题目: Streaming High Quality Video over the Internet: Challenge,Framework, and Evaluation 报告人： Yang Guo
Corporate Research, Thomson 时间：2008年6月12日 （周四） 上午 10：00-11：30 地点：蒙民伟楼109会议室
Video over Internet has been gaining popularity due to the fast
penetration of high-speed Internet access, expanding group of savvy
broadband users, and rich video contents available over the Internet.
Peer-to-Peer video streaming is emerging as a novel/effective paradigm
for video delivery with low infrastructure cost. Most of the video
contents broadcasted over Internet, however, are low/medium quality
content with the video bit-rate around 400 kpbs. Higher streaming rate
means better video quality. It also provides more cushion to absorb the
bandwidth variations caused by peer churn, network congestion, etc. A
P2P streaming framework that is capable of support high video bit-rate
while accommodating a large number of users is highly desired.
In this talk, I will present Hierarchically Clustered P2P Streaming
system (HCPS), a P2P streaming framework that is scalable and supports
the streaming rate approaching the global optimum upper bound. HCPS
employs the distributed queue-based chunk scheduling algorithm that
achieves the maximum streaming rate allowed by a P2P streaming system in
practical network setting. Various implementation issues are also
explored to handle dynamics in realistic network environments In the second part of the talk, I will touch upon the P2P scheduling
sensitivity problem. Considerable efforts have been made by academia
and industry on P2P streaming design. While academia mostly focus on
exploring the design space to approach theoretical performance bounds,
commercial P2P streaming systems are able to deliver good user Quality
of Experience(QoE) with seemingly simple designs. One intriguing
question is: how elaborate should a good P2P video streaming design be?
We address this question through an extensive experimental comparison
study of several representative P2P video streaming designs. Our study
unveils several key mechanisms contributing most to the successes of P2P
streaming on the Internet. We also identify regions where naive designs
are inadequate and more elaborate designs can improve things
considerably. The study not only brings us better understandings and
more insights into the operation of existing systems, but also sheds
lights on the design of future systems that can strike a good balance
between the performance and the complexity. Bio: Yang Guo is a senior member of technical staff at Corporate
Research, Thomson at Princeton, NJ. He received his B.S. and M.S.
degrees from Shanghai Jiaotong University, and his Ph.D. degree in
Electrical and Computer Engineering from University of Massachusetts at
Amherst. His research interests include peer-to-peer networking and
content distribution, media streaming, real-time system, and network
modeling and performance evaluation. His recent work focuses on users'
quality-of-experience (QoE) improvement in P2P based live streaming and
VoD services using optimization and intelligent scheduling algorithms.
|
