Projet ACI Sécurité Informatique : 2003 - 2006 Rapport

1 General Description ACI Rossignol is composed of teams from LIF (UMR 6166), LSV (UMR 8643 CNRS ENS Cachan and INRIA Futurs-project SECSI), INRIA Futurs (project Comète) and Verimag (UMR 9939). The Rossignol project aims at investigating issues involved in the verification of cryptographic protocols. The focus has been put on research directions that go beyond the classical approach devised for far, known as the Dolev-Yao model, which has been extensively studied in the last years. The contributions of the project cover the following topics : – Extensions of the Dolev-Yao model that retains the basics of this formalization, but modify the model of the insecure environment (which we refer below as the intruder capabilities) to reflect more closely the real world. This can be done by adding or modifying axioms that express the properties of the operations used by the protocols or to express weak secret that can be broken by guessing attacks. – To deal with other security properties than secrecy or authentication, like anonymity or opacity that are used for instance in e-voting protocols. – Models of probabilistic protocols in a variant of π-calculus and testing the model against real protocols. To devise sound and convincing models for this class of processes is a difficult task that leads to subtle errors when this is not handled properly. – Relationship between the formal world and the computational world. While the formal approach leads to a clean trace semantics and automated verification, the approach is questionned by the computational community. Therefore works linking the two approaches provide a better understanding of the issue and can make the formal approach more convincing.