An adaptive machine learning-based QoE approach in SDN context for video-streaming services

Abstract

In data service applications over the Internet, user perception and satisfaction can be assessed by quality of experience (QoE) metrics. QoE depends both on the users’ perception and the used service, which together form end-to-end metrics. While network optimization has traditionally focused on optimizing network properties such as QoS, we focus in this work on optimizing end-to-end QoE metrics with the aim to deliver to the client a good QoE that can be monitored in real time. We argue that end-user QoE is a relevant measurement for network operators and service providers. In this paper, we present a machine learning approach combined with adaptive video delivery service in order to provide a better QoE for video streaming services. This solution will be established using SDN architecture. The first part of the paper deals with a brief introduction of SDN networks, QoE requirement, and ML algorithms. Secondly, we expose the rating of the web application that we developed. This will help in conducting a subjective study to collect MOS on real-time as well as objective parameters SSIM, VQM, and PSNR. At the end, we expose our QoE-aware monitoring approach and explain what it is based on.