Symbolic Methods in Computational Cryptography Proofs - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year :

Symbolic Methods in Computational Cryptography Proofs

Abstract

Code-based game-playing is a popular methodology for proving security of cryptographic constructions and side-channel countermeasures. This methodology relies on treating cryptographic proofs as an instance of relational program verification (between probabilistic programs), and decomposing the latter into a series of elementary relational program verification steps. In this paper, we develop principled methods for proving such elementary steps for probabilistic programs that operate over finite fields and related algebraic structures. We focus on three essential properties: program equivalence, information flow, and uniformity. We give characterizations of these properties based on deducibility and other notions from symbolic cryptography. We use (sometimes improve) tools from symbolic cryptography to obtain decision procedures or sound proof methods for program equivalence, information flow, and uniformity. Finally, we evaluate our approach using examples drawn from provable security and from side-channel analysis-for the latter, we focus on the masking countermeasure against differential power analysis. A partial implementation of our approach is integrated in EASYCRYPT, a proof assistant for provable security, and in MASKVERIF, a fully automated prover for masked implementations.
Fichier principal
Vignette du fichier
paper.pdf (475.23 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-02404701 , version 1 (11-12-2019)

Identifiers

Cite

Gilles Barthe, Benjamin Grégoire, Charlie Jacomme, Steve Kremer, Pierre-Yves Strub. Symbolic Methods in Computational Cryptography Proofs. CSF2019 - 32nd IEEE Computer Security Foundations Symposium, Jun 2019, Hoboken, United States. pp.136-13615, ⟨10.1109/CSF.2019.00017⟩. ⟨hal-02404701⟩
330 View
661 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More