Rate-Distortion Optimization in Data Driven Peer-to-Peer Streaming
Date Issued
2008
Date
2008
Author(s)
Chiu, Ti-Chia
Abstract
Perceptual visual quality guarantee is of vital concern for the P2P streaming system. Most of existing works, nevertheless, focus on how to improve network-oriented Quality-of-Service (QoS) such as throughput, delay, and jitter in the P2P streaming system. Few of them concentrate on how to enhance the perceptual visual quality under rigorous network conditions. In this work, we first identify the new challenging issue, i.e. the control of distribution of received blocks, in a data-driven P2P streaming system. Specifically, we exploit an optimistic end-to-end bandwidth estimation approach to adapt time variant network conditions and varying source streaming rate so as to utilize the overall system bandwidth as much as possible. In addition, a rate-distortion prioritized block assignment is involved to enhance visual quality. That is, the estimated bandwidth and the prioritized blocks are taken into a distributed scheduler to control the distribution of received blocks so as to achieve the objective of visual quality optimization. The simulation results show that we can not only achieve better application-level deliver ratio, but also can improve visual quality up to 1.7 dB on average compared with existing P2P streaming systems.
Subjects
P2P Streaming
Rate-Distortion Optimization
Data Block Scheduling
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-97-R95922008-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):5c9ab294358544d48876036e239c917f