Shedding light on the structure of internet video quality problems in the wild

  • Authors:
  • Junchen Jiang;Vyas Sekar;Ion Stoica;Hui Zhang

  • Affiliations:
  • Carnegie Mellon University, Pittsburgh, USA;Stony Brook University, New York, USA;University of California, Berkeley, Berkeley, USA;Carnegie Mellon University, Pittsburgh, USA

  • Venue:
  • Proceedings of the ninth ACM conference on Emerging networking experiments and technologies
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

The key role that video quality plays in impacting user engagement, and consequently providers' revenues, has motivated recent efforts in improving the quality of Internet video. This includes work on adaptive bitrate selection, multi-CDN optimization, and global control plane architectures. Before we embark on deploying these designs, we need to first understand the nature of video of quality problems to see if this complexity is necessary, and if simpler approaches can yield comparable benefits. To this end, this paper is a first attempt to shed some light on the structure of video quality problems. Using measurements from 300 million video sessions over a two-week period, we identify recurrent problems using a hierarchical clustering approach over the space of client/session attributes (e.g., CDN, AS, connectivity). Our key findings are that: (1) a small number (2%) of critical clusters account for 83% of join failure problems (44--84% for other metrics); (2) many problem events (50%) persist for at least 2 hours; (3) a majority of these problems (e.g., 60% of join failures, 30--60% for other metrics) are related to content provider, CDN, or client ISP issues. Building on these insights, we evaluate the potential improvement by focusing on addressing these recurrent problems and find that fixing just 1% of these clusters can reduce the number of problematic sessions by 55% for join failures (15%--40% for other metrics).