Isolated proofs of knowledge and isolated zero knowledge

  • Authors:
  • Ivan Damgård;Jesper Buus Nielsen;Daniel Wichs

  • Affiliations:
  • University of Aarhus, Denmark;University of Aarhus, Denmark;New York University

  • Venue:
  • EUROCRYPT'08 Proceedings of the theory and applications of cryptographic techniques 27th annual international conference on Advances in cryptology
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

We consider proof of knowledge protocols where the cheating prover may communicate with some external adversarial environment during the run of the proof. Without additional setup assumptions, no witness hiding protocol can securely ensure that the prover knows a witness in this scenario. This is because the prover may just be forwarding messages between the environment and the verifier while the environment performs all the necessary computation. In this paper we consider an l-isolated prover, which is restricted to exchanging at most l bits of information with its environment. We introduce a new notion called l-isolated proofs of knowledge (l-IPoK). These protocols securely ensure that an l-isolated prover knows the witness. To prevent the above-mentioned attack, an l-IPoK protocol has to have communication complexity greater than l. We show that for any relation in NP and any value l, there is an l-IPoK protocol for that relation. In addition, the communication complexity of such a protocol only needs to be larger than l by a constant multiplicative factor.