一种互联网AS层拓扑的动态演化模型  

Novel dynamic evolution model for Internet AS-level topology

作　　者：郭虹[1] 卜佑军[1] 兰巨龙[1] 刘洛琨[1] 

机构地区：[1]国家数字交换系统工程技术研究中心,河南郑州450002

出　　处：《系统工程学报》2011年第2期152-159,共8页Journal of Systems Engineering

基　　金：国家高技术研究发展计划资助项目(863:2007AA01Z2a1);国家高技术研究发展计划资助项目(863:2008AA01A-323);国家重点基础研究发展计划资助项目(973:2007CB307100)

摘　　要：了解互联网的拓扑结构、理解互联网的演化机制对优化网络性能、指导网络协议设计至关重要.本文着重探讨了互联网AS(autonomous system,自治域)层拓扑的局部连接偏好性和簇度相关性,提出了局域世界的动态地改变大小和选取方法;并基于互联网规模的摩尔定理给出了网络指数增长的模拟细节;并在吸取局域世界模型思想和PFP(positive feedback preference,正反馈偏好连接)模型非线性择优概率的基础上,充分刻画出拓扑演化过程中节点与链接的生与灭随机事件,建立了一种互联网AS层拓扑动态演化模型(ITALW).计算机建模和数值仿真分析表明,该模型不仅能够模拟大部分网络宏观特征,还能够较为准确地模拟互联网AS拓扑的局部行为.Understanding Internet topological structure and investigating its evolving mechanisms are crucial for optimizing its transmission performance and providing theoretical foundation for high-performance routing protocol designing. In this paper, firstly, Internet AS（autonomous system）-level topology＇s local connectivity preference is explored and methods of the dynamic size and determination of local-world are presented. Secondly, considering Internet size growth in an exponential ＂form with the famous Moore＇s law, a method about exponential adding new nodes and dying out old nodes is given. Finally, a dynamic evolution model named ITALW is proposed to more accurately depict the Internet AS-level topology. This model based on the local-world evolving idea and PFP＇s（positive feedback preference） nonlinear probability is used to simulate some network stochastic events, such as nodes＇ and edges＇ boring and dying. Numerical simulations show that most of macroscopic statistics, closed to real networks＇, and the Internet-AS topology＇s local behavior could be simulated accurately.

关 键 词：Internet网络拓扑 自治域级拓扑建模 局域世界 指数增长 局部集聚系数 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]