Model-Based Run-Time Checking of Security Permissions Using Guarded Objects

In this paper we deal with the application of run-time checking to enforce requirements which, because of their nature, cannot be enforced statically. More specifically, it deals with the problem how to control access to objects within an object-oriented system at run-time in a way that enforces an overall security policy. It aims to improve on the ad-hoc (and often untrustworthy) way it is currently done in practice by automatically generating the run-time checks from a model-based specification of the system that captures the security policy. Concretely, the models are expressed in the UML security extension UMLsec, and the run-time checks that are generated for Java programs rely on Guarded Objects.