Re: [quicwg/base-drafts] Byte counting at Congestion avoidance (#3917)

[Jana Iyengar <[e-mail]>]

The text currently says that the congestion window MUST NOT increase by more than 1 max datagram size per acknowledged cwnd. The pseudo code currently reflects what the text says. In that sense, the pseudo code is correct.

This PR is an alternative implementation of the text. As such, it doesn't fix a problem in the text, it fixes a problem with implementations that use integers.

The pushback on this PR is that the behavior is different than the existing pseudo code. Broadly, I agree with that, though, as I have said earlier, it shouldn't matter in practice. That said, the difference in behavior exists. The pseudo code in the draft is meant as clarification of the text, and implementations will and do differ from the pseudo code in various places.

My suggestion then would be to drop this PR, and instead add a comment above the concerning line saying something like: "Implementers that use an integer representation for the congestion window should be careful with the following division, and can use the alternative approach suggested in Section 2.1 of RFC 3465."

-- 
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/pull/3917#issuecomment-687481216