基于二分图模型的通信信息网络故障联合定位

Network fault location based on bipartite graphs for communication and information networks

国家电网信息通信网络依靠两套运维系统,分别实现对信息网络与通信网络的故障定位与分析,然而通信网络故障往往会引发信息网络故障,如何高效精确地进行通信信息网络故障联合定位是亟需解决的问题。针对信息通信网络的联合故障定位问题,提出了基于二分图模型的故障联合定位算法。首先依据通信网网络节点的关联性对网络分簇,并将每一簇作为一个子域。其次在每个子域内建立基于二分图的故障关联影响模型,最终利用目标排序法并行地对多个子域内网络故障进行分析,从而实现通信信息网络关联故障高效精确的联合定位。实验结果表明,该联合故障定位分析方法的故障诊断率达85%~95%。

The current State Grid communication and information networks rely on two separate network management systems. Such management structure brings challenges to network fault location, which calls for joint network fault analyze method. Grasping the transmission correlation of network nodes, a joint network fault location mechanism based on bipartite graph model was proposed. Firstly, network nodes were clustered into multiple sub-domains based on their fault correlation. Then, the bipartite graph fault correlation model was established for each sub-domain. Finally, the target sorting method was applied to locate network faults in every sub-domain simultaneously. Simulation experiments show that this method can locate network faults with high accuracy.