Quantum Key Distribution Contingency in the Absence of the Classical Channel

It is an accepted paradigm in the already matured industry of Quantum Key Distribution (QKD) implementations that when the quantum channel is attacked or unresponsive, the system reverts to classical security. Thus, in times of crises, when the quantum system is severely damaged, the saving resort is considered to be the classical solution. This paper explores the opposite approach. In the case of disaster, when parts of the classical part of the key distribution system are broken, are there any possible crisis management options to give some limited functionality? The result of this research shows that if the classical channel fails, the quantum channel can still produce and distribute keys. The experimental results of the contingency QKD show that, using positive operator-valued measurements (POVMs), keys can still be produced and shared. The scheme described in this paper uses the quantum channel only to distribute imperfect keys. Any one distributed key has a theoretical overlap of approximately $75\%$ between Alice’s key and Bob’s key, respectively. The experimental POVM circuit is implemented with two different Naimark dilation approximations: one using $R_z$ gates and the other using $R_y$ gates. The practical implementation results are close to the theoretical analysis. As the keys have a partial overlap, the encryption/ decryption algorithm also needs to adjust to this reality. The encryption/decryption algorithm used in the experiments is a repetition algorithm that is simple but shows the resilience of the scheme. Ultimately, the classical channel is not used during the contingency QKD at all, while the quantum channel is assumed trusted under a restricted adversary model in which Eve is limited to individual attacks. Under this model, partial secrecy is retained for all non-zero channel error rates below a pre-agreed threshold.