[quicwg/base-drafts] recovery: Clarifying the priority of Probe packet (#3780)

[Junho Choi <[e-mail]>]

Currently "Sending Probe Packets" define the priority what to be in Probe packet as follows:

- A probe packet SHOULD carry new data when possible.
- A probe packet MAY carry retransmitted unacknowledged data when new data is unavailable.

And also says "Implementations MAY use alternative strategies for determining the content of probe packets, including sending new or retransmitted data based on the application's priorities."

When there is no new data, we need to choose which one to retransmit when there is a multiple data loss in the sender. Since QUIC's Probe packet approach is a combined experience of TCP TLP and RTO and I believe Probe packet is more similar to TLP. RACK draft says (https://datatracker.ietf.org/doc/html/draft-ietf-tcpm-rack-08#section-7.3) "The probe data segment can either be new, previously unsent data, or a retransmission of previously sent data just below SND.NXT" so it's pretty clear which one to send among the lost data (newest one).

Is there any reason that it's up to the implementers when it need to choose which one to send from lost data? It's better to clarify which one to send among retransmitted unacknowledged data, from oldest or newest when there is multiple data packet lost.

I think it's better to send newest (unacked) one first to follow the reasoning of TLP (new to the old), but there may be some reason why the current draft doesn't define it clearly.

https://github.com/cloudflare/quiche/pull/562#discussion_r441864077 has some background and my comment.

-- 
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/3780