No abstract

FALCON is a very efficient and compact lattice-based signature finalist of the NIST’s Post-Quantum standardization campaign. This work assesses Falcon’s sidechannel resistance by analyzing two vulnerabilities, namely the pre-image computation and the trapdoor sampling. The first attack is an improvement of Karabulut and Aysu (DAC 2021). It overcomes several difficulties inherent to the structure of the stored key like the Fourier representation and directly recovers the key with a limited number of traces and a reduced complexity. The main part of this paper is dedicated to our second attack: we show that a simple power analysis during the signature execution could provide the exact value of the output of a subroutine called the base sampler. This intermediate value does not directly lead to the secret and we had toadapt the so-called hidden parallelepiped attack initially introduced by Nguyen and Regev in Eurocrypt 2006 and reused by Ducas and Nguyen in Asiacrypt 2012. We extensively quantify the resources for our attacks and experimentally demonstrate them with FALCON’s reference implementation on the ELMO simulator (McCann, Oswald and Whitnall USENIX 2017) and on a ChipWhisperer Lite with STM32F3 target (ARM Cortex M4).These new attacks highlight the need for side-channel protection for one of the three finalists of NIST’s standardization campaign by pointing out the vulnerable parts and quantifying the resources of the attacks.

No abstract

Non-Centered Discrete Gaussian sampling is a fundamental building block in many lattice-based constructions in cryptography, such as signature and identity-based encryption schemes. On the one hand, the center-dependent approaches, e.g. cumulative distribution tables (CDT), Knuth-Yao, the alias method, discrete Zigurat and their variants, are the fastest known algorithms to sample from a discrete Gaussian distribution. However, they use a relatively large precomputed table for each possible real center in [0, 1) making them impracticable for non-centered discrete Gaussian sampling. On the other hand, rejection sampling allows to sample from a discrete Gaussian distribution for all real centers without prohibitive precomputation cost but needs costly floating-point arithmetic and several trials per sample. In this work, we study how to reduce the number of centers for which we have to precompute tables and propose a non-centered CDT algorithm with practicable size of precomputed tables as fast as its centered variant. Finally, we provide some experimental results for our open-source C++ implementation indicating that our sampler increases the rate of Peikert's algorithm for sampling from arbitrary lattices (and cosets) by a factor 3 with precomputation storage up to 6.2 MB.M.R.

No abstract