Summary: John Schmoller
"Analysis and Simulation of a Fair Queueing Algorithm" Reaction
The paper "Analysis and Simulation of a Fair Queueing Algorithm" examines a
new strategy for routers. Instead of the First-Come-First-Serve nature of today's current
routers, they propose that each source-destination pair passing through a router should
have its own queue, and each source should get a fair slice of the router's bandwidth.
This would be done in a round-robin fashion, where they weight the size and frequency
of packets to make sure that no source uses more than its fair share of bandwidth. They
hope that this will eliminate threats from ill-behaved hosts by limiting their bandwidth so
they only get a fair-share.
In my opinion, this paper has no merit. They don't even address how a router
could possibly have enough queues for every source-destination pair passing through
them at any time. And they don't discuss how a router could possibly know, keep track
of, and quickly process all this information without the internet grinding to a screeching
halt. While it's an interesting theoretical exercise, it's more of a math paper than
anything that could ever be applied.
From a clean slate perspective, I also believe that this paper does not offer any
ideas that could be used. As stated in the previous paragraph, it's just not technologically