Add PoC/example pure uring based UDP client and server #88
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
eb21339 to
b8bf495
Compare
|
Code coverage summary for 55bbabe: ✅ Region coverage 54% passes |
cbd1014 to
ef59f7e
Compare
3d6d4d1 to
daafa1a
Compare
These are not fully featured and are basically the bare minimum to get a working pair of UDP client and server. By building a pure I/O based event loop the hope was to cut out all the overhead relating to the use of async and make it possible to focus purely on the I/O bottlenecks.
daafa1a to
55bbabe
Compare
7 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Note
This is probably not for merging, more of a PoC.
Description
These are not fully featured and are basically the bare minimum to get a working pair of UDP client and server.
This builds on #87 since in this use case the allocations showed up in the flamegraphs and it was trivial to reuse buffers for both inside and outside I/O
This uses just 8 slots for rx of each of inside and outside traffic. Using any more degrades performance (I think because things get punted to iou-wrk threads instead of completing inline).
Motivation and Context
By building a pure I/O based event loop the hope was to cut out all the overhead relating to the use of async and make it possible to focus purely on the I/O bottlenecks.
How Has This Been Tested?
I ran some iperf tests using the containerized environment and also ran some speedtests with two machines directly connected with a cross over test against a nearby speedtest server.
iperf results were marginal/in the noise, speedtest results improved from ~1.4G/sec download to 1.6G/sec and ~700M/sec upload to 1.2G/sec. The latency was roughly halved though.