fix: handle network changes in multipath by divagant-martian · Pull Request #383 · n0-computer/noq

divagant-martian · 2026-01-29T21:55:02Z

Description

Adds proto an quinn level apis for dealing with netwokr changes.

Adds Connection::handle_network_change(hint, now) to let applications signal
that the underlying network has changed.

Behavior:

Non-multipath: Rotates the remote CID and pings for liveness (RFC9000
migration) This might require more love but RFC9000 migrations are out of
scope for this PR
Multipath, client-side: Classifies each path as recoverable or
non-recoverable. In multipath the standard behaviour is to consider paths non
recoverable and move then to new PathIds. The optional NetworkChangeHint
gives special treatment to paths that survive the network change so that
these retain the RFC9000 migration behaviour.
Multipath, server-side: Always attempts recovery (servers have stable
addresses in theory and should not migrate).
During handshake: Just updates the remote CID and pings.
All cases: local ip of the path is cleared. We need to figure it out from
scratch.

Public API:

Adds NetworkChangeHint
Connection::handle_network_change at proto level is added.
Endpoint::handle_network_change is added in quinn. This just broadcasts to
all connections using the above.

Also includes:

Bugfix: don't emit PathOpened events for non-multipath connections. This has
some pending discussions but seemed the best choice right now.
Tests for single-path recovery, multipath replacement, selective hints.

Breaking Changes

in proto::Connection local_address_changed is gone, and it's replaced with
handle_network_change.

Notes & open questions

n/a

github-actions · 2026-01-29T21:55:55Z

Performance Comparison Report

cdeb5e79319d77850ef0a9db2cc8d1fc894dae87 - artifacts

No results available

---

f14566eb40cbf7fde8a26bc4f40b73f7cd924bf6 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5303.4 Mbps	7897.0 Mbps	-32.8%	92.6% / 97.5%
medium-concurrent	5546.9 Mbps	7903.1 Mbps	-29.8%	96.9% / 122.0%
medium-single	4040.9 Mbps	4476.5 Mbps	-9.7%	94.8% / 122.0%
small-concurrent	3637.2 Mbps	5182.3 Mbps	-29.8%	92.6% / 100.0%
small-single	3494.8 Mbps	4690.7 Mbps	-25.5%	95.4% / 126.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2911.9 Mbps	3643.4 Mbps	-20.1%
lan	768.9 Mbps	796.4 Mbps	-3.4%
lossy	69.8 Mbps	69.8 Mbps	~0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 25.6% slower on average

---

21c45f304150b4893695759731af69bbb9608a07 - artifacts

No results available

---

284690ed7799ab3c9c5defe17d4d9caf810475c8 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5338.1 Mbps	7713.4 Mbps	-30.8%	91.9% / 97.2%
medium-concurrent	5255.7 Mbps	7606.4 Mbps	-30.9%	92.9% / 97.6%
medium-single	4071.3 Mbps	4189.9 Mbps	-2.8%	95.6% / 112.0%
small-concurrent	3681.2 Mbps	5120.1 Mbps	-28.1%	97.5% / 116.0%
small-single	3341.3 Mbps	4408.4 Mbps	-24.2%	91.4% / 99.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2992.8 Mbps	3687.6 Mbps	-18.8%
lan	782.4 Mbps	796.4 Mbps	-1.8%
lossy	69.8 Mbps	55.9 Mbps	+24.9%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 23.9% slower on average

---

2a4ec09a13dee967107f307fea859c547fee5b4f - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5291.7 Mbps	7835.9 Mbps	-32.5%	93.7% / 98.3%
medium-concurrent	5386.5 Mbps	7661.3 Mbps	-29.7%	98.3% / 177.0%
medium-single	3813.7 Mbps	4100.7 Mbps	-7.0%	87.1% / 96.5%
small-concurrent	3791.6 Mbps	5061.0 Mbps	-25.1%	96.3% / 128.0%
small-single	3475.0 Mbps	4377.8 Mbps	-20.6%	96.9% / 126.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2835.6 Mbps	3939.1 Mbps	-28.0%
lan	768.4 Mbps	801.7 Mbps	-4.2%
lossy	69.8 Mbps	55.9 Mbps	+25.0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 24.8% slower on average

