【摘要】 在分析现有网络拓扑推测算法的基础上,提出了一种基于端到端报文丢失性能的网络拓扑推测方法,该方法通过比较节点间的Hamming距离和节点接收到的报文数量推测网络拓扑,在推测拓扑过程中不需要网络内部节点的协助。通过理论比较和分析,证明该方法在推测精度上比现有的方法有显著的提高,最后利用仿真实验验证了该方法的有效性和准确性,推测精度比现有方法提高了20%。更多还原

【Abstract】 Based on the analysis of existing network topology identification algorithms,an improved network topology identification method based on end-to-end loss performance was proposed.The proposed algorithm identified network topology using Hamming distance of the sibling nodes and inferred number of received probe packets at sibling nodes.The proposed algorithm required no support from internal nodes.The theoretic analysis and comparison shown that the improved algorithm can significantly improve the inference accuracy.To validate its accuracy and efficiency,the pro-posed algorithm was implemented in a simulated network.The simulation result shown that the inference accuracy is im-proved more than 20%.更多还原