Automatic Support of the Generation and Maintenance of Assurance Cases

Abstract

One of the challenges in developing safety critical systems is to ensure software assurance which encompasses quality attributes such as reliability and security as well as functionality and performance. An assurance case, which lays out an argumentation-structure with supporting evidence to claim that software assurance in a system is achieved, is increasingly considered as an important means to gain confidence that a system has achieved acceptable safety when checking with emerging standards and national guidelines. However, the complexity of modern safety critical applications hinders the automatic integration of heterogeneous artifacts into an assurance case during a development process such as a V-model, let alone the automatic support of system evolution. In this paper, we present a novel framework to automatically generate assurance cases via safety patterns and further support the maintenance of them during a system’s evolution. The application of safety patterns not only enables reusability of previously successful argument structures but also directs the support of assurance maintenance caused by common types of modifications in safety critical domains. The framework is implemented as a prototypical tool built using Model Driven Architecture (MDA). We evaluated the framework with two case studies featuring two criteria and the preliminary experimental results not only show that the framework is useful in evaluation of safety critical systems but also reveal how different types of modification can affect a structure of an assurance case.