Paper 2025/110

Verification-efficient Homomorphic Signatures for Verifiable Computation over Data Streams

Gaspard Anthoine, IMDEA Software Institute, Universidad Politécnica de Madrid
Daniele Cozzo, IMDEA Software Institute
Dario Fiore, IMDEA Software Institute
Abstract

Homomorphic signatures for NP (HSNP) allow proving that a signed value is the result of a non-deterministic computation on signed inputs. At CCS'22, Fiore and Tucker introduced HSNP, showed how to use them for verifying arbitrary computations on data streams, and proposed a generic HSNP construction obtained by efficiently combining zkSNARKs with linearly homomorphic signatures (LHS), namely those supporting linear functions. Their proposed LHS however suffered from an high verification cost. In this work we propose an efficient LHS that significantly improves on previous work in terms of verification time. Using the modular approach of Fiore and Tucker, this yields a verifier-efficient HSNP. We show that the HSNP instantiated with our LHS is particularly suited to the case when the data is taken from consecutive samples, which captures important use cases including sliding window statistics such as variances, histograms and stock market predictions.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
Published elsewhere. Major revision. Financial Cryptography and Data Security
Keywords
Verifiable Computationhomomorphic signatureszero-knowledge proofsSNARKsData StreamsData Privacy
Contact author(s)
gaspard anthoine @ imdea org
daniele cozzo @ imdea org
dario fiore @ imdea org
History
2025-01-24: approved
2025-01-23: received
See all versions
Short URL
https://ia.cr/2025/110
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2025/110,
      author = {Gaspard Anthoine and Daniele Cozzo and Dario Fiore},
      title = {Verification-efficient Homomorphic Signatures for Verifiable Computation over Data Streams},
      howpublished = {Cryptology {ePrint} Archive, Paper 2025/110},
      year = {2025},
      url = {https://eprint.iacr.org/2025/110}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.