Evaluating System of Systems resilience using interdependency analysis

A System-of-Systems (SoS) is a collection of distributed independent individual systems that interact with one another to achieve an SoS capability requirement that cannot be achieved by individual systems alone (in contrast, in a monolithic system, hardware or software components are integrated to form a single entity). Interdependency between SoS systems, while enabling new capability, also means that failures can cascade throughout the SoS, creating development delays or additional system failures. Here we develop a method based on Bayesian networks to evaluate the resilience of SoS design alternatives to failures during operations. We propose a conditional resilience metric that measures each constituent system's contribution to overall SoS resilience, and a resilience pattern that shows how SoS performance degrades as systems fail. SoS resilience is determined by both the SoS architecture and the constituent system reliability. In a simple example based on the Littoral Combat Ship SoS, we determine the two most critical systems using the conditional resilience metric. Adding a communications link between these two systems increases the resilience, resulting in higher expected performance and slower expected performance degradation as a result of system failure. The conditional resilience and resilience pattern can be used by designers and other decision makers to identify SoS architectures and systems that increase SoS resilience to failure.