Research On Spreading Dynamics Based Key Nodes Identification In Inter-domain Routing System

Inter-domain routing system,which is of vital importance for country and network security,is key infrastructure of the internet.Cascading failure phenomenon is common within inter-domain routing system,based on which attacking the key node would crash the whole internet,so that identifying and protecting key nodes of inter-domain routing system is of great importance to keep it safe and secure.However current researches are limited to static structure attributes to evaluate importance of key nodes,leading to misunderstanding of influence introduced by key nodes during cascading failure process.Considering the above problem,this paper proposed a spreading dynamic based inter-domain routing system key node identification method.Firstly,two damage factors,causing cascading failure,are incorporated to construct inter-domain routing system cascading failure model,then the correlation between static attributes of nodes and failure propagation is analyzed after evaluating node importance,finally both dynamic and static inter-domain routing system key node identification methods,which are oriented to failure propagation,are summarized to provide theoretical foundation for protecting inter-domain routing system.This paper contributes as follows:1.A double damage factors based inter-domain routing system cascading failure model(DDF-CFM)is proposed after analyzing failure conditions caused by links and nodes respectively.At the same time,BFS based UPDATE packets propagation model and optimal path based traffic redistribution policy are incorporated to imitate cascading failure process comprehensively.The experiment was tested on existing cascading failure dataset and the result turns out that DDF-CFM could effectively describe cascading failure process of inter-domain routing system with the prediction error decreased by 34%.2.A spreading dynamic based inter-domain routing system key node identification method(SD-KNI)is proposed,within which node importance evaluation model is the core part that could quantify the impact caused by crashed nodes to evaluate the node importance during cascading failure process.SD-KNI could evaluate the node importance effectively and the result turns out that SD-KNI could identify key nodes with the precision increased by 16.7% and 21.9% at most on dataset collected from DDF-CFM model,when compared with existing methods.3.A double static attributes based inter-domain routing system key node identification method(DSA-KNI)is proposed,which is based on the strong correlation existed between degree centricity,routing path number and invalid effects after analyzing.Given these two static attributes,DSA-KNI could effectively identify key nodes within inter-domain routing system according to node importance.When compared with SD-KNI algorithm,the time complexity has been reduced to O(lmn).And the experiment result shows that the key node identification precision increased by 7.4% in average and 12.9% at most with 90% coverage ratio when nodes surpass 25.