[Resource Topic] 2024/909: Approximate CRT-Based Gadget Decomposition and Application to TFHE Blind Rotation

Welcome to the resource topic for 2024/909

Title:
Approximate CRT-Based Gadget Decomposition and Application to TFHE Blind Rotation

Authors: Olivier Bernard, Marc Joye

Abstract:

One of the main issues to deal with for fully homomorphic encryption is the noise growth when operating on ciphertexts. To some extent, this can be controlled thanks to a so-called gadget decomposition. A gadget decomposition typically relies on radix- or CRT-based representations to split elements as vectors of smaller chunks whose inner products with the corresponding gadget vector rebuilds (an approximation of) the original elements. Radix-based gadget decompositions present the advantage of also supporting the approximate setting: for most homomorphic operations, this has a minor impact on the noise propagation but leads to substantial savings in bandwidth, memory requirements and computational costs. A typical use-case is the blind rotation as used for example in the bootstrapping of the TFHE scheme. On the other hand, CRT-based representations are convenient when machine words are too small for directly accommodating the arithmetic on large operands. This arises in two typical cases: (i) in the hardware case with multipliers of restricted size, e.g., 17 bits; (ii) in the software case for ciphertext moduli above, e.g., 128 bits.

This paper presents new CRT-based gadget decompositions for the approximate setting, which combines the advantages of non-exact decompositions with those of CRT-based decompositions. Significantly, it enables certain hardware or software realizations otherwise hardly supported like the two aforementioned cases. In particular, we show that our new gadget decompositions provide implementations of the (programmable) bootstrapping in TFHE relying solely on native arithmetic and offering extra degrees of parallelism.

ePrint: https://eprint.iacr.org/2024/909

See all topics related to this paper.

Feel free to post resources that are related to this paper below.

Example resources include: implementations, explanation materials, talks, slides, links to previous discussions on other websites.

For more information, see the rules for Resource Topics .