Back to ProVerif
Bruno Blanchet. Automatic Proof of Strong Secrecy for Security Protocols. In Dagstuhl seminar Language-Based Security, October 2003. Tool feature


Seminar homepage:


We present a new automatic technique for proving strong secrecy for security protocols. Strong secrecy means that an adversary cannot see any difference when the value of the secret changes. Strong secrecy detects implicit flows and flows of partial information. It is a particular case of observational equivalence, which gives nice compositionality properties and makes it easier to combine manual and automatic proofs. Our technique relies on an automatic translation of the protocol into Horn clauses, and a resolution algorithm on the clauses. It requires important extensions with respect to previous work for (syntactic) secrecy and authenticity. This technique can handle a wide range of cryptographic primitives, and yields proofs valid for an unbounded number of sessions and an unbounded message space; it is also flexible and efficient. We have proved its correctness, its termination on the large class of tagged protocols, and implemented it.


  AUTHOR = {Bruno Blanchet},
  TITLE = {Automatic {P}roof of {S}trong {S}ecrecy for {S}ecurity {P}rotocols},
  BOOKTITLE = {Dagstuhl seminar "Language-Based Security"},
  YEAR = 2003,