Paper Review:
Quality Adaptation for Congestion Controlled Video Playback over the Internet

Reviewer: Jie Zhou

Problem

Internet has been experiencing explosive growth of audio and video streaming. To ensure fairness among applications, the streaming applications have to adopt effective congestion control mechanism. However, the congestion control mechanisms usually will cause varying available bandwidth. Since rapid fluctuation of quality is very disturbing to users, how to avoid it while performing congestion control becomes an important issue.

Contribution

The paper presents a mechanism for using layered video in the context of unicast congestion control. This quality adaptation mechanism bridges the gap between short-term changes in transmission rate and the need for stable quality in streaming applications.

Main Ideas

  • Buffering can effectively help cope with short-term changes in bandwidth due to AIMD congestion control.
  • Increasing the smoothing factor can limit the frequency of quality variation.

    Critique

    This paper proposes an effect machanism for quality adaption in context of congestion control. It somehow contributes to the incorporation of congestion control into streaming applications. So I give it rate 3 (modest contribution). Sound logic and mathematical analysis, together with extensive simulation results, makes this paper quite convincing.

    Limitations:
    The mechanism is built on the assumption that the congestion control employs an additive increase, multiplicative decrease algorithm. If the dominant congestion control is no longer AIMD in future, the mechanism will become invalid.

    Lessions:

    The paper chooses the Rate Adaptation Protocol as the underlying congestion control mechanism because it's properties are relatively simple to predict. This is a good lesson. When studying some aspect of a subject, we'd better keep other aspects as simple as possible so that we can achieve our goal much easier.