Paper Review: < RHE99 >
Quality Adaptation for Congestion Controlled Video Playback over the Internet
Reviewer: Oleg Elkhunovich
The problem addressed by this paper is adapting the compression without requiring video-servers to re-encode the data, and fitting the resulting stream into a rapidly varying bandwidth.
Authors present a mechanism for using layered video in the context of unicast congestion control. The scheme allows the server to trade short-term improvement for long-term smoothing of quality. The problem is very important as video-straming is going through explosive growth and is expected to keep growing.
Quality adaptation - ability of the server to adjust quality of the stream it plays back so that the perceived quality is as high as possible with available bandwidth.
Layered quality adaptation with hiercarchical encoding is used to achieve the goals of designing this scheme.
End-to-end congestion control done is TCP-friendly and follows AIMD scheme.
This paper is significant because it addresses and important problem of creating a good scheme for streaming video and audio. This is very important because those applications do not perform end-to-end congestion control now and their growth can lead to the collapse of congestion control.
Methodology of the paper is very good. Authors motivate their scheme, then describe it in good detail and analize it theoretically. They then present results of their experiments showing that scheme works well
Experiments are not all convincing and some assumptions do not represent true Internet.
Applying basic concepts from previous problems can lead to good solutions to seemingly very differnt problems.