Re: [quicwg/base-drafts] Document TCP RTO vs QUIC PTO (#3441)

[Martin Thomson <[e-mail]>]

martinthomson commented on this pull request.



> @@ -228,6 +228,23 @@ QUIC endpoints measure the delay incurred between when a packet is received and
 when the corresponding acknowledgment is sent, allowing a peer to maintain a
 more accurate round-trip time estimate (see Section 13.2 of {{QUIC-TRANSPORT}}).
 
+### Probe Timeout Replaces RTO and TLP
+
+QUIC uses a probe timeout (see {{pto}}), with a timer based on TCP's
+RTO timeout.  QUIC PTO includes the peer's max_ack_delay in the calculation,
+instead of relying upon a fixed minimum RTO.  Unlike TCP's RTO, which collapses
+the congestion window upon expiry, QUIC does not change the congestion window
+and instead allows sending probe packets whenever the timer expires.  This is
+similar to TCP with F-RTO, but it does allow more packets to be sent when the
+congestion window was not fully utilized when the probe timeout expires. Though
+this is slightly more aggressive than TCP RTO, it's less aggressive than the
+congestion controller allowed.

Phrasing here suggests "than the congestion controller allows" or "than the congestion controller in TCP allowed" or something else.

-- 
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/3441#pullrequestreview-355639387