Publications

Detailed Information

Faster TFHE Bootstrapping with Block Binary Keys

Cited 0 time in Web of Science Cited 0 time in Scopus
Authors

Lee, Changmin; Min, Seonhong; Seo, Jinyeong; Song, Yongsoo

Issue Date
2023
Publisher
Association for Computing Machinery (ACM)
Citation
Proceedings of the ACM Conference on Computer and Communications Security, pp.2-13
Abstract
Fully Homomorphic Encryption over the Torus (TFHE) is a homomorphic encryption scheme which supports efficient Boolean operations over encrypted bits. TFHE has a unique feature in that the evaluation of each binary gate is followed by a bootstrapping procedure to refresh the noise of a ciphertext. In particular, this gate bootstrapping involves two algorithms called the blind rotation and key-switching. In this work, we introduce several optimization techniques for the TFHE bootstrapping. We first define a new key distribution, called the block binary distribution, where the secret key can be expressed as a concatenation of several vectors of Hamming weight at most one. We analyze the hardness of (Ring) LWE with a block binary secret and provide candidate parameter sets which are secure against the best-known attacks. Then, we use the block key structure to simplify the inner working of blind rotation and reduce its complexity. We also modify the RLWE key generation and the gadget decomposition method to improve the performance of the key-switching algorithm in terms of complexity and noise growth. Finally, we use the TFHE library to implement our algorithms and demonstrate their benchmarks. Our experimentation shows that the execution time of TFHE bootstrapping is reduced from 10.5ms down to 6.4ms under the same security level, and the size of the bootstrapping key decreases from 109MB to 60MB.
ISSN
1543-7221
URI
https://hdl.handle.net/10371/201192
DOI
https://doi.org/10.1145/3579856.3595804
Files in This Item:
There are no files associated with this item.
Appears in Collections:

Related Researcher

  • College of Engineering
  • Dept. of Computer Science and Engineering
Research Area Cryptography, Privacy, Security

Altmetrics

Item View & Download Count

  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.

Share