Nodes and links jointed critical region identification based network vulnerability assessing

Abstract

Large-scale regionally-correlated failures resulting from natural disasters or intentional attacks pose a great threat to the physical backbone networks, since the impact of large-scale failures can cause network nodes and links co-located in a large geographical area to fail. When the same intensity of threats occurs at different physical locations, the damage to the network performance varies greatly. In the network vulnerability assessment, the critical region is defined as the destructed area which would lead to the highest network disruption. Traditional critical region identification for network vulnerability assessment is only determined by nodes, without considering the failures of the links. To this end, this paper proposes a critical region identification method that joints nodes and links to find the critical region. We study this vulnerability assessment problem in two cases, the special case of the failure center constrained at a network node and the general one of that at an arbitrary location, and propose two algorithms for these two cases respectively. The simulation-based experiment on synthetic network is given with different criticality metrics. The simulation results verify the feasibility and effectiveness of our proposed critical region identification method in comparison to others.