Towards a Cluster Based Incentive Mechanism for P2P Networks

Abstract

Free riders in Peer-to-peer (P2P) networks are the nodes only consume services but provide little or nothing. They seriously degrade the fault-tolerance and scalability of the P2P networks. A Cluster-Based Incentive Mechanism (CBIM) is proposed in this paper to prevent free riding problem in P2P networks regardless of their topologies and service diversity. Nodes with asymmetric interests are organized in clusters that consist of service exchange rings. A node in a ring can receive a service from its predecessor by providing a requested service to its successor. Free riders can not complete their requested services since a ring will collapse once free riding is detected. We firstly identify five design requirements, namely, adaptability, service diversity, reward and penalty, altruism and performance. Second, we describe the cluster formation process and a graph theory based ring identification algorithm. Finally, we describe our coarse-grained probability-based free riding prevention algorithm. Through a set of simulations, we find that the CBIM is feasible and outperforms other incentive mechanisms.