基于负载估计和子域消减的并行网络模拟拓扑划分算法

摘要

针对目前研究Internet的主要方法--传统的基于拓扑划分的大规模并行网络模拟的效率低、划分不均衡的问题,提出并实现了基于负载估计和子域消减的并行网络模拟拓扑划分策略.基于启明星辰探测获得的实际拓扑进行的DDoS攻击模拟表明,该划分方法可适用于实际拓扑,并可进行大规模网络安全事件的模拟.实验结果表明,该拓扑划分方法可缩减路由表规模约19.8%,减少模拟时间约17.2%.%Considering the low performance of the traditional topology partition-based parallel network simulation, the main method for Internet research at present, the paper puts forward a topology partition policy for parallel network simulation based on load estimation and subdomain subtraction. The results of the simulation based on the real topology obtained by Venusense prove that this partition policy can be used for real topology and large scale security event simulation. The test results prove that this partition policy can reduce the scale of router table by 19.8 percent, the simulation time by 17.2 percent, and the performance of simulation can be improved.