一种基于时空可信度推断AS商业关系的方法

Method for Inferring AS Relationship Based on Temporal and Spatial Reliability

摘要: 本文提出了一种基于时空可信度的AS商业关系推断方法，该方法选取多个时间点的不同监测点路由表，基于路由策略和层次结构初步推断每个时刻不同监测点的AS商业关系，并基于空间一致性和时间稳定性计算其可信度值，根据给定阈值选取可信的AS商业关系。同时，本文还分析了路由监测点数量与阈值对有效AS商业关系推断的影响。实验结果表明，选取恰当数量的路由监测点与阈值可以有效提高AS商业关系的精准度。

Abstract: In this paper, a new method of inferring AS relationship by analyzing the temporal and spatial reliability is proposed. Firstly, the method selects the routing tables which originate from different monitored points in different intervals. Based on the routing strategy and the hierarchical structure, the AS relationship of different monitored points is inferred in every time window. Moreover, we define and calculate the value of the temporal and spatial reliability by analyzing the spatial consistency and time stability. Furthermore, we set a threshold and regard any AS relationship with reliability value greater than the threshold as trusted relationship. In the meanwhile, we analyze the number of monitored points and the threshold to infer the influence on AS relationship’s validity. Experiment result shows that the accuracy of AS relationship deduction can be largely improved by selecting appropriate number of routing monitored points and the threshold value through our proposed methods.