Multigraph critical infrastructure model

Abstract

Interdependencies between critical infrastructures have been studied widely, but largely at the abstract and structural levels with an emphasis on large infrastructure networks and frequently their stochastic properties. However, an in-depth understanding of infrastructure interdependencies and the likely impact of degradation of selected elements are important for an adversary intent on maximizing attack efficiency. This chapter describes a simple multigraph model for several classes of interdependent critical infrastructure elements and an attack tree model with attribute domains extended by acyclic phase-type distributions to capture temporal dependencies. The efficacy of this modeling approach is demonstrated via a case study involving regional interdependent infrastructures that include the electric power, water and telecommunications sectors in a bounded region. The case study uses extensive simulations to demonstrate that an adversary with access only to publicly-available information and the ability to analyze a multigraph model can cause severe harm.