Security-Aware Data-Driven QKD Routing and Admission Control in SDN-Controlled Quantum-Classical Optical Networks

Abstract Quantum key distribution (QKD) is gradually evolving from dedicated-link deployment toward networked integration with existing optical network infrastructures. In quantum–classical co-propagation scenarios, the quality of quantum links is readily affected by noise interference and device non-idealities, making efficient routing and resource allocation in multi-node networks a critical issue. To address this problem, this paper proposes a security-aware, data-driven QKD routing and admission control framework for software-defined networking (SDN)-controlled quantum–classical integrated optical networks. In the proposed framework, a variational quantum classifier (VQC) is employed as a link-state pre-screening mechanism, and the detection results are further used to support candidate path screening, routing decision-making, and resource allocation. In addition, the actual key generation capability of each path is further validated by incorporating the QKD protocol procedure into the decision process. Simulation results show that, compared with conventional baseline methods, the proposed method achieves stable performance under varying traffic loads, non-ideal conditions, and larger network scales, while yielding moderate improvements in system blocking ratio, average quantum signal-to-noise ratio (QSNR), and final key length.