Run-Time Enforcement of Nonsafety Policies

Reviewer: Bernard Kuc

I have always viewed digital security and protection as more of an art than a science, the art of staying only one step behind the ingenuity of the next kid trying to break into some system or release that small bit of code that will bring the Internet to its knees. The authors look at the enforcement of security policies from a theoretical perspective. Although this is a common theme of much research, their contribution is that of looking at nonsafety policies. Safety policies ensure that nothing bad happens. After providing a few definitions of property and policy, and how security automata work, the authors detail what sorts of policies can be enforced by monitors. The authors show that while truncation automata can only enforce safety properties, edit automata-automata that can modify the actions that are to be executed-can enforce a wider set of properties, such as liveness and renewal. Each of the above terms is clearly defined and explained using simple examples. Finally, the authors look at the constraints and caveats that limit the applicability of the research, followed by a quick look at alternative ways of defining security automata and even a description of a Java-based implementation. A security system is only as secure as its weakest link, and the phrase, "under the following assumptions," is tantamount to admitting defeat. However, any formal analysis or proof is usually only valid once certain assumptions have been made. Does this mean we should abandon theoretical research__?__ No. Each step that brings more science to the art should always be welcomed. Online Computing Reviews Service