---

dd78fc69458144461b072070ed8ef5f1f47019b1 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5309.8 Mbps	8012.5 Mbps	-33.7%	90.3% / 97.0%
medium-concurrent	5346.0 Mbps	7665.7 Mbps	-30.3%	99.0% / 189.0%
medium-single	3548.0 Mbps	4748.6 Mbps	-25.3%	85.7% / 95.6%
small-concurrent	3736.5 Mbps	5392.5 Mbps	-30.7%	88.0% / 97.8%
small-single	3363.4 Mbps	4733.5 Mbps	-28.9%	97.1% / 133.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	3006.7 Mbps	N/A	N/A
lan	782.4 Mbps	N/A	N/A
lossy	69.8 Mbps	N/A	N/A
wan	83.8 Mbps	N/A	N/A

Summary

iroh-quinn is 30.3% slower on average

---

1a4f55324b3e2da13c940ff5ca01f0dbf3227258 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5345.3 Mbps	7940.6 Mbps	-32.7%	97.1% / 149.0%
medium-concurrent	5428.9 Mbps	7919.1 Mbps	-31.4%	96.5% / 149.0%
medium-single	4261.3 Mbps	4749.7 Mbps	-10.3%	91.9% / 101.0%
small-concurrent	3872.7 Mbps	4920.5 Mbps	-21.3%	95.1% / 103.0%
small-single	3516.8 Mbps	4799.7 Mbps	-26.7%	97.6% / 153.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	N/A	3701.8 Mbps	N/A
lan	N/A	796.5 Mbps	N/A
lossy	N/A	58.8 Mbps	N/A
wan	N/A	83.8 Mbps	N/A

Summary

iroh-quinn is 26.1% slower on average

---

1e9e567583f5e924185ab6a0d67174147636f197 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5385.5 Mbps	7732.1 Mbps	-30.3%	96.3% / 98.2%
medium-concurrent	5134.5 Mbps	7460.9 Mbps	-31.2%	95.0% / 97.0%
medium-single	4089.1 Mbps	4072.2 Mbps	~0%	96.2% / 98.4%
small-concurrent	3585.0 Mbps	5098.4 Mbps	-29.7%	94.9% / 97.3%
small-single	3453.1 Mbps	4431.5 Mbps	-22.1%	96.9% / 98.7%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2835.5 Mbps	4057.9 Mbps	-30.1%
lan	782.5 Mbps	810.4 Mbps	-3.4%
lossy	69.9 Mbps	69.9 Mbps	~0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 24.8% slower on average

---

218441309342d61fa655f0c6929ced7804e317d1 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5440.6 Mbps	7898.7 Mbps	-31.1%	95.7% / 101.0%
medium-concurrent	5424.5 Mbps	7890.6 Mbps	-31.3%	94.7% / 101.0%
medium-single	4086.1 Mbps	4742.2 Mbps	-13.8%	99.1% / 153.0%
small-concurrent	3668.4 Mbps	5246.6 Mbps	-30.1%	90.8% / 98.8%
small-single	3564.5 Mbps	4797.3 Mbps	-25.7%	95.2% / 104.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2835.3 Mbps	3882.9 Mbps	-27.0%
lan	772.2 Mbps	796.4 Mbps	-3.0%
lossy	69.9 Mbps	55.9 Mbps	+25.0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 26.7% slower on average

---

adb3c993738d022cb12ecf667fbfb539dc619aa3 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5601.9 Mbps	7652.7 Mbps	-26.8%	96.8% / 98.5%
medium-concurrent	5343.5 Mbps	7874.9 Mbps	-32.1%	95.6% / 97.6%
medium-single	3833.8 Mbps	4652.4 Mbps	-17.6%	95.2% / 97.7%
small-concurrent	3592.7 Mbps	5236.1 Mbps	-31.4%	95.2% / 99.3%
small-single	3536.2 Mbps	4779.8 Mbps	-26.0%	95.5% / 97.9%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	3042.7 Mbps	3701.5 Mbps	-17.8%
lan	782.4 Mbps	796.4 Mbps	-1.8%
lossy	69.9 Mbps	63.5 Mbps	+9.9%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 25.7% slower on average

---

c456958584d19f87849daca43d719131f6fcfebe - artifacts

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	N/A	3687.7 Mbps	N/A

---

e84b1821ea761aa5cd9153b28c075a4a77d0f45e - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5911.2 Mbps	8028.6 Mbps	-26.4%	96.4% / 98.2%
medium-concurrent	5669.7 Mbps	7669.1 Mbps	-26.1%	96.3% / 98.0%
medium-single	4228.0 Mbps	4623.5 Mbps	-8.6%	96.4% / 98.1%
small-concurrent	3681.1 Mbps	5097.3 Mbps	-27.8%	96.0% / 100.0%
small-single	3581.8 Mbps	4669.5 Mbps	-23.3%	96.9% / 99.3%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2779.1 Mbps	4136.3 Mbps	-32.8%
lan	774.8 Mbps	810.4 Mbps	-4.4%
lossy	69.8 Mbps	55.9 Mbps	+25.0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 23.9% slower on average

