Multisignatures secure under the discrete logarithm assumption and a generalized forking lemma

  • Authors:
  • Ali Bagherzandi;Jung-Hee Cheon;Stanislaw Jarecki

  • Affiliations:
  • University of California at Irvine, Irvine, CA, USA;Seoul National University, Seoul, South Korea;University of California at Irvine, Irvine, CA, USA

  • Venue:
  • Proceedings of the 15th ACM conference on Computer and communications security
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Multisignatures allow n signers to produce a short joint signature on a single message. Multisignatures were achieved in the plain model with a non-interactive protocol in groups with bilinear maps, by Boneh et al, and by a three-round protocol under the Discrete Logarithm (DL) assumption, by Bellare and Neven, with multisignature verification cost of, respectively, O(n) pairings or exponentiations. In addition, multisignatures with O(1) verification were shown in so-called Key Verification (KV) model, where each public key is accompanied by a short proof of well-formedness, again either with a non-interactive protocol using bilinear maps, by Ristenpart and Yilek, or with a three-round protocol under the Diffie-Hellman assumption, by Bagherzandi and Jarecki. We improve on these results in two ways: First, we show a two-round O(n)-verification multisignature secure under the DL assumption in the plain model, improving on the three-round protocol of Bellare-Neven. Second, we show a two-round O(1)-verification multisignature secure under the DL assumption in the KV model, improving on assumptions and/or communication rounds of the schemes of Ristenpart and Yilek and Bagherzandi and Jarecki. Exact security of both schemes matches (in ROM) that of Schnorr signatures. The reduced round complexity is due to a new multiplicatively homomorphic equivocable commitment scheme which can be of independent interest. Moreover, our KV model scheme is enabled by a generalized forking lemma, which shows that standard non-interactive zero-knowledge (NIZK) proofs of knowledge in ROM admit efficient simultaneous post-execution extraction of witnesses of all proof instances. As a consequence of this lemma, any DL-based multisignature secure in so-called Knowledge-of-Secret-Key model can be implemented in the KV model using standard ROM-based NIZK's of DL as proofs of key well-formedness.