带恶意节点和服务节点故障可修的P2P网络安全性能分析

P2P Network Security Performance Analysis with Malicious Nodes and Repairable Service Node Failures

摘要: Peer-to-Peer (P2P)网络是逻辑层分布式结构，在互联网内容分发与业务共享领域应用优势显著，每天在全球范围内有大量文件经由P2P业务上传和下载，使得P2P流量大幅增长甚至造成流量拥堵，降低节点在线能耗是提高P2P网络安全稳定性的关键所在。为控制节点不必要在线行为，本文基于现行混合P2P网络中存在的请求节点和恶意节点，为服务节点引入同步多重工作休假策略，建立带有负顾客、启动期、关闭期、工作休假且故障可修的M/M/c排队模型，研究P2P共享系统的机制及安全性能。采用矩阵几何解法以及Gauss-Seidel迭代法得到模型稳态分布及关键性能指标，通过数值实验分析各指标对系统性能的影响，通过纳什均衡和社会最优策略，得到P2P节点合理收费的社会效益最优值，为P2P网络提供更为有效的降低节点在线能耗的方案，从而提高网络安全稳定性。

Abstract: Peer-to-peer (P2P) network is a logical layer distributed structure, which has significant advantages in the field of Internet content distribution and service sharing, and a large number of files are up-loaded and downloaded through P2P services around the world every day, making P2P traffic in-crease significantly and even causing traffic congestion, and energy consumption has become a hot issue that has attracted widespread attention. In order to control the unnecessary online behavior of nodes and reduce the energy consumption of the system, based on the current hybrid P2P net-work node online mechanism, this paper introduces a synchronous multiple work leave policy for the server-side node, establishes an M/M/c queuing model with malicious nodes, startup period, shutdown period, work leave and fault repair, and studies the mechanism and performance of P2P sharing system. The steady-state distribution and key performance indicators of the model are obtained by using the matrix geometric solution method and the Gauss-Seidel iterative method, and the influence of each index on the system performance is analyzed through numerical experiments, and the optimal value of social benefits of reasonable charging of P2P nodes is obtained through the Nash equilibrium and social optimal strategy, which provides a more effective solution for P2P net-work to reduce the online energy consumption of nodes, so as to improve the network security stability.