Towards Automated Security Design Flaw Detection

Abstract

Efficiency of security-by-design has become an important goal for organizations implementing software engineering practices such as Agile, DevOps, and Continuous Integration. Software architectures are (often manually) analyzed at design time for potential security design flaws, based on natural language descriptions of security weaknesses (e.g., CWE, CAPEC). The use of natural language hinders the application of such knowledge bases in an automated fashion. In this paper, we analyze an existing catalog of 19 security design flaws in order to identify conceptual, technology-independent requirements on architectural models that enable automatically detecting these flaws. This constitutes the first step towards automated assessment of design-level security. Our findings are illustrated on an IoT-based smart home system.