Title
Fast Fourier Orthogonalization over NTRU Lattices
Abstract
FALCON is an efficient and compact lattice-based signature scheme. It is also one of the round 3 finalists in the NIST PQC standardization process. The core of FALCON is a trapdoor sampling algorithm, which has found numerous applications in lattice-based cryptography. It needs the fast Fourier orthogonalization algorithm to build an LDL tree. But the LDL tree needs much RAM to store, which may limit the application of FALCON on memory-constrained devices. On the other hand, if building the LDL tree dynamically, the signature cost will almost double. In this work, we discover the LDL tree of FALCON has some symmetric structure, and prove why this phenomenon occurs. With this property, we can reduce the generation time and storage of the LDL tree by almost half without affecting the efficiency of FALCON. We verify the correctness and validity of our way in the implementations of FALCON. In addition, the result applies to the cyclotomic field $$\mathbb {Q}[x]/(x^n-x^{n/2}+1)$$ with $$n = 3 \cdot 2 ^\kappa $$ . But we can not apply it to NTRU module lattices so far.
Year
DOI
Venue
2022
10.1007/978-3-031-15777-6_7
Information and Communications Security
Keywords
DocType
ISSN
fast Fourier orthogonalization, lattice-based cryptography, NTRU, FALCON, trapdoor sampling
Conference
0302-9743
Citations 
PageRank 
References 
0
0.34
0
Authors
6
Name
Order
Citations
PageRank
Sun Shuo100.34
Zhou Yongbin200.34
Rui Zhang375.87
Tao Yang400.34
Qiao Zehua500.34
Ming Jingdian600.34