A Modular Approach to Non-deterministic Dynamic Fault Trees

Abstract

Dynamic Fault Trees (DFTs) are powerful tools for deriving fault-tolerant system designs. However, deterministic approaches to DFTs suffer from semantic struggles with problems such as spare races. In this paper, we discuss the added complexity in the state-space representation of a non-deterministic DFT model and propose a modularized approach for synthesizing recovery automata. Finally, we give an implementation and evaluate it on the Fault tree FOResT (FFORT) benchmark. The results show that non-deterministic semantics with modularization can scale for literature case studies.