From: Jonas Lindberg (jonaslin@kth.se)
Date: Sun Oct 24 2004 - 16:43:18 PDT
Review of A. Demers, S. Keshav, and S. Shenker's "Analysis and Simulation of
a Fair Queueing Algorithm"
By: Jonas Lindberg
This paper has two major contributions: first the authors present
improvements to Nagel's Fair Queuing algorithm (e.g. to consider packet
length); second they present results from simulations in which FQ and FCFS
is compared.
An interesting discussion about fairness is held throughout the paper and
the authors explain why they do not consider FCFS fair. Demers also present
a lot of interesting advantages with FQ, such as increased throughput as a
consequence of shorter delay for small ACK packets.
The simulations that are presented consists of six scenarios: "Underloaded
Gateway", "Overloaded Gateway", "Ill-Behaved Source", "Mixed Source",
"Multihop Path", and "Complicated Network". The chosen scenarios show
differences between FQ and FCFS and each scenario is followed by a
discussion. It is clear that in these scenarios the FQ algorithm performs
better than FCFS.
When it comes to how the paper can be improved, I think the authors'
suggestions for future work and improvements are good. However, when reading
the paper I felt that it sometimes is not clear where to find the different
arguments and I therefore believe that it might be room for improvements in
the structuring of the paper.
I think this paper is relevant since it has contributed to a better design
of FQ algorithms. The discussion on fairness makes it even more interesting.
This archive was generated by hypermail 2.1.6 : Mon Oct 25 2004 - 00:43:48 PDT