QoE‐aware traffic monitoring based on user behavior in video streaming services
Fatima Laiche,Asma Ben Letaifa,Taoufik Aguili
DOI: https://doi.org/10.1002/cpe.6678
2021-11-08
Concurrency and Computation: Practice and Experience
Abstract:In recent years, online video content has gained a lot of popularity and the exponential growth of video traffic continues in every area of the connected world. Thus, understanding the quality of experience (QoE) perceived by end‐users of video streaming services is important for both network operators and over the top providers since estimating end user's QoE has become one of the main points to meet user expectations. However, it is not trivial to ensure an adequate QoE since user experience is affected by various influencing factors (e.g., context factors, human factors, and system factors), and it is still challenging to identify the QoE key influencing factors. To address these challenges, we focused on improving QoE estimation and management strategies by exploiting valuable human and context information because the influence of social contextual information and user behavior on the perceptual quality is often neglected. In this article, we first proposed a classification of influence factors into four categories which are: system factors, human factors, context factors, and social‐behavioral factors. We developed a monitoring web application where video content is played to end‐users so that subjective and objective video metrics are collected. We built a new machine learning (ML) based model for QoE prediction. We used well‐known supervised ML algorithms like decision tree, k‐nearest neighbors, and support vector machine. Finally, we proposed a QoE management approach in the context of software defined network/multi‐access edge computing that implements the proposed QoE prediction model to optimize the video delivery transmission chain.