Schnorr-Like Identification Scheme Resistant to Malicious Subliminal Setting of Ephemeral Secret

In this paper we propose a modification of the Schnorr Identification Scheme (\({\mathsf {IS}}\)), which is immune to malicious subliminal setting of ephemeral secret. We introduce a new strong security model in which, during the query stage, we allow the adversary verifier to set random values used on the prover side in the commitment phase. We define the \({\mathsf {IS}}\) scheme to be secure if such a setting will not enable the adversary to impersonate the prover later on. Subsequently we prove the security of the modified Schnorr \({\mathsf {IS}}\) in our strong model. We assume the proposition is important for scenarios in which we do not control the production process of the device on which the scheme is implemented, and where the erroneous pseudo-random number generators make such attacks possible.