Finite state transducers for policy evaluation and conflict resolution

Abstract

This work presents a formal framework for a policy representation model based on finite state transducers (FSTs). The main motivation for this approach is to produce the formalities to represent the body of policies in an unambiguous way. Using well-known entities such as finite state machines, we propose interpretations and adaptations to the basic theory to fit in the domain of policy based management. Policies are modeled as finite state transducers that consume events, and a function we call a tautness function is defined to be on the transitions. The operations of finite state transducers are revised accordingly. In particular, we present determinization and intersection operations for FSTs that will mimic the modality conflict resolution process between policies. We also demonstrate how the composition of transducers could be used to express constraints or meta-policies. In our approach, all the tasks associated with the conflict resolution process can be done "a priori," and the computing of a policy-evaluation is linear in the number of events, and independent of the number of policies.