Cryptology ePrint Archive: Report 2013/281

Adapting Lyubashevsky’s Signature Schemes to the Ring Signature Setting

Carlos Aguilar-Melchor and Slim Bettaieb and Xavier Boyen and Laurent Fousse and Philippe Gaborit

Abstract: Basing signature schemes on strong lattice problems has been a long standing open issue. Today, two families of lattice-based signature schemes are known: the ones based on the hash-and-sign construction of Gentry et al.; and Lyubashevsky’s schemes, which are based on the Fiat-Shamir framework. In this paper we show for the first time how to adapt the schemes of Lyubashevsky to the ring signature setting. In particular we transform the scheme of ASIACRYPT 2009 into a ring signature scheme that provides strong properties of security under the random oracle model. Anonymity is ensured in the sense that signatures of different users are within negligible statistical distance even under full key exposure. In fact, the scheme satisfies a notion which is stronger than the classical full key exposure setting as even if the keypair of the signing user is adversarially chosen, the statistical distance between signatures of different users remains negligible. Considering unforgeability, the best lattice-based ring signature schemes provide either unforgeability against arbitrary chosen subring attacks or insider corruption in log-sized rings. In this paper we present two variants of our scheme. In the basic one, unforgeability is ensured in those two settings. Increasing signature and key sizes by a factor k (typically 80 − 100), we provide a variant in which unforgeability is ensured against insider corruption attacks for arbitrary rings. The technique used is pretty general and can be adapted to other existing schemes.

Category / Keywords: public-key cryptography / Ring signatures, lattices

Date: received 14 May 2013

Contact author: carlos aguilar at unilim fr

Available format(s): PDF | BibTeX Citation

Version: 20130516:133120 (All versions of this report)

Discussion forum: Show discussion | Start new discussion


[ Cryptology ePrint archive ]