Local Contribution (LC) and Trustworthiness Factors to Induce Fairness in P2P Networks

A peer-to-peer (P2P) network is a distributed system in which the autonomous peers participate at their motivation and resources are shared in distributed manner. P2P systems can be classified in two major categories: centralized and distributed. Distributed P2P systems can be further classified in two types: structured and unstructured. With immense increase in Internet, the P2P networks have found increased interest in the research community since the different protocols for these networks can be gainfully utilized in the resource discovery process for Internet related applications. However, there are several challenges in designing an efficient protocol for the P2P networks. These networks suffer from problems such as fake content distribution, free riding, whitewashing, poor search scalability, lack of a robust trust model, etc. Moreover, the peers can join and leave the network (churn) at any time which makes trust management and searching in these networks more challenging. Free riding nature of the peers damages overall performance of the P2P systems. There are several mechanisms proposed for promoting fairness in the P2P systems like tit-for-tat policy, point-based incentive policy, PageRank, Eigen Trust, layered taxation, advertisement of incentive, etc. We have observed that in some schemes like point-based incentive scheme, peers are stricken in the process of requesting data download which leads to the starvation. We have proposed a new scheme which collects the statics of resources shared by the participating peers in P2P networks and calculated local contribution and trustworthiness factors based on the statistics. The starvation problem is minimized in the proposed scheme. We have done analytical modeling and analysis of different costs involved in downloading a file with the costs in existing scheme and found that the costs involved in proposed scheme is less than the existing scheme.