Re: [tsvwg] Gorry Fairhurst Individual thoughts on choosing whether/howto advance ECN work.

[Paul Vixie <[e-mail]>]

On Saturday, 23 May 2020 22:09:21 UTC Sebastian Moeller wrote:
> ...
> 
> I hope you agree that there is no clear significant complexity difference
> between L4S and fq_codel if measured by lines of code. I am not claiming
> that that is an ideal measure, but it sure beats lines of pseudocode.

complexity is better measured in terms of state variables and transitions; 
there's nothing published i can find which compares these two competing 
approaches in those terms, and so i'm inclined to think in terms of network 
carried vs. edge carried complexity. it's the placement in this case, not the 
magnitude, that concerns me. i liked the fq_codel approach because it helped 
immediately and measurably at every increment of deployment, nobody had to 
deploy on faith as in ipv6 and dnssec that benefits would come later when 
others also deployed. so if you're counting fan boys, i'm an edge-complexity 
guy, even if the mass of all those edges each carrying all that flow state is 
higher overall than a smart-core approach.

i realize that nobody is proposing the fr/atm model (cir, br, per flow) but i 
still get cold chills down my spine when i hear a proposal for core complexity 
-- of whatever kind.

-- 
Paul