This paper shows strong evidence that this efficiency gap cannot be bridged. Namely, whilst there are efficient protocols allowing a prover to cheat only with negligibly small probability in the case of known order groups, we provide strong evidence that for hidden order groups this probability is bounded below by $1/2$ for all efficient $\Sigma$-protocols not using common reference strings or the like.
We prove our results for a comprehensive class of $\Sigma$-protocols in the generic group model, and further strengthen them by investigating certain instantiations in the plain model.
Category / Keywords: cryptographic protocols / Generic Group Model; $\Sigma$-Protocols; Proofs of Knowledge; Error Bounds; Publication Info: Extended abstract appears in TCC 2010. Date: received 3 Dec 2009, last revised 4 Dec 2009 Contact author: stephan krenn at bfh ch Available formats: PDF | BibTeX Citation Version: 20091204:162333 (All versions of this report) Discussion forum: Show discussion | Start new discussion