On the vulnerability of multi-level communication network under catastrophic events

Various cyber-physical infrastructures such as communication networks and power grids are known to be vulnerable to large-scale stressors ranging from natural disasters to intentional attacks such as those effected by weapons of mass destruction and high-altitude electromagnetic pulses. The stresses instigated by these events can cause damage to critical components of the network infrastructure. In this paper, a general probabilistic model is developed for assessing the vulnerability of a communication network under various catastrophic events. A multi-level scalable network framework is proposed to capture the inter-dependencies across various communication networks in the infrastructure. For a given large-scale stressor, the initial-failure probability of each network component is formulated independently and then by taking into account the failure of the components that it depends upon. This enables the modeling of a shared failure among network components. Detailed simulations of a three-level network model are performed and key network-performance metrics are computed including the total network capacity, the maximum flow and the number of node failures. This work paves the way to model and evaluate the reliability of critical communication networks under massive stressor events.