Summary

A quantum key distribution network (QKDN) is expected to maintain stable operations and meet the requirements of various cryptographic applications efficiently. Due to the advantages of machine learning (ML) related to autonomous learning, it can help to overcome the challenges of QKDN in terms of quantum layer performances, key management layer performances and QKDN control and management efficiency. Based on the functional requirements and architecture of QKDN stated in Recommendations ITU-T Y.3801 and ITU-T Y.3802, this Recommendation specifies one possible set of functional requirements and a possible architecture for an ML-enabled QKDN (QKDNml), including an overview and the functional requirements, architecture and operational procedures of QKDNml.