Access control to objects in common object-oriented languages is statically verified but cannot be changed at run-time. However, dynamic authorization is required by most applications and it would be desirable to check more flexible access control policies also statically, at least partially. In this work, we introduce a model where "views" to object references represent the current access control policy of a principal for a given object, and first class authorizations support dynamic modification of those policies. To demonstrate our concepts, we have developed a core language, equipped with a provably correct type and effect system capable of detecting unauthorized method calls at compile-time, and defined and implemented a typechecking algorithm.
|Title of host publication||Lecture Notes in Computer Science|
|Publication status||Published - 1 Jan 2010|
|Event||Automated Reasoning for Security Protocol Analysis and Issues in the Theory of Security (ARSPA-WITS) - |
Duration: 1 Jan 2010 → …
|Conference||Automated Reasoning for Security Protocol Analysis and Issues in the Theory of Security (ARSPA-WITS)|
|Period||1/01/10 → …|