Network performance evaluation criterion model based on large connections for low latency in industrial 50G-PON network

Abstract

• This study proposed the new evaluation criterion model for user and service satisfaction to optimize the resource scheduling strategy of the 50G-PON network. The combination of this satisfaction model and the MF-DWBA algorithm not only meets the practical needs of smart factories, but also brings important inspirations to other industrial manufacturing entities for network applications in Industry 4.0. • We propose a Multi Factor Dynamic Wavelength Bandwidth Allocation (MF-DWBA) algorithms, to meet the new standard of the 50G-PON. The MF-DWBA algorithm can flexibly adjust the weight of each service, according to the real-time changes in the network status, to maximize the overall network performance and resource utilization. • On the basis of the traditional single-index user satisfaction, we propose a multi-index model method based on reinforcement learning to enhance the existing evaluation indicators. This approach is designed to address the challenge of accommodating large-scale connectivity while maintaining low latency. • A 4 evaluation criterion matrix(4,M) service satisfaction evaluation model based on utility weight is introduced, which integrates response time, throughput, number of service connections and average packet delay, supplemented by coefficient weight. This comprehensive evaluation model is designed to focus on end-user business satisfaction. • The findings of this study are of great significance for promoting the performance evaluation and optimization technology of 50G-PON networks,provide a basis for further research and development of optical networks in the future. In this paper, a new evaluation criterion model for user and service satisfaction is proposed. The application of 50G-PON networks in Industry 4.0 is explored to enhance user experience and network efficiency by integrating 50G-PON in Industry 4.0. Facing the demand for high-bandwidth and low-latency networks for robotic assembly, laser cutting, sensors, and other devices in production lines. We propose the MF-DWBA (Multi Factor Dynamic Wavelength Bandwidth Allocation) algorithm, which aims to enhance user experience and network efficiency in industrial environments. Then, a user satisfaction evaluation criterion based on reinforcement learning is constructed through a mathematical model, and the satisfaction value calculated by this model can accurately identify network performance bottlenecks and further enhance the resource scheduling efficiency in the production environment. In addition, a (4,M) service satisfaction evaluation criterion based on utility function is proposed, which makes the satisfaction assessment of different service types in industrial scenarios more comprehensive and can reflect the service priority between critical and non-critical devices. Finally, the calculated satisfaction values are compared with the actual user feedback, and the simulation results prove the accuracy and effectiveness of the proposed satisfaction model. This study provides an important reference and guidance for the optimization of large-connection 50G-PON networks, and also provides a feasible path to solve the problem of network performance optimization in the smart shop environment. The combination of this satisfaction model and the MF-DWBA algorithm not only meets the practical needs of smart factories, but also brings important inspirations to other industrial manufacturing entities for network applications in Industry 4.0.