---

eb650b604b67d2492e015799d58ba16b3eb44d39 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	6000.0 Mbps	7897.6 Mbps	-24.0%	97.4% / 98.7%
medium-concurrent	5540.2 Mbps	7969.0 Mbps	-30.5%	95.3% / 97.2%
medium-single	4024.2 Mbps	4712.3 Mbps	-14.6%	95.8% / 98.2%
small-concurrent	3647.1 Mbps	5168.7 Mbps	-29.4%	95.1% / 97.6%
small-single	3547.3 Mbps	4693.2 Mbps	-24.4%	96.0% / 98.4%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2849.0 Mbps	4104.7 Mbps	-30.6%
lan	771.3 Mbps	810.3 Mbps	-4.8%
lossy	69.9 Mbps	69.8 Mbps	~0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 25.3% slower on average

---

4be21998c085b16dec8ec9fd5830bd0cda45e9a0 - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5571.6 Mbps	7873.6 Mbps	-29.2%	96.0% / 150.0%
medium-concurrent	5516.2 Mbps	7864.8 Mbps	-29.9%	96.3% / 151.0%
medium-single	4124.2 Mbps	4749.0 Mbps	-13.2%	90.8% / 100.0%
small-concurrent	3778.8 Mbps	5026.7 Mbps	-24.8%	95.3% / 104.0%
small-single	3494.0 Mbps	4760.8 Mbps	-26.6%	97.9% / 154.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2898.5 Mbps	4078.6 Mbps	-28.9%
lan	772.2 Mbps	810.3 Mbps	-4.7%
lossy	69.9 Mbps	55.9 Mbps	+25.1%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 25.5% slower on average

---

424fc3668bf4865ef8d6b439dff43eac2e04212d - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5481.2 Mbps	7795.3 Mbps	-29.7%	96.1% / 104.0%
medium-concurrent	5334.3 Mbps	7139.6 Mbps	-25.3%	95.7% / 104.0%
medium-single	4149.2 Mbps	4155.5 Mbps	~0%	98.9% / 155.0%
small-concurrent	3842.4 Mbps	5028.5 Mbps	-23.6%	92.9% / 107.0%
small-single	3475.4 Mbps	4358.4 Mbps	-20.3%	95.7% / 105.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	N/A	4080.5 Mbps	N/A
lan	N/A	810.3 Mbps	N/A
lossy	N/A	55.9 Mbps	N/A
wan	N/A	83.8 Mbps	N/A

Summary

iroh-quinn is 21.8% slower on average

---

4c4996d46e85b24c4b94439021785bf1bf09191a - artifacts

Raw Benchmarks (localhost)

Scenario	iroh-quinn	upstream	Delta	CPU (avg/max)
large-single	5583.5 Mbps	7786.0 Mbps	-28.3%	96.8% / 98.3%
medium-concurrent	5479.7 Mbps	7888.2 Mbps	-30.5%	95.2% / 97.0%
medium-single	4346.6 Mbps	4602.9 Mbps	-5.6%	96.8% / 99.2%
small-concurrent	3634.1 Mbps	5164.7 Mbps	-29.6%	96.1% / 98.7%
small-single	3477.6 Mbps	4707.4 Mbps	-26.1%	98.5% / 100.0%

Netsim Benchmarks (network simulation)

Condition	iroh-quinn	upstream	Delta
ideal	2855.5 Mbps	3705.1 Mbps	-22.9%
lan	782.4 Mbps	796.5 Mbps	-1.8%
lossy	69.8 Mbps	55.9 Mbps	+25.0%
wan	83.8 Mbps	83.8 Mbps	~0%

Summary

iroh-quinn is 24.4% slower on average

github-actions · 2026-01-29T21:57:58Z

Documentation for this PR has been generated and is available at: https://n0-computer.github.io/quinn/pr/383/docs/iroh_quinn/

Last updated: 2026-02-10T18:48:32Z

codecov-commenter · 2026-02-03T16:15:54Z

Codecov Report

❌ Patch coverage is 66.66667% with 31 lines in your changes missing coverage. Please review.
✅ Project coverage is 77.03%. Comparing base (d29726a) to head (4c4996d).

