Applications of Machine Learning in Quantum Key Distribution Networks

Zhao, Yongli; Zhang, Kaixin; Zhu, Qingcheng; Wang, Hua; Yu, Xiaosong; Zhang, Jie

doi:10.1109/ogc52961.2021.9654412

Cited by 3 publications

(2 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Artificial intelligence (AI), machine learning (ML), deep learning (DL), and optimisation techniques can play significant roles in enhancing the performance of existing QKD networking techniques [6]. Consequently, there is a growing interest in employing machine learning to enhance the performance of quantum communication networks [36]. One of the highly used algorithms in this regard is the reinforcement learning (RL) algorithm which is used to explain and conclude how an intelligent agent learns and improves its strategies through interacting with its environment [55], [17], [15], [49].…”

Section: Introductionmentioning

confidence: 99%

A Novel Security Survival Model for Quantum Key Distribution Networks Enabled by Software-Defined Networking

Shirko

Askar

2023

IEEE Access

View full text Add to dashboard Cite

Quantum-key-distribution (QKD) networks can theoretically generate quantum secret keys that are completely secure for an entire communication system. The rate of key distribution is low and decreases exponentially with increasing distance. A classic trusted relay (CTR) uses additional keys to enhance security distance in QKD networks. In practice, the assurance of security for certain relay nodes is still lacking, despite the fact that CTR requires that all nodes be trusted. Owing to channel unreliability, system faults accumulate during the key relay, thereby increasing the probability of CTR failing to distribute the secret key. The failure of a successful key relay would then result in the subsequent destruction of all the keys involved in the process, which leads to the wasting of the quantum secret key and reduction system encryption. Accordingly, reducing the impact of CTR failures and realising the key security distribution of remote quantum networks have become new challenges that require the development of a new survivability scheme. In this study, a software-defined networking (SDN) technique is introduced to circumvent this drawback by utilising the flexibility provided by the SDN paradigm for better QKD network management efficiency. In particular, a novel survivability model called software-defined quantum key relay failure (SDQKRF) is proposed in this paper in which a new function is developed and added to the SDN controller. According to the simulation results, SDN over a QKD network using the SDQKRF model is more reliable and performs better in terms of the key generation ratio, key utilisation rate, recovery after failure, avalanche effect, and service blocking rate than a regular QKD network without the SDQTRF model. INDEX TERMSQuantum key distribution (QKD), software-defined network (SDN), survivability, classical trusted relay(CTR).

show abstract

Section: Introductionmentioning

confidence: 99%

A Novel Security Survival Model for Quantum Key Distribution Networks Enabled by Software-Defined Networking

Shirko

Askar

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The key-management layer helps in key storage, verifying, routing and deleting of the previously established key, this layer is also in charge of quality of service (QoS).It has been suggested that QKDN management and controller layers are central to end-to-end QKD in the holistic control architecture (ITU-TY.3800, 2019;Zhao et al, 2021).Machine learning methods have been employed in the selection of optimal QKD protocol and applied random forest (RF) algorithm (Ren et al, 2021). The authors have compared RF with other machine learning algorithms and obtained an accuracy of 98 percent with the testing data set with a good receiver operating characteristic.…”

Section: Introductionmentioning

confidence: 99%

A Framework for Improving the Performance of QKDN using Machine Learning Approach

Arthi,

Saravanan,

Nayana

et al. 2023

ADCAIJ

View full text Add to dashboard Cite

A reliable secure communication can be given between two remote parties by key sharing, quantum key distribution (QKD) is widely concentrated as the information in QKD is safeguarded by the laws of quantum physics. There are many techniques that deal with quantum key distribution network (QKDN), however, only few of them use machine learning (ML) and soft computing techniques to improve QKDN. ML can analyze data and improve itself through model training without having to be programmed manually. There has been a lot of progress in both the hardware and software of ML technologies. Given ML’s advantageous features, it can help improve and resolve issues in QKDN, facilitating its commercialization. The proposed work provides a detailed understanding of role of each layer of QKDN, addressing the limitations of each layer, and suggesting a framework to improve the performance metrics for various applications of QKDN by applying machine learning techniques, such as support vector machine and decision tree algorithms.

show abstract