Performance Analysis Of P2P Network Based On Two-stage Service With Failure Repairability

Along with the rapid development of the information technology revolution,P2P tech-nology has been more and more widely and deeply researched in the field of mobile wire-less,and mobile P2P has become an emerging and popular research field.The particularity of the mobile environment has extremely high requirements on the P2P network topology,resource transition paths and network security.According to the actual situation of dynamic changes of peers,various strategies are introduced in the two stage queueing system,two P2P queueing models are established based on a hybrid mobile network model combining the semi-distributed topology and the fully-distributed topology,and performance analysis of system is developed to provide decision foundation for the mobile P2P network model.Firstly,the data of P2P network is described according to the data information of the blockchain websites,and the~2test is used to verify that the request arrival interval obeys an exponential distribution,which provides a basis for the distribution hypothesis of the request arrival process of the penalty strategy model.Secondly,the requesting peer in P2P is abstracted as a customer,and the search and transmission process of the peer is abstracted as the service process,and a two-stage queue-ing model with negative customers and repairable faults is established.Using the matrix-geometric solution and Gauss-Seidel iterative method,the probability distribution of the net-work model in steady-state is obtained,and expressions for performance indicators such as the average number of peers in the two stages are given.Numerical experiments were con-ducted using programming software to analyze the effect of parameters on each indicator.The optimal parameters were solved by constructing the average cost unit time function and social benefit function to provide decision foundation for the mobile P2P network model.Finally,a two-stage queueing model is built in an almost invisible situation,which con-siders impatient peers dropping out and peers that fail to accept the service initiating a retry.Using the matrix geometry solution method and Gauss-Seidel iterative method,the num-ber of requested peers in the system for each stage and space and the leaving probability of peers for each type of reason are obtained.Numerical algorithms are derived by computer programming where the influence of the system parameters on the performance indicators is analyzed.The average cost per unit time function and the social utility function of the system are constructed and the socially optimal parameters of the system are analyzed by using Nash equilibrium.