Security of Identification Schemes

Adriana Palacio
Department of Computer Science and Engineering, University of California, San Diego

Abstract: An identification (ID) scheme enables a party holding a secret key to identify itself to another party without revealing the secret key. The former party is called the prover, and the latter is called the verifier. The main notion of security for ID schemes is security against impersonation under active attack. In such an attack, the adversary's goal is impersonation: playing the role of the prover, but denied the secret key, it attempts to make the verifier accept. Before the impersonation attempt, the adversary can play the role of cheating verifier, interacting with the prover numerous times to try to obtain information about the secret key.

The Guillou-Quisquater (GQ) and Schnorr ID schemes are among the most efficient and best known such schemes. However, the question of whether they are secure against impersonation under active attack had remained open for more than 10 years. In this talk I will discuss our results about the security properties of these schemes. We show that GQ is provably secure based on an assumption about the RSA function called RSA security under one more inversion. We also provide such a proof for the Schnorr scheme based on a corresponding discrete-logarithm-related assumption.