Using security robustness analysis for early-stage validation of functional security requirements

Security is nowadays an indispensable requirement in software systems. Traditional software engineering processes focus primarily on business requirements, leaving security as an afterthought to be addressed via generic "patched-on" defensive mechanisms. This approach is insufficient, and software systems need to have security functionality engineered within in a similar fashion as ordinary business functional requirements. Functional security requirements need to be elicited, analyzed, specified and validated at the early stages of the development life cycle. If the functional security requirements were not properly validated, then there is a risk of developing a system that is insecure, deeming it unusable. Acceptance testing is an effective technique to validate requirements. However, an ad hoc approach to develop acceptance tests will suffer the omission of important tests. This paper presents a systematic approach to develop executable acceptance tests that is specifically geared for model-based secure software engineering processes. The approach utilizes early-stage artifacts, namely misuse case and domain models, and robustness diagrams. The feasibility of the proposed approach is demonstrated by applying it to a real-world system. The results show that a comprehensive set of security acceptance tests can be developed based upon misuse case models for early-stage validation of functional security requirements.