基于滑动地址序列的IPv6网络拓扑发现引擎

IPv6 topology discovery engine based on a sequence of sliding addresses

为满足IPv6网络的拓扑发现需求,提出了一种基于滑动地址序列的IPv6网络拓扑发现算法。算法根据并发进行的分布式探针数量,将待探测的目标地址集等分为若干子集,在每个周期内每个探针只按序针对其中的一个子集进行探测,同时探针之间互相通告各自探测的结果;此后每个探针按序对下一个子集进行探测,如果发现过程抵达前面的探测周期中其他探针已经发现的共同路径时,则停止继续向前探测。基于本算法实现的系统对一个真实的纯IPv6主干网环境进行了全面发现实验。实验结果表明,该引擎减少了约35%的IPv6探测报文,达到设计目的。

An IPv6 network router-level topology discovery algorithm is developed where the target addresses are divided into m subsets, which is the number of parallel distributed probes. In each round, each probe only detects one of the subsets. The probes then notify each other of the detection results at the end of each round. Then, each probe begins to detect the next subset in order. If a probe finds that it is encountering the same path that was detected by another probe in the previous round, it will stop the probing process for the current target. A prototype based on the algorithm was implemented in a pure IPv6 network environment. Tests show that the algorithm eliminates about 35% of the probed IPv6 packets to achieve better efficiency than previous approaches.