Paper 2023/1774

Decentralized Private Steam Aggregation from Lattices

Uddipana Dowerah, Chalmers University of Technology
Aikaterini Mitrokotsa, University of St. Gallen
Abstract

As various industries and government agencies increasingly seek to build quantum computers, the development of post-quantum constructions for different primitives becomes crucial. Lattice-based cryptography is one of the top candidates for constructing quantum-resistant primitives. In this paper, we propose a decentralized Private Stream Aggregation (PSA) protocol based on the Learning with Errors (LWE) problem. PSA allows secure aggregation of time-series data over multiple users without compromising the privacy of the individual data. In almost all previous constructions, a trusted entity is used for the generation of keys. We consider a scenario where the users do not want to rely on a trusted authority. We, therefore, propose a decentralized PSA (DPSA) scheme where each user generates their own keys without the need for a trusted setup. We give a concrete construction based on the hardness of the LWE problem both in the random oracle model and in the standard model.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
Published elsewhere. ACNS 2024
Keywords
Private Stream AggregationDecentralizedLearning with Errors
Contact author(s)
dowerahuddipana @ gmail com
katerina mitrokotsa @ unisg ch
History
2023-11-17: approved
2023-11-16: received
See all versions
Short URL
https://ia.cr/2023/1774
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2023/1774,
      author = {Uddipana Dowerah and Aikaterini Mitrokotsa},
      title = {Decentralized Private Steam Aggregation from Lattices},
      howpublished = {Cryptology ePrint Archive, Paper 2023/1774},
      year = {2023},
      note = {\url{https://eprint.iacr.org/2023/1774}},
      url = {https://eprint.iacr.org/2023/1774}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.