Méthodes pour la vérification des protocoles cryptographiques dans le modèle calculatoire

摘要

Critical and private information are exchanged on public environment. To protect it from dishonest users, we use cryptographic tools. Unfortunately, bad conception, poorly written security properties and required security hypothesis lead to attacks, and it may take years before one discover the attack and fix the security schemes involved. In this context, provable security provides formal definitions for security objectives and implied mathematical proofs that these objectives are fullfilled. On another hand, complexity and variety of cryptographic systems are increasing, and proofs by hand are too complicated to write and to verify (Bellare& Rogaway 2004, Shoup 2004, Halevi 2005). Thus, we need computer-assisted verification methods for cryptographic systems. The aim of this thesis is to progress in this direction. More precisely we want significant progress over formal proofs on cryptographic protocols. To verify cryptographic protocols we need to develop a theoritical framework providing: - a precise modelisation for cryptographic protocols and security properties we want to prove in the computationnal model, - designing tactics to automate proofs, - taking into account realistic models for adversary (side-channels...). By the end of the thesis we have enhanced a theoretical framework and computing tools helping verifying cryptographic protocols.