Real-Time Diagnostics on a QKD Link via QBER Time-Series Analysis

Area of Science:

• Quantum Information Science
• Network Security
• Machine Learning Applications

Background:

• Integrating Quantum Key Distribution (QKD) systems into metro optical networks presents significant technical challenges.
• Existing methods are insufficient for real-time identification of QKD link impairments within communication networks.

Purpose of the Study:

• To devise a methodology for real-time identification of various impairments affecting QKD links in communication networks.
• To develop a supervised machine learning model for this identification task.

Main Methods:

• A supervised machine learning model was designed and trained.
• The model utilizes time-series data of Quantum Bit Error Rate (QBER) and Secure Key Rate (SKR) as input.
• The model's design ensures independence from specific QKD protocols or systems.

Main Results:

• The developed methodology enables real-time identification of QKD link impairments.
• The model successfully processes QBER and SKR data to determine link status.
• The output provides critical information about the QKD link's working conditions.

Conclusions:

• The proposed machine learning approach offers a versatile solution for monitoring QKD link health in metro optical networks.
• This method enhances the reliability and manageability of QKD systems.
• The insights gained are valuable for users and key management systems, improving overall network security.