Survivability analysis of reconfigurable systems

This paper presents a new probabilistic approach to analyze the survivability of a system. We show similarities and differences between survivability and reliability analysis. In a reliability model, we can describe a k-out-of-n:G system (k <= n), in which the component system is good only if any k or more components work. The system can also be configured into an initial k-out-of-n:G model with m backup components (0 <= m <= n). If the system cannot perform its intended function, the m backup components will be reconfigured with the remaining working components into a new form to sustain system function. We refer to such studies to calculate the system successful probabilities as the survivability analysis. In this paper, we focus on the survivability analysis of a simple piping system. This study could potentially be used to analyze the survivability of power network systems, dependable secure computing systems, military reconfgurable information systems, and other large reconfigurable network systems.