What's New

Predicting the Unpredictable - A New Approach to Providing Statistical Performance Guarantees in Mobile Video Streaming

Project Description
With the growth in smartphone adoption and world-wide deployment of high-speed mobile data networks such as 3G/4G, mobile video streaming is not only feasible but also commonplace nowadays. According to a recent study, global mobile data traffic will increase 13-fold between 2012 and 2017, with video traffic accounting for over 66 percent of all mobile data traffic by the end of the forecasted period. Despite rapid advances in mobile networks, streaming high-quality video to mobile users remains a challenge. While modern mobile networks can offer very high peak bandwidth, their nature of wireless transmission unavoidably exhibits bandwidth fluctuations, which are far more significant than their wired counterparts. Moreover, with the recent widespread adoption of streaming video over HTTP/TCP, the throughput of a video session is further modulated by TCP's built-in flow and congestion control algorithms. In fact one cannot expect reliable and consistent performance when streaming video over mobile networks - the encounter of playback rebuffering is the norm rather than the exception.
   One problem is the lack of network or transport-level QoS control in mobile networks. Most, if not all, of the industry have largely given-up on the hope of offering predictable performance in mobile video streaming, leading to the switch of focus to adaptive video streaming. However, our investigations show that while adaptive video streaming can improve streaming performance, its performance is still highly unpredictable. In this project we argue that predicting the streaming performance over mobile networks is not only possible, but also feasible in today's mobile networks. The key insight is that the streaming community has been using the wrong metric - throughput statistics, to directly inform the choice of video bitrate. By contrast, we propose a novel framework called post-streaming rate analysis to extract the correlation between traffic flow statistics and video bitrate selection, and quantify their impact to streaming performance. Our pilot study shows that the proposed framework can (a) quantify the tradeoffs between video bitrate selection and streaming performance; and (b) enable online video bitrate selection with predictable streaming performance - to the best of the PI's knowledge it is the first video bitrate selection algorithm possessing such property. Achieving predictable performance has always been one of the holy-grails in mobile video streaming. With the rapid migration from physical media to online media streaming, we expect results from this research project to have a far-reaching impact to the global mobile/media industry.

Project Funding
This project is funded by a General Research Fund grant (GRF 14203714) from the HKSAR Research Grant Council.

PI: Prof. Jack Lee


  1. Guanghui Zhang and Jack Y. B. Lee, "On Data Wastage in Mobile Video Streaming," Proc. IEEE Wireless Communications and Networking Conference (WCNC), Barcelona, Spain, April 15-18, 2018.
  2. Yan Liu and Jack Y. B. Lee, "Post-Streaming Rate Analysis - A New Approach to Mobile Video Streaming with Predictable Performance," IEEE Transactions on Mobile Computing, vol.16(12), Dec 2017, pp.3488-3501.
  3. Ke Liu and Jack Y. B. Lee, "On Improving TCP Performance Over Mobile Data Networks," IEEE Transactions on Mobile Computing, vol.15(10), Oct 2016, pp.2522-2536.
  4. Victor K.C. Wu, Yan Liu, and Jack Y. B. Lee, "Exploiting Trace Data for Adaptive Mobile Video Streaming with Performance Guarantees," Proc. Second IEEE International Conference on Multimedia Big Data (BigMM), Taipei, Taiwan, 20-22 Apr 2016.
  5. Yan Liu and Jack Y. B. Lee, "A Unified Framework for Automatic Quality-of-Experience Optimization in Mobile Video Streaming," Proc. IEEE INFOCOM 2016. 15 April 2016, San Francisco, CA, USA.
  6. Yan Liu and Jack Y. B. Lee, "Streaming Variable Bitrate Video Over Mobile Networks with Predictable Performance," Proc. IEEE Wireless Communications and Networking Conference (WCNC), 3-6 April 2016, Doha, Qatar.
  7. Yan Liu and Jack Y. B. Lee, "An Empirical Study of Throughput Prediction in Mobile Data Networks," Proc. IEEE GLOBECOM 2015, San Diego, CA, USA, December 6-10, 2015.
  8. Victor K. C. Wu, Yan Liu, and Jack Y. B. Lee, "Mobile Video Streaming with Video Quality and Streaming Performance Guarantees", Proc. IEEE International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob), 19-21 Oct 2015, Abu Dhabi, UAE.
  9. K. M. Chan and Jack Y. B. Lee, "Improving Adaptive HTTP Streaming Performance with Predictive Transmission and Cross-Layer Client Buffer Estimation," Multimedia Tools and Applications, Springer, 19 Mar 2015.
  10. Yan Liu and Jack Y. B. Lee, "On Adaptive Video Streaming with Predictable Streaming Performance," Proc. IEEE GLOBECOM 2014, Texas, USA, December 8-12, 2014.


Contact Webmaster with comments or questions regarding this site.
© Copyright 2020, Multimedia Communications Laboratory, All rights reserved.