Implements FlexFEC-03 (RFC 8627) end to end in the SFU, in both directions:
- Upstream: decode FlexFEC repair packets from publishers and recover lost
media in the receive buffer before forwarding (pkg/sfu/flexfec/decoder.go,
pkg/sfu/buffer). Recovered packets are surfaced via Prometheus counters.
- Downstream: generate FlexFEC for subscribers off the downtrack, paced
alongside media (pkg/sfu/downtrack_fec.go, pkg/sfu/pacer).
- Negotiation: advertise and match flexfec-03 in the media engine and
transport SDP, gated by new config knobs (pkg/rtc, pkg/config,
config-sample.yaml).
- Telemetry: livekit_fec_* metrics for sent/received/recovered packets
(pkg/telemetry/prometheus/packets.go).
Tests: unit coverage for the decoder, buffer recovery, downtrack generation,
and transport negotiation, plus an end-to-end integration test and the
test-client support it needs (test/flexfec_test.go, test/client).
* Add TURN permission handler.
- Turn off permissions to private/link local/multicast and internal IPs
- Add a list of CIDRs that can be used for more things to deny
permission to.
* unused
* add config for allowing private IPs, used in testing
* add a TTL to user name and use it to auth
* allow list for restricted peer CIDRs
* Added optional "Power of Two Random Choices" algorithm for the node selector sort_by feature. The current, default behavior of picking the lowest-valued node remains.
* Support for participant attributes
* move metadata setters to LocalParticipant
* address feedback
* forward error
* update go mod
* update attributes first
* Buffer size config for video and audio.
There was only one buffer size in config.
In upstream, config value was used for video.
Audio used a hard coded value of 200 packets.
But, in the down stream sequencer, the config value was used for both
video and audio. So, if video was set up for high bit rate (deep
buffers), audio sequencer ended up using a lot of memory too in
sequencer.
Split config to be able to control that and also not hard code audio.
Another optimisation here would be to not instantiate sequencer unkess
NACK is negotiated.
* deprecate packet_buffer_size
* Add control of playout delay
Add config to enable playout delay. The delay will be limited by
[min,max] in the config option and calculated by upstream & downstream
RTT.
* check protocol version to enable playout delay
* Move config to room, limit playout-delay update interval, solve comments
* Remove adaptive playout-delay
* Remove unused config
* API execution timeout is now configurable
In certain environments, it can take longer than the default 2s to
fully execute API requests. Making execution timeout a configurable option.
* do not expose api to YAML. internal for now.
David Chen