P2P节点在线机制的性能评估与定价策略的研究

发布时间：2019-04-11 13:34

【摘要】：近年来随着P2P技术迅速发展，P2P网络逐渐成为互联网的一个重要分支。如何有效调节P2P节点的自私行为，提高网络资源利用率和服务性能，成为重要的研究课题。本文依据P2P节点的在线机制，建立连续时间Markov模型，进行性能评估，通过博弈分析，，给出定价策略，实现P2P网络社会最优。首先，依据移动P2P节点在线机制，考虑P2P节点移动性强，随时上下线的特点，建立服务台数随机变化的二维连续时间Markov模型。构建系统的状态转移率矩阵，采用矩阵几何解法，进行系统模型的稳态分析。针对移动P2P节点给出失效率，节点平均延迟等系统性能评价指标的定义及表达式。其次，依据混合P2P节点在线机制，将节点状态分为上线状态和离线状态，结合请求节点个数及在线节点个数，建立一个服务台数随机变化的三维连续时间Markov模型。构建系统的状态转移率矩阵，利用矩阵几何解法，对模型进行稳态分析。针对混合P2P节点给出失效率，节点平均延迟等系统性能评价指标的表达式。然后，基于移动和混合P2P网络，针对不同的请求节点到达率、离去率以及服务节点到达率、离去率等系统参数分别进行数值实验与仿真实验。定量刻画节点失效率和节点平均延迟等性能指标随服务节点个数和请求节点到达率等参数的变化趋势，对网络性能进行系统评估。最后，考虑传输回报和逗留成本，构造收益函数，分别针对移动P2P网络和混合P2P网络给出纳什均衡和社会最优策略。通过数值结果验证纳什均衡节点到达率高于社会最优节点到达率。针对请求节点给出合理的定价方案，最终达到P2P网络的社会最优。
[Abstract]:In recent years, with the rapid development of P2P technology, P2P network has gradually become an important branch of the Internet. How to effectively regulate the selfish behavior of P2P nodes, improve the utilization of network resources and service performance, has become an important research topic. In this paper, based on the online mechanism of P2P nodes, a continuous-time Markov model is established to evaluate the performance of P2P networks. Through game analysis, pricing strategies are given to realize the social optimization of P2P networks. Firstly, according to the on-line mechanism of mobile P2P nodes, a two-dimensional continuous-time Markov model with random changes in the number of service stations is established, considering the strong mobility of P2P nodes and the characteristics of both the up and down lines at any time. The state transition rate matrix of the system is constructed, and the steady state analysis of the system model is carried out by using the geometric solution of the matrix. In this paper, the definition and expression of system performance evaluation indexes such as loss efficiency, average delay of nodes and so on are given for mobile P2P nodes. Secondly, according to the online mechanism of hybrid P2P nodes, the node states are divided into on-line state and offline state. Combined with the number of requesting nodes and online nodes, a three-dimensional continuous-time Markov model with random changes in the number of service stations is established. The state transition rate matrix of the system is constructed, and the steady state analysis of the model is carried out by using the geometric solution of the matrix. For hybrid P2P nodes, the expressions of system performance evaluation indexes, such as loss efficiency, average delay of nodes, are given. Then, based on mobile and hybrid P2P networks, numerical and simulation experiments are carried out for different system parameters, such as arrival rate of request node, departure rate, arrival rate of service node, removal rate and so on. The network performance is evaluated by quantificationally describing the changing trend of node failure rate and node average delay with the number of service nodes and the arrival rate of requesting nodes. Finally, considering the return of transmission and the cost of stay, a revenue function is constructed, and Nash equilibrium and social optimal strategies are given for mobile P2P networks and hybrid P2P networks, respectively. The numerical results show that the Nash equilibrium node arrival rate is higher than the social optimal node arrival rate. This paper gives a reasonable pricing scheme for the requesting node, and finally achieves the social optimum of P2P network.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02

【参考文献】