Multi-characteristic Subnets Discovery and Analysis Based on Traceroute
Wei Yao,
Hai Zhao,
Jian Zhu,
Xiangyi Chen
July, 2020
Abstract
互联网测量的研究促进了路由器级拓扑发现的发展,而网络层的子网能为其提供更详细的中间互补视图.针对子网边界条件以及完整性考虑不足引起的准确率较低问题,提出了一种多特征结合子网发现算法.研究了同一子网IP的traceroute路径特征,将多个特征结合设计更精准的子网边界判定条件.通过筛选子网的完整性,缩小候选子网的搜索空间,启发式求解子网发现问题.实验结果表明,本文算法与现有其他算法相比,能更准确地发现子网,有效地减少子网误报情况,同时效率有所提高.最后,对六个地理上分散的ISP进行子网推断,并分析了这些ISP之间常见的各种子网特征.
Publication
In Journal of Northeastern University (Natural Science)
Ph.D in Computer Science
My research interests include include edge computing, edge AI, software-defined networking, deep reinforcement learning, federated learning, etc.