Trusted Repeater Placement in QKD-enabled Optical Networks Enables Scalable Key Distribution with Reliability-Aware Node Ranking

Quantum Key Distribution offers an unparalleled level of security for communications, but its practical implementation in optical networks faces a fundamental limitation, distance. Arup Kumar Marik, Basabdatta Palit, and Sadananda Behera from the National Institute of Technology Rourkela address this challenge with a new approach to placing ‘trusted repeater’ nodes, essential for extending the reach of secure quantum communications. Their work moves beyond simply assuming the security of these repeaters, recognising the potential for vulnerabilities from software flaws or malicious insiders, and instead proposes a system that assesses and integrates the reliability of each node into the network design. By ranking nodes based on both their network position and a calculated ‘trust score’, the team demonstrates a significant improvement in network coverage, achieving over 10% more secure connections compared to conventional methods while using a similar number of repeater nodes, paving the way for more robust and scalable quantum communication networks.

The team addressed a key challenge: long-distance QKD suffers from signal loss, requiring relay stations, but assuming complete trustworthiness of these stations introduces vulnerabilities. Their novel framework considers both network topology and the reliability of each node, using a composite score combining node reliability, betweenness centrality, and eigenvector centrality to rank potential TRN locations for optimal coverage and security. Tested on a 28-node metropolitan optical network with randomized reliability scores, the method consistently achieves higher path coverage than a baseline approach using only degree centrality, demonstrating a 10. 77% improvement with approximately eight TRNs. Recognizing that assuming complete trustworthiness of all relay nodes introduces security vulnerabilities, the team prioritized node reliability alongside network performance. Their approach assigns a trust score to each node, reflecting resilience to threats, and integrates this score into a modified Dijkstra algorithm, weighting paths to favour more reliable nodes during key distribution. To determine optimal TRN placement, the team employed a composite scoring system combining betweenness centrality and eigenvector centrality, identifying nodes strategically positioned within the network and well-connected to critical infrastructure.

This composite score enables a scalable deployment strategy, allowing efficient TRN placement even in complex topologies. Results demonstrate a 10. Addressing the limitation of existing QKD systems restricted by signal loss, this new approach moves beyond assuming fully trusted relay nodes, incorporating node reliability into the path selection process to mitigate risks from software exploits and insider threats. The team modified Dijkstra’s algorithm to utilize weighted links reflecting each node’s trust score, ensuring routes favour more reliable infrastructure. Experiments demonstrate a 10.77% increase in shortest path coverage compared to traditional TRN placement strategies based solely on degree centrality, utilizing approximately eight TRNs. Researchers designed a composite centrality metric combining betweenness centrality and eigenvector centrality, effectively assessing a node’s influence within the network. Researchers addressed a critical limitation of existing approaches, which assume complete reliability of all nodes, by incorporating a node’s trustworthiness into the network planning process. The team developed an algorithm that combines Dijkstra’s algorithm with weighted links reflecting both distance and node reliability, identifying the most secure and efficient paths for key distribution. Evaluation demonstrates a 10. Future work will focus on optimizing parameters to maximize key rates, success rates, and minimize the number of required repeaters, further refining the framework’s performance and practicality.