Résumé: The Java Security Architecture includes a dynamic mechanism for enforcing access control checks, the so-called stack inspection process. While the architecture has several appealing features, access control checks are all implemented via dynamic method calls. This is a highly nondeclarative form of specification that is hard to read, and that leads to additional run-time overhead. This article develops type systems that can statically guarantee the success of these checks. Our systems allow security properties of programs to be clearly expressed within the types themselves, which thus serve as static declarations of the security policy. We develop these systems using a systematic methodology: we show that the security-passing style translation, proposed by Wallach et al. [2000] as a dynamic implementation technique, also gives rise to static security-aware type systems, by composition with conventional type systems. To define the latter, we use the general HM(X) framework, and easily construct several constraint- and unification-based type systems.

Langue: Anglais