Security policy testing via automated program code generation

Ting Yu,Dhivya Sivasubramanian,Tao Xie
DOI: https://doi.org/10.1145/1558607.1558623
2009-01-01
Abstract:Access control is one of the fundamental security mechanisms for information systems. It determines the availability of resources to principals, operations that can be performed, and under what circumstances. Traditionally the enforcement of access control is often hardcoded in applications or systems; such hardcoding makes it hard to verify the correctness of access control and to accommodate changes of security requirements. Recently, access control policies have been increasingly separated from enforcement mechanisms. An access control policy is explicitly specified using certain policy languages with well-defined syntax and semantics. An application then consults the policy during runtime to determine whether an access request from a principal should be allowed or denied. There are two main advantages of this approach. First, security officers can now perform systematic and formal security analysis on access control policies. Second, by separating policies from enforcement mechanisms, it is possible to change policies without affecting the underlying mechanisms, and vice versa.
What problem does this paper attempt to address?