Research On Free-riding Behavior And Optimization On Incentive Mechanism In P2P Network

The user behavior analysis and design of incentive mechanism in P2P systems has become the hotspot in recent years.However,the existing incentive mechanism still has many shortcomings,mechanism in P2P systems such as BitTorrent can't stop free-riding behavior,free-rider can get download rate from the seeds.Be faced with the problems of nodes in the network with asymmetry of interest,the small possibility of repeat transaction,and the difficulties in creating mutual relations,the existing research implemented the transfer for mutual relations among the nodes by sharing history information to solve these problems.However,these works haven't considered the characteristics of the cluster of interest in nodes and referenced to successful experience in TFT strategy of BitTorrent system.The research purposes in this dissertation is improving the performance of P2P networks through researching the incentive mechanism of BitTorrent file sharing system, referencing to the successful experience of it,and considering the influence of the node history trades to obtaining the shared resource in the P2P network and suppressing the free-riding behavior in the peer-to-peer network..The dissertation took BitTorrent file sharing system as research background.It analyzed free-riding behavior with the fluid mathematical model,and provided the basis for improving the incentive mechanism in BitTorrent system.It proposed an improved strategy based on seed bandwidth assignment for the incentive mechanism in BitTorrent, and theoretically analyzed that the strategy can achieve the Nash-balance.In the foundation of TFT strategy in BT system,it has introduced group-trusts,proposed incentive mechanism based on the group.It designed and realized a serial simulator BTsim which taken the BitTorrent system as the background and actuated based on discrete events,and made the improvement to enable it to simulate the incentive mechanism based on the group.The simulating experiments demonstrated that the incentive mechanism base on the seed bandwidth assignment has punished free-riders successfully by extending its downloading time,and reduced contributor's average downloading time.The strategy has provided incentive effectively.It simulated incentive mechanism GIM based on group. Compared with the RIT mechanism,the GIM is superior to the RIT mechanism in system expandability,processing node change ID and whitewasher.