Re: [quicwg/base-drafts] Should we allow ACK-only packets to be declared lost? (#3451)

[Igor Lubashev <[e-mail]>]

The text is slightly ambiguous on the topic of whether pure-ACK packets are to be acknowledged.

Section "13.2.  Generating Acknowledgements" implies that they are ACKed later:

> Packets that are not ack-eliciting are only acknowledged when an ACK frame is sent for other reasons.

Section "13.2.1.  Sending ACK Frames" implies that they SHOULD but not MUST be ACKed:

> Every packet SHOULD be acknowledged at least once

and

> [all] packets marked with the ECN Congestion Experienced (CE) codepoint in the IP header SHOULD be acknowledged immediately

It seems strange NOT to acknowledge all packets when sending an ACK frame. What benefit is there to NOT doing so? And if there is no benefit, why not require the protocol implementations to be more predictable?

What specific response a specific congestion controller has as a reaction to a lost / EC-marked pure-ACK packet is an implementation matter, of course.

-- 
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/3451#issuecomment-586744717