On the Analysis of YouTube QoE in Cellular Networks through in-Smartphone Measurements

Cellular-network operators are becoming increasingly interested in knowing the Quality of Experience (QoE) of their customers. QoE measurements represent today a main source of information to monitor, analyze, and subsequently manage operational networks. In this paper, we focus on the analysis of YouTube QoE in cellular networks, using QoE and distributed network measurements collected in real users' smartphones. Relying on YoMoApp, a well-known tool for collecting YouTube smartphone measurements and QoE feedback in a crowdsourcing fashion, we have built a dataset covering about 360 different cellular users around the globe, throughout the past five years. Using this dataset, we study the characteristics of different QoE-relevant features for YouTube in smartphones. Measurements reveal a constant improvement of YouTube QoE in cellular networks over time, as well as an enhancement of the YouTube video streaming functioning in smartphones. Using the gathered measurements, we additionally investigate two case studies for YouTube QoE monitoring and analysis in cellular networks: the machine-learning-based prediction of QoE-relevant metrics from network-level measurements, and the modeling and assessment of YouTube QoE and user engagement in cellular networks and smartphone devices. Last but not least, we introduce the YoMoApp cloud dashboard to openly share smartphone YouTube QoE measurements, which allows anyone using the YoMoApp smartphone app to get immediate access to all the raw measurements collected at her devices.