Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

More text on ACK rate in the applicability draft #386

Merged
merged 4 commits into from Jun 30, 2021
Merged

More text on ACK rate in the applicability draft #386

Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
d8a08a8
More text on ACK rate in the applicability draft
mirjak Jun 2, 2021
fce878f
white space of course...
mirjak Jun 2, 2021
36fcd33
Apply suggestions from Ian
mirjak Jun 2, 2021
2ac301e
More or less proposed text by MT
mirjak Jun 30, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
14 changes: 11 additions & 3 deletions draft-ietf-quic-applicability.md
View file
Open in desktop
Expand Up @@ -565,11 +565,19 @@ RESET_STREAM frames, which contain only an error code.

# ACK-only packets on constrained links

The stragegy on sending acknowlegdements in QUIC version 1 is adopted from
TCP, meaning that usually an QUIC implementation is expected to send one
packet containing an ACK frame for each two packets received.
The cost of sending acknowledgments - in processing cost or link
utilization - could be a significant proportion of available resources if
these resources are constrained. Reducing the rate at which acknowledgments
are generated can preserve these resources and improve overall performance,
for both network processing as well as application-relevant metrics.
these resources are constrained. "ACK-only" packets are not reliably
detectable in QUIC, so network ACK thinning done for TCP is not suitable
for QUIC traffic. Instead QUIC endpoint need to control the ACK rate
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

/instead QUIC endpoint need/instead a QUIC endpoint needs/

appropriately and some implenentation may even provide an interface to
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

/some implementation may even provide an interface to impact/implementations provide an interface that can reduce/ ??? ... although I like MT's comment also which avoids this.

impact the amount of ACK traffic sent. Reducing the rate at which
acknowledgments are generated in constrained situations can preserve resources
and improve overall performance, for both network processing as well as
application-relevant metrics.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I can't suggest over the diff in the way I'd like, but here is my take on this:

# Acknowledgment Efficiency

QUIC version 1 without extensions uses an acknowledgment strategy based on TCP.
That is, every other packet is acknowledged.  However, generating and processing
QUIC acknowledgments can consume significant resources, both in terms of link
utilization and processing time.  Some applications might be able to improve
overall performance by using alternative strategies that reduce the rate of
acknowledgments.

This drops the "thinning" text (that's for the other doc).

There is a lot more to unpack here, but the transport doc already talks about this topic in more precise terms.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That looks good, OK also to me.


# Port Selection and Application Endpoint Discovery {#ports}

Expand Down