From: Ioannis Giotis (giotis@cs.washington.edu)
Date: Sun Oct 24 2004 - 23:23:08 PDT
The problem at hand is congestion control. The authors point out that
queuing algorithms at the gateways can also serve as congestion control
mechanisms.
The proposed algorithm is a variation of a round-robin system. This
algorithm ensures fairness among the sources and also manages to enforce
congestion control among sources. This is achieved by decoupling bandwidth ,
promptness and queues.
The authors present clear arguments about why their algorithm performs much
better than the other known ones and present some simulations to support
their algorithms.
The major question was, why didn't they implement this earlier ? The
algorithm is simple, intuitive and seems to perform well. I was impressed by
the completeness of the paper.
Of course today, the usefulness of the algorithm is questionable as we would
like to implement more complex flow controlling in order to guarantee Quos,
security and protection against misbehaving hosts.
On the other hand, one of the major strengths of the algorithm is not its
performance but the fact that it lies entirely in the gateways which are
more secure and trustworthy. It also allows easy upgradeability, a feature
that would be most useful if this algorithm had prevailed. I suppose that it
would be really easy to add extensions to this algorithm to handle QoS, by
modifying its round-robin pattern.
Overall, the paper successfully addressed the issue at hand and the authors
managed to do this in a way that would allow their algorithm to be
implemented easily.
This archive was generated by hypermail 2.1.6 : Sun Oct 24 2004 - 23:23:02 PDT