A self adaptive super peer based topology for unstructured P2P system

Abstract

Peer to peer network (P2P) has emerged as a popular model for file sharing because of high scalability, low cost and easy maintenance. The success of a P2P network is highly dependent on its “Search Mechanism” and “Topology Formation” strategy. Many researchers have dealt with super peer based on overlay topology where the key factor is selection of a super peer from all the peers. In this paper, a novel mechanism called “Adherence Model” is proposed which describes a new selection method for super peers within a P2P network. The selection process in the “Adherence Model” is different comparing to the general super peer selection processes as it considers the most significant factors like bandwidth, processing capability and storage space of a peer. “Adherence Model” selects super peers by considering the adherence factor of peers in the network. The capacity, lifetime, upload-download ratio and trust value of peers have taken into consideration while selecting these super peers. Moreover, this model introduces “Acting super peer” to balance the heterogeneity and dynamism in the network. Simulation results show that how search efficiency is improved by selecting proper “Super Peers” and how it helps in achieving proper “Load Balance” through the overall network. Finally, this model is also capable of minimizing the effect of “Free riders” and “Malicious peers”.