Search Results

The research area of cryptographic protocol analysis contains a number of innovative algorithms and procedures for checking various security properties of protocols. Most of these procedures focus on solving one of several "knowledge problems" that model intruder knowledge. Solving these problems can demonstrate the ability of the intruder to obtain some forbidden information of the protocol, such as secret keys. Two important examples of these problems are the deduction problem and the static equivalence problem. Deduction is concerned with the ability to derive a term from a set of terms (or knowledge) obtained from the observation of a protocol instance. Static equivalence, on the other hand, is concerned with distinguishing between two runs of a protocol based on two sets of knowledge. These two knowledge problems at first inspection appear to be very close to the older automated reasoning problems of matching and unification. However, this first impression is wrong, and there have been a few results that have shown theories where one problem, such as unification, is undecidable but another problem, such as deduction, is decidable. These existing dichotomy results were, however, incomplete, and not all cases had been examined, thus leaving the possibility of some connection between the problems for those unexamined cases. In this paper, we consider the missing dichotomy cases. For each of the remaining cases, we demonstrate a theory that separates the two problems. In addition, once the dichotomy results are completed, it leaves open the question of the existence of non-trivial classes of theories for which all four of the problems are decidable. One example for which this is true is the well-known class of subterm convergent term rewrite systems. In this paper, we develop another example, a class of restrictive permutative theories for which all problems are likewise decidable.