单测量源点的位置对网络拓扑测量结果的影响

Influence of single measurement source placement on network topology measurement

在用网络测量手段发现网络拓扑中,测量源点位置的选择通常被认为起着至关重要的作用。单点测量中源点的位置是否会导致测量结果有显著差异是值得研究的问题。通过在仿真环境中产生不同的拓扑模型,分析了单侧量源点在发现链路(节点)数量方面的统计特性。实验表明,在随机和无标度拓扑中,单个源点发现的节点或链路数量都可近似地用正态分布描述;不同源点在获取链路(节点)数上的最大差值具有稳定性,与目标节点的比例无关;源点的度对其测量结果没有影响。单测量源点在拓扑发现中的测量差异并不明显。

The selection of measurement source placement is usually thought to be important in discovermg network topology by measurement method. In single source measurement, whether the source placement leads to marked difference in the measurement result is a problem that deserves attention. By generating different topology models and simulating measurement process in simulation environment, the statistical properties of single source in discovered links or nodes were analyzed. The simulation experiments show that the number of links or nodes discovered by single source in both random and scale-free topologies may be described by normal distribution. The maximum difference value among different sources in the number of discovered links or nodes is stable, which is independent of the destination ratio. The degree of a source does not influence its measurement results. The source＇s measurement diversity is not obvious in discovering topology.