Paper Review: <<>BBSS01><>>
Dynamic Behavior of Slowly-Responsive Congestion Control Algorithms

Reviewer: Oleg Elkhunovich

Problem

To explore the properties of TCP-friendly algorithms under more realistic dynamic conditions, addressing the fundamental question of whether these algorithms are safe to deploy in the Internet.

Contribution

The paper discusses several types of purposed TCP-friendly protocols, and analyzes their behavior using a series of tests. It concludes that most protocols would be safe to deploy.

Main Ideas

  • Most of the TCP-compatible algorithms are safe for deployment
  • Incorporating the principle of packet conservation is crucial in dynamic settings to ensure safety
  • In return for smoother transmission rates, slowly-responsive algorithms lose throughput to faster ones like TCP

    Critique

    Significance: 3
    This paper is significant as it tries to answer an important question of whether it is safe to deploy SlowCC protocols in the Internet.
    Methodology:
    Methodology of the paper is pretty good with different metrics defined, measured and discussed in relation to the problem addressed. However, while this work addresses analysis of the SlowCC protocols in dynamic environment, the tests are still not realistic and do not represent Internet conditions. Authors conclude that protocols are safe for the Internet only based on some stress tests.
    Limitation:
    Paper lacks theoretical analysis of the protocols and their properties. Also, tests do not entirely convince me of the conclusion that the authors make.

    Lessons:
    It can be possible to create new protocols with totally new properties, yet they can co-exist by preserving certain characteristics and being fair when interacting between each other and old protocols.