Dynamic Fault Tree Analysis Using Input/Output Interactive Markov Chains

Dynamic Fault Trees (DFT) extend standard fault trees by allowing the modeling of complex system components’ behaviors and interactions. Being a high level model and easy to use, DFT are experiencing a growing success among reliability engineers. Unfortunately, a number of issues still remains when using DFT. Briefly, these issues are (1) a lack of formality (syntax and semantics), (2) limitations in modular analysis and thus vulnerability to the state-space explosion problem, and (3) lack in modular model-building. We use the input/output interactive Markov chain (I/O-IMC) formalism to analyse DFT. I/O-IMC have a precise semantics and are an extension of continuous-time Markov chains with input and output actions. In this paper, using the I/OIMC framework, we address and resolve issues (2) and (3) mentioned above. We also show, through some examples, how one can readily extend the DFT modeling capabilities using the I/O-IMC framework.

• Type: Conference
• Year: 2007
• Where: DSN
• Authors: Hichem Boudali, Pepijn Crouzen, Mariëlle Stoelinga
• Keywords: Computer Networks, DSN 2007, Dynamic Fault Trees, Fault Trees, Markov Chains