Link Importance Evaluation of Data Center Network Based on Maximum Flow

Abstract

Recently, data center networks (DCN) have gained considerable attention from both the academia and industry. However, the main research of DCN is limited to load balancing or the improvement of network architectures. Moreover, several of them focus on the security of DCN architecture, especially for the vulnerability evaluation of DCN. In our previous solutions, we fill this void by presenting an algebraic method and contribution matrix of information entropy from the perspective of node importance. Differ from the existing ones; we engaged in the evaluation of links importance in this paper. Above all, we use a directed and weight graph to describe the topology and then formulate and convert the performance analysis into a typical maximum-flow problem by adding a supersource and a supersink. What is more, we get the value of maximum-flow and the corresponding links by using Edmonds and Goldberg algorithm. Last but not the least, according to the theory of maximum-flow and minimal cut sets, we get the critical links of each architecture. Extended experimental results and analysis show that our method is effective. Furthermore, the issues observed in this paper are generic and can be widely used for newly proposed architectures.