Skip to Main Content
Supervisory Control and Data Acquisition (SCADA) systems control and monitor industrial and critical infrastructure functions, such as the electricity, gas, water, waste, railway and traffic. Recently, SCADA systems have been targeted by an increasing number of attacks from the Internet due to its grow- ing connectivity to Enterprise networks. Traditional techniques and models of identifying attacks, and quantifying its impact cannot be directly applied to SCADA systems because of their limited resources and real-time operating characteristics. The paper introduces a novel framework for evaluating survivability of SCADA systems from a service-oriented perspective. The framework uses an analytical model to evaluate the status of services performance and the survivability of the overall system using queuing theory and Bayesian networks. We further discuss how to learn from historical or simulated data automatically for building the conditional probability tables and the Bayesian networks.