Paper review: < Analysis of the Increase and Decrease Algorithms for Congestion Avoidance in Computer Networks [CJ89] >

Reviewer: <Ryan Gehl>

1. State the problem the paper is trying to solve.

This paper concentrates on a detailed analysis of the increase/decrease algorithms for dealing with the problem of network congestion.

2. State the main contribution of the paper: solving a new problem, proposing a new algorithm, or presenting a new evaluation (analysis). If a new problem, why was the problem important? Is the problem still important today? Will the problem be important tomorrow?  If a new algorithm or new evaluation (analysis), what are the improvements over previous algorithms or evaluations? How do they come up with the new algorithm or evaluation? 

The main contribution of this paper is to abstractly characterize the wide class of increase/decrease algorithms and compare them using several different performance metrics.

3. Summarize the (at most) 3 key main ideas (each in 1 sentence.) 

(1) Congestion avoidance and congestion control are two different things. Congestion avoidance attemnpts to keep the network from congesting by "operating at the knee" whereas congestion control shemes help improve performance after congestion has occurred.
(2) The criteria for selecting controls are efficiency, fairness, distributedness, and convergence. (3) What are all the possible solutions that converge to efficient and fair states? How do we compare those controls that converge.

4. Critique the main contribution
   • Rate the significance of the paper on a scale of 5 (breakthrough), 4 (significant contribution), 3 (modest contribution), 2 (incremental contribution), 1 (no contribution or negative contribution). Explain your rating in a sentence or two.

   I would give this paper a 4 because it seems to be a significant contribution to the series of papers on congestion [7,8,10,11]. This paper's analytical approach in describing how to characterize different controls really strengthens their results.

   • Rate how convincing the methodology is: how do the authors justify the solution approach or evaluation? Do the authors use arguments, analyses, experiments, simulations, or a combination of them? Do the claims and conclusions follow from the arguments, analyses or experiments? Are the assumptions realistic (at the time of the research)? Are the assumptions still valid today? Are the experiments well designed? Are there different experiments that would be more convincing? Are there other alternatives the authors should have considered? (And, of course, is the paper free of methodological errors.)

   The authors of this study use analytical models to classify the set of possible solutions. The authors first characterize the set of all linear controls that converge and narrow this set down based on the distributedness criterion. Next, they try to find a subset that represents the optimal trade-off of responsiveness and smoothness. Finally, the authors discuss how to extend the results to nonlinear controls as well as discussing practical considerations.

   • What is the most important limitation of the approach?

   The biggest limitation of this paper was the fact that the authors had to abstract away several details in order to classify a set of problems. Specifically, the model assumes that all users sharing the same bottleneck will receive the same feedback and react to it.

5. What lessons should researchers and builders take away from this work. What (if any) questions does this work leave open?

A simple additive increase and multiplicative decrease algorithm satisfiees the sufficient conditions for convergence to an efficient and fair state regardless of the starting state of the network.