Re: [Cake] packet mass, ecn, and a fractional count

[Jonathan Morton <chromatix99@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 788 bytes --]

I have no idea how Tracy-Widom applies here, or what it would mean if it
did. Pure mathematics is not my strong point. If there's a concrete point
you're trying to make, I might grasp it better if explained a different way.

What I do understand is that fq_codel and cake already provide a higher
signalling rate in the presence of multiple flows, when compared to plain
codel. Each codel instance signals at a given rate once triggered, and the
number of instances scales more or less linearly with the number of flows.
Further, the signals are directed specifically to the flows which need them.

So I don't see the need to increase that signalling rate still further, in
the FQ case. It might be a good idea for plain codel, but that's not the
subject of this list.

- Jonathan Morton

[-- Attachment #2: Type: text/html, Size: 880 bytes --]