Extending access control models with break-glass

Abstract

Access control models are usually static, i.e, permissions are granted based on a policy that only changes seldom. Especially for scenarios in health care and disaster management, a more flexible support of access control, i.e., the underlying policy, is needed.Break-glass is one approach for such a flexible support of policies which helps to prevent system stagnation that could harm lives or otherwise result in losses. Today, break-glass techniques are usually added on top of standard access control solutions in an ad-hoc manner and, therefore, lack an integration into the underlying access control paradigm and the systems' access control enforcement architecture.We present an approach for integrating, in a fine-grained manner, break-glass strategies into standard access control models and their accompanying enforcement architecture. This integration provides means for specifying break-glass policies precisely and supporting model-driven development techniques based on such policies.