Modelling to Safety

Abstract

In this position paper we outline the challenges that face safety critical systems. We identify the need to shift the validation emphasis from process to product, and state how formal proofs would be of great benefit by providing stronger evidence for safety case arguments. We also argue that a successful solution for the incremental certification problem could bring benefit to all levels of system design. A better understanding of the non-functional behaviour and methods for expressing, embedding into design, and managing properties that play a role in this aspect of safety critical systems are urgently needed. We also speculate that research into trusted components and compositional architectures is vital for the future of safety critical systems design. In partnership with BAE SYSTEMS, Rolls Royce, and QinetiQ, we have developed a framework and an architectural description language that addresses these issues.