Quantum identity authentication without entanglement

Abstract

An interesting protocol for quantum identity authentication on a basis of the classic shared secret has been presented recently (Hong et al. in Quantum Inf Process 16(10): 236, 2017). It requires few resources and it is technologically feasible. Its seminal analysis focuses on average eavesdropper’s information gain per single protocol run—a parameter suitable for estimation of the security margin applicable to the protocols aiming at provision of confidentiality. However, the security requirements of the authentication process are very stringent—no single bit of the secret can be revealed, otherwise the eavesdropper can collect subsequent bits of the secret in successive protocol runs. The failure of the seminal version to meet this requirement is demonstrated. The sequential processing of qubits and the static nature of the compared data is the rationale behind the improvement. The introduced changes make that partial information gained in some authentication attempt gives the eavesdroppers no advantage in breaking the next ones. They are forced to consider each authentication transaction as a separate puzzle that can be solved according to all-or-nothing paradigm. The improvement retains conceptual simplicity and technological feasibility of the seminal version of the protocol.