Paper Review: < DKS89 >>
Analysis and Simulation of a Fair Queueing Algorithm
Reviewer: Oleg Elkhunovich
Problem
Methods of congestion control in datagram networks have two points of implementation - the source and the gateway. This paper analyses Fair Queueing in gateways.
Contribution
FCFS queueing has some major flows in terms of fairness, and this paper presents an intersting alternative that might be leading to a better scheme.
Main Ideas
FCFS is inherently not fair, alternative algorithm can help with congestion control
Fair queueing is based on the round-robin scheduling
Fair queueing prevents ill-behaved sources from aquiring extra bandwidth
Critique
Significance: 4
This paper presents a good alternative to FCFS scheme with some promising results.
Methodology:
Methodology is very good as authors describe their motivations based on previous research, add constraints, and present their algorithm They do a good job discussing their algorithm analytically and then present interesting simulation results.
Limitation:
Simulation has been done on only limited number of scenerios. Some of the results are ignored and dismissed and too complicated. It is not clear if scheme proposed is economically feasible.
Lessons:
When solving a problem, it is good to keep in mind different aspects that come together. Authors go beyond end-to-end congestion control to solve congestion problem.