feat(net): multi-relay failover, BBRv3 congestion, wider flow windows - #21
Merged
Merged
Conversation
Internet best-practice pass on the transport layer: - --relay is now repeatable on rds and rds-agent. iroh probes all configured relays, homes on the lowest-latency one and fails over automatically — the upstream-recommended production topology is >=2 relays in different failure domains, now expressible. - Congestion control: BBRv3 (noq_proto::congestion::Bbr3Config) on both backends instead of loss-based Cubic — paced sends resist bufferbloat, which is what keeps desktop interactive latency low when a sync transfer shares the path. - stream_receive_window 4 MiB / send_window 32 MiB (upstream tunes for ~100 Mbps x 100 ms): a large keyframe or sync chunk stream no longer stalls on high-BDP links. - EndpointConfig.relay -> relays: Vec<RelayUrl>; with_relay() kept, with_relays() added; noq-proto is now a direct dep (already in the tree via iroh — zero new packages). - Verified live: agent+client on two --relay flags (one dead) — failover works, multipath selects the faster direct path.
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Internet best-practice pass on the transport layer (iroh production docs + quinn/noq tuning):
--relayrepeatable onrds/rds-agent— iroh probes all configured relays, homes on lowest-latency, fails over automatically. Upstream recommends ≥2 dedicated relays for production.noq_proto::congestion::Bbr3Config) on both backends instead of loss-based Cubic — paced sends resist bufferbloat, keeping desktop interactive latency low when a sync transfer shares the path.stream_receive_window4 MiB /send_window32 MiB (upstream defaults target ~100 Mbps × 100 ms) — large keyframes and sync chunk streams no longer stall on high-BDP links.EndpointConfig.relay→relays: Vec<RelayUrl>;with_relaysadded;noq-protopromoted to direct dep (already in tree via iroh — zero new packages).Test plan
--relayflags (one dead) — failover works, multipath selects faster direct pathGenerated with Devin