Peer Selection for Reliability Improvement in P2P Networks by Fuzzy-Based and Ns-3 Simulation Systems

In Peer-to-Peer (P2P) networks, the peers should collaborate with trustworthy neighboring peers. However, it should be noted that in real applications, the peer may be faulty, which may make a wrong decision. For safe communication, the peers should be reliable. The reliability can be evaluated by the interaction of a peer with other peers but this procedure needs many parameters. In this paper, in order to make a qualitative evaluation, we implement a Fuzzy-based system and for quantitative evaluation we build another simulation system by taking into account the correlation coefficients and ns-3 simulator. Three parameters Packet Loss (PL), Delay Time (DT) and Throughput (TG) are considered for the evaluation of both systems. For Fuzzy-based system, the output is Peer Selection Decision (PSD). From the simulations results of FL-based system, we conclude that when PL and DL are increasing, the PSD is decreased. While, when TG is increased, the PSD is increased. By ns-3 simulator results, we found that by using the value of correlation coefficients is possible to select the best reliable peers.