Free-rider Behavior In The Network Analysis And Inhibition Mechanism

Peer-to-Peer(P2P) is a newly emerging and popular network model in the research area of international network technology.Compared with traditional Client/Server model,each peer node in P2P network is equal because firstly,there are no strict distinctions between network service provider and consumer and secondly,the P2P networks are free from the authoritative central control.P2P technology is well received as a result of its openness, distribution and anonymity.The P2P networks are built with the intention that all the peers can share their own resource with other peers of the networks.However,resource sharing may bring some disadvantages such as bandwidth occupation,processing time delay, and even potential safety hazard.As network resources are nonexclusive,most of the peers in P2P networks prone to benefit from the resources without any contributions in return.Such behavior is called "free-riding".Those peers behaving in this way are called "free-riders". With the rapid development of the P2P networks,network resources are in high demand. Some studies show that only a small number of peers share their files with other peers and respond to their queries.Free-riding researchers have observed a high degree of free riding in P2P networks and they argue that free-riding has become a serious threat for the robustness and effective operation of P2P networks.In 2000,Adar and Huberman et al first stated the phenomenon of free-riding in most peers in P2P networks.Researchers in P2P field show great concerns for this issue and propose some mechanisms to monitor and restrain free-riding.The guiding principle of mechanisms is that the quantity and quality of the shared peer service are determined by the contributions they have made to the networks.Peers with fewer contributions can not search and download as much as those with more contributions.Mechanisms of restraining free-riding fall into two categories.One is incentive-based and the other is reciprocity-based.The first type of mechanism employs utility function to set the direct proportion between the service peers can access and their contributions to the P2P networks.The essential part of the mechanism is how to define and realize the utility function so as to ensure the service fairness of peers.The second type of mechanism manages to enhance the co-operation between peers.A peer chooses its counterpart by the degree of shared information of other peers.To be specific,peers connect with other peers based on the historical records of exchanging resources.Reciprocity can be further divided into two categories:One is direct reciprocity and the other is indirect reciprocity.The current typical mechanisms restraining free-riding include reputation mechanism and game theory etc.In short,P2P networks must take some safeguard measures to ensure enough sharing resources and high quality of the network service.Some problems in the traditional mechanisms to restrain free-riding in p2p networks are as follows:1.Utility functions are used to quantify the contribution peers have made to the networks.However,a reliable third party is needed in calculating the utility functions to ensure fairness.All peers are equal in the peer-to-peer networks and there are no servers.It is difficult to implement this kind of calculation without the third party.2.Peers' individual behaviors and the way they are connected have great influence on the whole performance of the system.Former mechanisms did not study intensively on individual behaviors of peers.The main work of this paper is to put forward a P2P Hieraxchy-Based Connection Management Protocol(HBCM) based on the exiting P2P confining free-riding models and different peers' behavior.HBCM protocol applies a distributed way independent of the third party to do the measurement job.P2P networks under the management of this new HBCM protocol can effectively isolate the free-riders and encourage peers to make contributions to the P2P networks.The thesis is divided into five chapters.Chapter one states the background of our study, the problem of free-riding and the structure of the thesis.Chapter two gives a detailed analysis of popular schemes to restrain free-riding.A novel Hierarchy-Based Connection Management protocol is proposed in the third chapter.HBCM protocol is designed to supervise the connections between peers and restrain free-riding in P2P networks.Chapter four provides simulation results to confirm the effectiveness of the HBCM protocol for improving the co-operations between the peers of the P2P networks.The fifth chapter is the conclusion part and presents the future work.