SMASH: A Supervised Machine Learning Approach to Adaptive Video Streaming over HTTP

The growth of online video-on-demand consumption continues unabated. Existing heuristic-based adaptive bit-rate (ABR) selection algorithms are typically designed to optimise video quality within a very narrow context. This may lead to video streaming providers implementing different ABR algorithms/players, based on a network connection, device capabilities, video content, etc., in order to serve the multitude of their users' streaming requirements. In this paper, we present SMASH: a Supervised Machine learning approach to Adaptive Streaming over HTTP, which takes a tentative step towards the goal of a one-size-fits-all approach to ABR. We utilise the streaming output from the adaptation logic of nine ABR algorithms across a variety of streaming scenarios (generating nearly one million records) and design a machine learning model, using systematically selected features, to predict the optimal choice of the bitrate of the next video segment to download. Our evaluation results show that SMASH guarantees a high QoE with consistent performance across a variety of streaming contexts.

I. Introduction

Video-on-demand providers continue to grow their dominance in the area of subscription-based online video streaming services. These providers target a variety of customers distributed around the world. This diversity of customers poses a huge technical challenge in providing a high level of Quality of Experience (QoE) globally. As the underlying transmission medium characteristics can vary quite dramatically, these providers encode their video content into a discrete number of quality levels and implement an adaptive video content delivery mechanism to maximise QoE. Currently, the most prominent adaptive video streaming technology is HTTP Adaptive Streaming.