Model-based validation of safety-critical embedded systems

Safety-critical systems have become increasingly software reliant and the current development process of ¿build, then integrate¿ has become unaffordable. This paper examines two major contributors to today's exponential growth in cost: system-level faults that are not discovered until late in the development process; and multiple truths of analysis results when predicting system properties through model-based analysis and validating them against system implementations. We discuss the root causes of such system-level problems, and an architecture-centric model-based analysis approach of different operational quality aspects from an architecture model. A key technology is the SAE Architecture Analysis & Design Language (AADL) standard for embedded software-reliant system. It supports a single source approach to analysis of operational qualities such as responsiveness, safety-criticality, security, and reliability through model annotations. The paper concludes with a summary of an industrial case study that demonstrates the feasibility of this approach.