Provisioning of Quality-of-Service in Video Streaming

Project Description

The current best-effort Internet does not guarantee the bandwidth availability between a receiver and a sender, and so renders any quality-of-service (QoS) control difficult, if not impossible. This research project investigates reactive and predictive approaches to providing QoS control in streaming video over the best-effort Internet.

Reactive approaches refer to algorithms that monitors the network resources availability and then perform content adaptation or delivery routing to adapt to variations in the resources available. For example, this project developed a new adaptation algorithm to adjust the bit-rate of video data [1] in response to the network bandwidth available to improve playback continuity. Unlike previous works, the proposed algorithm requires no parameter tuning and yet can achieve better performance in most cases [2]. In addition to performing adaptation within a video stream, we also investigate the feasibility to perform bandwidth reallocation across multiple video streams in a TCP-friendly manner. The first step is to develop algorithms to reallocat bandwidth across a group of TCP flows sharing the same network bottleneck [3], and then the second step will be to develop data and bandwidth allocation algorithms to dynamically adjust the assigned bandwidth to improve QoS.

Predictive approaches refer to algorithms that not only react to observed changes in network resources availability, but also estimate and predict their future availability to support admission control, buffering, or even combined with adaptation. In particular, with the rapid growth of peer-to-peer applications, more and more applications will generate many-to-one rather than one-to-one traffic flows. Our investigation revealed that the aggregate bandwidth from multiple senders, while still varies randomly, is far more predictable than a single sender. Thus this increased predictability could open a new way to achieve probabilistic performance guarantees in multimedia applications.

The early measurement experiments conducted in the PlanetLab and in the Internet have been very promising [4-5]. Using these measurement data in trace-driven simulations we found that the statistical properties of many-to-one data flows can be effectively exploited for playback adaptation [6] as well as prefetch buffering [7-8].

Publication

1. Peter H. W. Wong, Robert T. W. Hung, Jack Y. B. Lee, S. C. Liew, C. S. Kim, and Roland T. Chin, "Rate Estimation For H.264/AVC Spatial Resolution Reduction," Proc. IEEE International Conference on Image Processing (ICIP 2004), Oct 24-27, 2004, Singapore.
2. L. S. Lam, Jack Y. B. Lee, S. C. Liew, and W. Wang, "A Transparent Rate Adaptation Algorithm for Streaming Video over the Internet," Proc. 18th International Conference on Advanced Information Networking and Applications (AINA 2004), Fukuoka, Japan, March 29-31, 2004.
3. K. H. Lau and Jack Y. B. Lee, "Sender-Driven Bandwidth Differentiation for Transmitting Multimedia Flows Over TCP," Proc. the 13th Annual Multimedia Computing and Networking (MMCN'06), January 18-19, 2006, San Jose, California, USA.
4. S. C. Hui and Jack Y. B. Lee, "Modeling of Aggregate Available Bandwidth in Many-to-One Data Transfer," Proc. of the Fourth International Conference on Intelligent Multimedia Computing and Networking, July 21-26, 2005, Utah, USA.
5. V. T. Sam, P. Y. Ho, and Jack Y. B. Lee, "A First Look at the Properties of Many-to-One Data Flows," to appear in the Workshop on Parallel and Distributed Multimedia held in conjunction with the International Conference on Parallel Processing (ICPP 2006), Ohio, USA, August 14, 2006.
6. S. C. Hui and Jack Y. B. Lee, "Playback-Adaptive Multi-Source Video Streaming," Proc. of the Fourth International Conference on Intelligent Multimedia Computing and Networking, July 21-26, 2005, Utah, USA.
7. P. Y. Ho, V. T. Sam, and Jack Y. B. Lee, "Providing Probabilistic Performance Guarantees in Multi-Source Video Streaming over the Internet," Proc. Workshop on Internet Protocol TV (IPTV) Services over World Wide Web held in conjunction with the 2006 World Wide Web Conference, 23-26 May 2006, Edinburgh, Scotland. (Invited Paper)
8. P. Y. Ho and Jack Y. B. Lee, "QoS-aware Adaptive Buffering for Multi-Source Video Streaming over the Internet", to appear in IEEE GLOBECOM 2006.

Contact Webmaster with comments or questions regarding this site.
© Copyright 2006, Multimedia Communications Laboratory, All rights reserved.