, and collects their load state. Receiving query request, the original node works as a coordinator to balancing the load. Simulation results show that: LSB-Chord can avoid data lost effectively; if the load upon data objects follows uniform distribution, LSB-Chord can achieve the similar balancing to CRA; LSB-Chord can achieve better balancing, if the load upon data objects follows Zipf distribution.