A Public Dataset for YouTube's Mobile Streaming Client

Abstract

Datasets are a valuable resource to analyze, model and optimize network traffic. This paper describes a new public dataset for YouTube's popular video streaming client on mobile devices. At the moment, we are providing 374 hours of time-synchronous measurements at the network, transport and application layer from two controlled environments in Europe. After describing our experimental design in detail, we discuss how to use our dataset for the analysis and optimization of HTTP Adaptive Streaming (HAS) traffic and point to specific use cases. To assure reproducibility and for community benefit, we publish the dataset at [1].