Research Article
A BP Neural Network Based Self-tuning for QoS Support in AVB Switched Ethernet
@INPROCEEDINGS{10.1007/978-3-319-78139-6_47, author={Dong Chen and Ang Gao}, title={A BP Neural Network Based Self-tuning for QoS Support in AVB Switched Ethernet}, proceedings={Communications and Networking. 12th International Conference, ChinaCom 2017, Xi’an, China, October 10-12, 2017, Proceedings, Part II}, proceedings_a={CHINACOM}, year={2018}, month={4}, keywords={AVB BP neural networks Self-tuning Machine learning QoS}, doi={10.1007/978-3-319-78139-6_47} }- Dong Chen
Ang Gao
Year: 2018
A BP Neural Network Based Self-tuning for QoS Support in AVB Switched Ethernet
CHINACOM
Springer
DOI: 10.1007/978-3-319-78139-6_47
Abstract
To support QoS of time-sensitive services in Ethernet, IEEE has proposed a set of standards for transporting and forwarding real-time content over Ethernet known as Audio Video Bridging (AVB) with bandwidth reservation and priority isolation. AVB traffic is granted highest priority to ensure its transmission while low-priority traffic follows Strict Priority (SP). However, due to restrictions of SP algorithm, low-priority traffic may suffer a problem of starvation. To solve the problem, we propose a BP neural network based self-tuning controller (BPSC) over a probability selector to manage the transmission of best effort (BE) traffic in AVB switched Ethernet. This paper introduces the model of BPSC, followed by an simulation to demonstrate that BPSC could operate effectively and dynamically. The result shows that BPSC not only has the ability to manage the transmission precisely, but also shows both effectiveness and robustness.