Re: [quicwg/base-drafts] Define under-utilizing the congestion window (#2675)

[Jana Iyengar <[e-mail]>]

janaiyengar commented on this pull request.

This still misses an important question: when a sender goes to increase the congestion window (upon receiving an ACK), what does it do? Should it not increase the cwnd because inflight < cwnd? There's discussion in #2554 which is still relevant to this.

> @@ -848,9 +850,10 @@ and not fully utilize the congestion window due to this delay. A sender
 should not consider itself application limited if it would have fully
 utilized the congestion window without pacing delay.
 
-Bursting more than an intial window's worth of data into the network might
-cause short-term congestion and losses. Implemementations SHOULD either use
-pacing or reduce their congestion window to limit such bursts.
+Sending multiple packets into the network without any delay between them
+creates a burst of load that might cause short-term congestion and losses.

```suggestion
creates a packet burst that might cause short-term congestion and losses.
```

> @@ -848,9 +850,10 @@ and not fully utilize the congestion window due to this delay. A sender
 should not consider itself application limited if it would have fully
 utilized the congestion window without pacing delay.
 
-Bursting more than an intial window's worth of data into the network might
-cause short-term congestion and losses. Implemementations SHOULD either use
-pacing or reduce their congestion window to limit such bursts.
+Sending multiple packets into the network without any delay between them
+creates a burst of load that might cause short-term congestion and losses.
+Implemementations SHOULD either use pacing or reduce their congestion window

```suggestion
Implementations SHOULD either use pacing or reduce their congestion window
```

> @@ -848,9 +850,10 @@ and not fully utilize the congestion window due to this delay. A sender
 should not consider itself application limited if it would have fully
 utilized the congestion window without pacing delay.
 
-Bursting more than an intial window's worth of data into the network might
-cause short-term congestion and losses. Implemementations SHOULD either use
-pacing or reduce their congestion window to limit such bursts.
+Sending multiple packets into the network without any delay between them
+creates a burst of load that might cause short-term congestion and losses.

I don't think this description of burst is necessary though. It doesn't help with #2686, if that's what you were trying to address. I would drop this, but I realized that the existing statement is potentially misleading in what it says. (more discussion on this in #2686)

-- 
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/2675#pullrequestreview-236344073