Assuring Safety and Security

Large technological systems produce new capabilities that allow innovative solutions to social, engineering and environmental problems. This trend is especially important in the safety-critical systems (SCS) domain where we simultaneously aim to do more with the systems whilst reducing the harm they might cause. Even with the increased uncertainty created by these opportunities, SCS still need to be assured against safety and security risk and, in many cases, certified before use. A large number of approaches and standards have emerged, however there remain challenges related to technical risk such as identifying inter-domain risk interactions, developing safety-security causal models, and understanding the impact of new risk information. In addition, there are socio-technical challenges that undermine technical risk activities and act as a barrier to co-assurance, these include insufficient processes for risk acceptance, unclear responsibilities, and a lack of legal, regulatory and organisational structure to support safety-security alignment. A new approach is required. The Safety-Security Assurance Framework (SSAF) is proposed here as a candidate solution. SSAF is based on the new paradigm of independent co-assurance, that is, keeping the disciplines separate but having synchronisation points where required information is exchanged. SSAF is comprised of three parts - the Conceptual Model defines the underlying philosophy, and the Technical Risk Model (TRM) and Socio-Technical Model (STM) consist of processes and models for technical risk and socio-technical aspects of co-assurance. Findings from a partial evaluation of SSAF using case studies reveal that the approach has some utility in creating inter-domain relationship models and identifying socio-technical gaps for co-assurance. The original contribution to knowledge presented in this thesis is the novel approach to co-assurance that uses synchronisation points, explicit representation of a technical risk argument that argues over interaction risks, and a confidence argument that explicitly considers co-assurance socio-technical factors.