Files with missing lines	Patch %	Lines
quinn-proto/src/connection/mod.rs	76.92%	18 Missing ⚠️
quinn/src/endpoint.rs	0.00%	10 Missing ⚠️
quinn/src/connection.rs	40.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #383      +/-   ##
==========================================
- Coverage   77.24%   77.03%   -0.22%     
==========================================
  Files          81       81              
  Lines       23142    23228      +86     
==========================================
+ Hits        17877    17893      +16     
- Misses       5265     5335      +70

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

flub

I know it's still in draft, but you said you were only really working on tests so I already had a look at the non-test code... hope that's helpful.

quinn-proto/src/connection/mod.rs

quinn-proto/src/tests/multipath.rs

quinn/src/endpoint.rs

quinn/src/lib.rs

flub

Thanks, I think this all makes sense. The wording of the comments is better now as well.

quinn-proto/src/connection/mod.rs

deal with network changes

cdeb5e7

divagant-martian added the multipath QUIC Multipath extension label Jan 29, 2026

divagant-martian self-assigned this Jan 29, 2026

n0bot bot added this to iroh Jan 29, 2026

github-project-automation bot moved this to 🚑 Needs Triage in iroh Jan 29, 2026

dignifiedquire moved this from 🚑 Needs Triage to 🏗 In progress in iroh Jan 30, 2026

dignifiedquire added this to the quinn: iroh 0.97 milestone Jan 30, 2026

divagant-martian linked an issue Feb 3, 2026 that may be closed by this pull request

bug: deal with network changes in multipath #376

Closed

divagant-martian added 4 commits February 3, 2026 09:55

Merge remote-tracking branch 'n0/main' into handle-network-changes

374ab69

expose new trait

f14566e

adjust test

21c45f3

spelling

284690e

Merge remote-tracking branch 'n0/main' into handle-network-changes

2a4ec09

flub reviewed Feb 3, 2026

View reviewed changes

divagant-martian added 9 commits February 5, 2026 11:50

add public api

814d528

update behaviour based on review

2d399e0

Merge remote-tracking branch 'n0/main' into handle-network-changes

eb47216

remove the need for a default hint value

93b3144

provide methods to get or not a hint

dd78fc6

so many annoying changes

1a4f553

fixes

1e9e567

remove mergetool file

2184413

reduce awkwardness

adb3c99

divagant-martian marked this pull request as ready for review February 5, 2026 19:15

divagant-martian requested a review from flub February 5, 2026 19:16

divagant-martian added 2 commits February 5, 2026 14:29

go back to a single method for even less awkwardness

c456958

clippy and fmt

e84b182

flub reviewed Feb 6, 2026

View reviewed changes

divagant-martian added 10 commits February 9, 2026 11:18

Merge remote-tracking branch 'n0/main' into handle-network-changes

eb650b6

make docs clearer (?), include client-side note

c0d0b2d

update API name to match that of proto

b9eed46

cleanup proto api by moving async bounds to the async crate

80d4621

aim for more clarity

4be2199

more failsafes

ceb5d85

public import

4d9a495

add network change test for the vanilla quic case

3ea13cb

bugfix: do not send pathopen events for non-multipath

43a4649

add test for multipath network changes without a hint

6b05835

flub approved these changes Feb 10, 2026

View reviewed changes

quinn-proto/src/connection/mod.rs Show resolved Hide resolved

divagant-martian added 3 commits February 10, 2026 11:03

Merge remote-tracking branch 'n0/main' into handle-network-changes

424fc36

add a test for a selective network change hint

8b9e666

add some logs

4c4996d

divagant-martian added this pull request to the merge queue Feb 10, 2026

Merged via the queue into main with commit d5580a5 Feb 10, 2026
36 checks passed

divagant-martian deleted the handle-network-changes branch February 10, 2026 19:15

github-project-automation bot moved this from 🏗 In progress to ✅ Done in iroh Feb 10, 2026

Conversation

divagant-martian commented Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Breaking Changes

Notes & open questions

Uh oh!

github-actions bot commented Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Performance Comparison Report

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Netsim Benchmarks (network simulation)

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Raw Benchmarks (localhost)

Netsim Benchmarks (network simulation)

Summary

Uh oh!

github-actions bot commented Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov-commenter commented Feb 3, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

flub left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

flub left a comment

Choose a reason for hiding this comment

divagant-martian commented Jan 29, 2026 •

edited

Loading

github-actions bot commented Jan 29, 2026 •

edited

Loading

github-actions bot commented Jan 29, 2026 •

edited

Loading

codecov-commenter commented Feb 3, 2026 •

edited

Loading