Polytopes in the Fiat-Shamir with Aborts Paradigm - Archive ouverte HAL
Proceedings/Recueil Des Communications Lecture Notes in Computer Science Année : 2024

Polytopes in the Fiat-Shamir with Aborts Paradigm

Résumé

The Fiat-Shamir with Aborts paradigm (FSwA) uses rejection sampling to remove a secret’s dependency on a given source distribution. Recent results revealed that unlike the uniform distribution in the hypercube, both the continuous Gaussian and the uniform distribution within the hypersphere minimise the rejection rate and the size of the proof of knowledge. However, in practice both these distributions suffer from the complexity of their sampler. So far, those three distributions are the only available alternatives, but none of them offer the best of all worlds: competitive proof of knowledge size and rejection rate with a simple sampler. We introduce a new generic framework for FSwA using polytope based rejection sampling to enable a wider variety of constructions. As a matter of fact, this framework is the first to generalise these results to integral distributions. To complement the lack of alternatives, we also propose a new polytope construction, whose uniform sampler approaches in simplicity that of the hypercube. At the same time, it provides competitive proof of knowledge size compared to that obtained from the Gaussian distribution. Concurrently, we share some experimental improvements of our construction to further reduce the proof size. Finally, we propose a signature based on the FSwA paradigm using both our framework and construction. We prove it to be competitive with Haetae in signature size and with Dilithium on sampler simplicity.
Fichier non déposé

Dates et versions

hal-04688010 , version 1 (04-09-2024)

Identifiants

Citer

Henry Bambury, Hugo Beguinet, Thomas Ricosset, Éric Sageloli. Polytopes in the Fiat-Shamir with Aborts Paradigm. Advances in Cryptology – CRYPTO 2024, Lecture Notes in Computer Science, 14920, pp.339-372, 2024, Lecture Notes in Computer Science, 978-3-031-68376-3. ⟨10.1007/978-3-031-68376-3_11⟩. ⟨hal-04688010⟩
18 Consultations
0 Téléchargements

Altmetric

Partager

More