Using Abstract Interpretation for the Safe Verification of Security Protocols

AbstractIn this paper we extend the results proposed in [5] so as to provide an approach for finding and using a safe abstraction for the verification of general security properties. Similar abstractions based on abstract interpretation techniques have been developed for the verification of temporal properties expressed using various branching-time temporal logics (eg. [10,8,15,9,13,14,19]). Here we transpose some of the results of [8,14,19] to the verification of security properties. We also automate the construction of the abstract model and the translation of security properties into abstract ones for a large class of practical situations. As opposed to other uses of abstraction which typically guarantee the preservation of a whole logic or of a whole class of properties, here a specific abstraction function is selected for each given property and is thus only to guarantee the property at hand. The requirements are consequently much less demanding and the model reduction can be much more important. The proposed approach is currently being applied for the verification of large electronic commerce protocols