Cross-Layer Verification of Type Flaw Attacks on Security Protocols

Abstract

Security protocols are often specified at the application layer; however, application layer specifications give little detail regarding message data structures at the presentation layer upon which some implementation-dependent attacks rely. In this paper we present an approach to verifying security protocols in which both the application and presentation layers are modelled. Using the Group Domain of Interpretation protocol as an example, our application layer specification of the protocol is used as input to the AVISPA model checking tool for analysis. Two type flaw attacks are found via model checking which are then verified against the corresponding presentation layer specification, thus identifying the minimal requirements to prevent the attacks.