MiRitH: Efficient Post-Quantum Signatures from MinRank in the Head
DOI:
https://doi.org/10.46586/tches.v2024.i2.304-328Keywords:
Digital Signature, MinRank, MPCitH, Post-Quantum, ZKPoK, Quantum AnalysisAbstract
Since 2016’s NIST call for standardization of post-quantum cryptographic primitives, developing efficient post-quantum secure digital signature schemes has become a highly active area of research. The difficulty in constructing such schemes is evidenced by NIST reopening the call in 2022 for digital signature schemes, because of missing diversity in existing proposals. In this work, we introduce the new postquantum digital signature scheme MiRitH. As direct successor of a scheme recently developed by Adj, Rivera-Zamarripa and Verbel (Africacrypt ’23), it is based on the hardness of the MinRank problem and follows the MPC-in-the-Head paradigm. We revisit the initial proposal, incorporate design-level improvements and provide more efficient parameter sets. We also provide the missing justification for the quantum security of all parameter sets following NIST metrics. In this context we design a novel Grover-amplified quantum search algorithm for solving the MinRank problem that outperforms a naive quantum brute-force search for the solution.
MiRitH obtains signatures of size 5.7 kB for NIST category I security and therefore competes for the smallest signatures among any post-quantum signature following the MPCitH paradigm.
At the same time MiRitH offers competitive signing and verification timings compared to the state of the art. To substantiate those claims we provide extensive implementations. This includes a reference implementation as well as optimized constant-time implementations for Intel processors (AVX2), and for the ARM (NEON) architecture. The speedup of our optimized AVX2 implementation relies mostly on a redesign of the finite field arithmetic, improving over existing implementations as well as an improved memory management.
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Gora Adj, Stefano Barbero, Emanuele Bellini, Andre Esser, Luis Rivera-Zamarripa, Carlo Sanna, Javier Verbel, Floyd Zweydinger
This work is licensed under a Creative Commons Attribution 4.0 International License.