Re: [quicwg/base-drafts] Define idle period for congestion control (#2555)

[Jana Iyengar <[e-mail]>]

RFC7661 is already cited in the section, and it was added there to defer to it so we didn't have to delve into details in this doc.

I think what we're really discussing here is whether we need the MUST in the sentence "When sending data after becoming idle, a sender MUST reset its congestion window to the initial congestion window".  I'm in agreement with what I'm sensing here -- that we don't need a requirement, but we can discuss the implications of various decisions here and current best practices.

Higher-level note, yeah, we always wanted to document current best practices, rather than what was in TCP or even linux TCP.  I think most things came from linux TCP because that was what we were familiar with.

FWIW, agreed that Cubic wasn't an RFC at the time, but it was close to becoming one, and we did discuss it.  We decided that it was too complex and adds a ton more complexity to this document. We wanted to have something simple but good enough that implementations could have as a starting point, and then move on to more esoteric algorithms as deployments needed.

(I'll disagree that Cubic is the best known CC at the moment -- unless you're talking about DC environments, it's not.)

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/issues/2555#issuecomment-481936247