Cascading Failures in Interdependent Infrastructures: An Interdependent Markov-Chain Approach

Many critical infrastructures are interdependent networks in which the behavior of one network impacts those of the others. Despite the fact that interdependencies are essential for the operation of critical infrastructures, such interdependencies can negatively affect the reliability and fuel the cascade of failures within and across the networks. In this paper, a novel interdependent Markov-chain framework is proposed that enables capturing interdependencies between two critical infrastructures with the ultimate goal of predicting their resilience to cascading failures and characterizing the effects of interdependencies on system reliability. The framework is sufficiently general to model cascading failures in any interdependent networks; however, this paper focuses on the electric-cyber infrastructure as an example. Using this framework it is shown that interdependencies among reliable systems, i.e., systems with exponentially distributed failure sizes, can make the individually reliable systems behave unreliably as a whole with power-law failure-size distributions.