Paper 2022/1082

Assisted Private Information Retrieval

Natnatee Dokmai, Indiana University Bloomington
L. Jean Camp, Indiana University Bloomington
Ryan Henry, University of Calgary
Abstract

Private Information Retrieval (PIR) addresses the cryptographic problem of hiding sensitive database queries from database operators. In practice, PIR schemes face the challenges of either high computational costs or restrictive security assumptions, resulting in a barrier to deployment. In this work, we introduce Assisted Private Information Retrieval (APIR), a new PIR framework for keyword-value databases generalizing multi-server PIR and relaxing its database consistency assumption. We propose the construction of Synchronized APIR, an efficient hybrid APIR scheme combining black-box single-server PIR and non-black-box multi-server PIR. To evaluate the scheme, we apply it to a proof-of-concept privacy-preserving DNS application. The experiment results demonstrate that Synchronized APIR outperforms the baseline single-server PIR protocol in communication and computational cost after the initial one-time cost.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
Preprint.
Keywords
private information retrievalDNS
Contact author(s)
natnatee dokmai @ gmail com
ljeanc @ gmail com
ryan henry @ ucalgary ca
History
2023-03-17: last of 2 revisions
2022-08-19: received
See all versions
Short URL
https://ia.cr/2022/1082
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2022/1082,
      author = {Natnatee Dokmai and L. Jean Camp and Ryan Henry},
      title = {Assisted Private Information Retrieval},
      howpublished = {Cryptology ePrint Archive, Paper 2022/1082},
      year = {2022},
      note = {\url{https://eprint.iacr.org/2022/1082}},
      url = {https://eprint.iacr.org/2022/1082}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.