[Bloat] Arris thoughts on bufferbloat in the CMTS

[Dave Taht]

 I finally got my SCTE proceedings today.

In there were a presentation and paper by Arris's CTO and several of
his senior engineers...

http://snapon.lab.bufferbloat.net/~d/trimfat/Cloonan_Paper.pdf
http://snapon.lab.bufferbloat.net/~d/trimfat/Cloonan_Presentation.pdf

There are some very good ideas and analysis in here, a nice
enhancement to SFQ easily implementable in hardware, and a new version
of RED, called "LRED".

In a first reading, there are a few things I quibble with (they missed
the SQF-like portion of fq_codel entirely, for example), but it is
really nice to have seen Jim's
paper go from "so-called" in 2011, to "seminal" in 2013.

These are worth reading thoroughly. I will wait until saturday to
comment further.



-- 
Dave Täht