Researches On Inhibition Mechanism In Resource Sharing P2P Networks Based On Game Theory

P2P network broke the traditional C/S and B/S network model and became the mostpopular Internet application technologies soon. However, plenty researches which usedvarious ways and took long time found that there are a large number of nodes in P2P networkdownload amount of resources from P2P system but never contribute anything. Thisphenomenon which is called free-ride brought a heavy burden to enthusiastic nodes in thenetwork. Free-ride phenomenon brought tremendous negative impact to P2P system.Moreover, P2P system may become vulnerable or even collapse because of the seriousfree-ride phenomenon. The large number of researchers began to explore a potent inhibitormechanism to resolve free-rider problem which became worse and worse. The main free-riderbehavior inhibition mechanism is divided into four categories: Monetary-Based Mechanisms,Incentive Mechanisms, Reputation-Based Mechanisms and Game Theory Mechanisms. In thisthesis, we proposed two free-rider behavior inhibition mechanism: FEM （FutureExpectation-based Mechanism）and ITTM(Improved Tit for Tat Inhibition Mechanism).In the FEM, when a node decided whether choose cooperate, it will take into account thehistorical performance of the other node. This historical performance is called reputation inthis mechanism. According to the fact that each node has different expectation for future inreal networks, we put forward another important reference factor-the discount rate. In thisthesis, we analyze this inhibitory mechanism in three different aspects: First, in the wholenetwork, there is only a single node change its own reputation. We can found that if theaverage reputation value remains unchanged, this node will always adjust its reputation valueuntil1. Second, if nodes in the network all the nodes were trying to adjust their reputation toincrease their benefit from the system, we can found that the average reputation of all nodes inthe network will hold an upward trend. In other words, the nodes in the network will share itsresource to other nodes in order to obtain more benefit. Third, by comparing two differentnodes with different discount rates which shows these nodes have different futureexpectations. According to a rigorous mathematical analysis we can find that the node whichhad bigger discount rates can get more benefit from system in the end. Those three pointsprove the effectiveness of Future Expectation-Based Mechanisms to inhibit free-rider behavior in P2P system.Improved Tit for Tat Inhibition Mechanism is the improvement of a tit for tat mechanism.This mechanism adopted a more "tolerant" attitude which means that node will be publishedafter betrayed n times. The simulation results showed that the Improved Tit for Tat InhibitionMechanism has a higher efficiency and greater network throughput than Tit for Tat InhibitionMechanism in the same environment.Simulation experiments in small world network showed the correctness and effectivenessof the two free-rider inhibition mechanisms. However, the assumed cond itions of thesimulation experiments is relatively strict, the experiments was run in a laboratory and theirmuch different from the real network. Therefore, there are lots of works to do in the future inorder to propose a practical free-rider behavior inhibition mechanism.