Re: [quicwg/base-drafts] congestion window increase on every ACKed packet could result in bursty sends (#3094)

[ianswett <[e-mail]>]

For the WG:  The proposed PR attempts to include a mechanism to limit CWND increase during slow start when not pacing.  I'll note that by itself, it does not guarantee we follow the existing MUST: 
 "Implementations MUST either use pacing or limit such
   bursts to the initial congestion window, which is recommended to be
   the minimum of 10 * max_datagram_size and max(2* max_datagram_size,
   14720)), where max_datagram_size is the current maximum size of a
   datagram for the connection, not including UDP or IP overhead."

I don't think we should add this to the pseudocode, because we say SHOULD pace packets and the pseudocode currently only covers SHOULDs, not MAYs.  Adding all MAYs to the pseudocode is too complex.

I don't have a concern about adding a reference to the existing RFC as a mechanism that MAY help limit bursts during slow start.

I'd like a decision from the WG on whether to include this change and if so, some criteria on what should be included going forward.

-- 
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/3094#issuecomment-555302900