NetworkStateEstimator is not used by WebRTC from receive side.
ReceiveSidesCongestionController::SetTransportOverhead is not needed either since NetworkStateEstimator is removed.
Note, CongestionControlFeedbackGenerator is used with RFC 8888 only and feedback frequency will be refactored in later cl.
Bug: webrtc:42220808
Change-Id: I08980aa19117e1de7a9b7896d05d07715dd9f962
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/375460
Auto-Submit: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Commit-Queue: Björn Terelius <terelius@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43821}
With this CL, the decision if an RTP packet should be sent as ect(1) is made in RtpControllerSend depending on if RFC 8888 has been negotiated and if CCFB is received with ECN enabled.
Since webrtc does not yet adapt to ECN feedback, packets are sent as ECT(1) until the first feedback is received.
Change-Id: Iddf63849328afbe54a7c8f921f2e8db134aeff6a
Bug: webrtc:42225697
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/367388
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43609}
If max allocated bitrate change, default max limit probe to 2x current
BWE.
Bug: webrtc:369044000, b/370883514
Change-Id: Ibaf79fff94157186002728828d6574bea21afd24
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/368820
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43448}
This will turn on RFC 8888 feedback messages if "ack ccfb" is negotiated.
This should eliminate the need for the "force" flag in the field trial.
Bug: webrtc:42225697
Change-Id: Iec7a894c244a417a8499200861550a33f89966a2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/367400
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43398}
Increased the number of errors the automation is fixing to 150 from
75 in this commit.
Bug: webrtc:370878648
Change-Id: If6e6a5f40db7eb54c27c1a85fb7031838e478c70
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/366205
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Dor Hen <dorhen@meta.com>
Cr-Commit-Position: refs/heads/main@{#43337}
Prior to this change, we unnecessarily re-compute AverageReportedPacketLossRatio a few times for the same observation.
Bug: webrtc:12707
Change-Id: I5de3d06bf3a32a39638a6f72baf7ec12144c6399
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/366583
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Diep Bui <diepbp@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43298}
In the `MaybeProcess` method, a new variable `time_until_cc_rep` is introduced
to track the time until congestion control feedback generation is processed.
The minimum of this value and the times until RBE and transport sequence number
feedback processing are calculated.
Co-authored-by: Shridhar Majali <smajali@nvidia.com>
Bug: webrtc:42225697
Change-Id: I44173062d8f8f84bf7e7791e05578c0ffc4fd017
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/365273
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43271}
Directly use RtpPacketToSend instead of RtpPacketSendInfo
For RFC8888 we will need to match SSRC and RTP sequence number with the transport sequence number that only exist on the sending side.
For retransmitted packets, RTX, RtpPacketSendInfo contain original SSRC and original sequence number.
Bug: webrtc:42225697
Change-Id: Iafa5d851ca5c51c85e4607ed4c1919d96da6084a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/366000
Auto-Submit: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Björn Terelius <terelius@webrtc.org>
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43264}
The split shows that some places don't need it at all. Most other
places will depend on both send and receive stream targets.
Bug: webrtc:373151158
Change-Id: I788136a2ee84180c16345a7929b7f7bf3f97507b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/365460
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43230}
This reverts commit 37458ce40a1741f9d5358d49fe49beed20140502.
Reason for revert: Will be wired up as an experiment instead.
Original change's description:
> Per defaul probe max to 2x current BWE if max total allocated bitrate change
>
> This aligns to probe limits in ALR for example.
>
> Bug: webrtc:369044000, b/369021234
> Change-Id: I3823b308cf97a8b7060b35b2ac38864e75d6f983
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/363301
> Reviewed-by: Jonas Oreland <jonaso@webrtc.org>
> Reviewed-by: Diep Bui <diepbp@webrtc.org>
> Commit-Queue: Per Kjellander <perkj@webrtc.org>
> Cr-Commit-Position: refs/heads/main@{#43074}
Bug: webrtc:369044000, b/369021234
Change-Id: I22b457254c9c21d2d951af2bda01a349ef83b3c7
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/364242
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Ranveer Aggarwal <ranvr@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43130}
Those trigger new warnings when importing the Chromium roll
Bug: None
Change-Id: Ica71cc83f5bbfd8fec4736185d389b9e82f2276e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/363740
Commit-Queue: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Harald Alvestrand <hta@webrtc.org>
Auto-Submit: Florent Castelli <orphis@webrtc.org>
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43080}
If the upper limit is infinite, dont probe.
Bug: webrtc:42224658
Change-Id: Ia662cceded83969ec11ee013adb2100f983fbd13
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/363660
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43079}
Add field trial parameter for setting scale factor of max allocated bitrate used for deciding when to skip probing.
Currently, a factor of 2 is used in most places for max allocated bitrate but not if the field trial skip_estimate_larger_than_fraction_of_max is used.
The purpose of this new field parameter is to be able to harmonize and always use the same factor.
Bug: webrtc:42224658
Change-Id: I5e1580b9bb18ef881b819affc0b4038094e94316
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/363400
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43078}
This aligns to probe limits in ALR for example.
Bug: webrtc:369044000, b/369021234
Change-Id: I3823b308cf97a8b7060b35b2ac38864e75d6f983
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/363301
Reviewed-by: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43074}
Ensure probing is not instantiated again until after timeout if a probe has been sent to max rate.
Bug: webrtc:42224658
Change-Id: I7d0d2edcfa81b1b454ea5748962af5a2070b347c
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/363240
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#43068}
If all packets are dropped for a period of time, an observation window will have the same length as the period when packets are dropped.
If later, no packets are lost, there is no point in loss based bwe backing down.
Therefore, ignore the observation with most loss and least loss when calculating an instant upper bound.
Bug: webrtc:42222865
Change-Id: I1d0125d6c76e68018b2aec1ecaa9b65729963136
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/356380
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@google.com>
Cr-Commit-Position: refs/heads/main@{#42772}
The idea is to reduce the risk of calculating a packet as lost if a
packet is reordered between two feedback reports.
It works as long as the recevied feedback does not complete an
observation.
Bug: webrtc:42222865 b/349765923
Change-Id: Iaf1595e624f546951baf3998d161f4cd1d5d491b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/355942
Reviewed-by: Diep Bui <diepbp@google.com>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42562}
std::optional<T>::emplace() without an initializer is broken on clang++
with gnu libstdc++. this workarounds the bug by removing the
absl::optional wrapping, which is actually pointless.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101227
Bug: chromium:41455655
Change-Id: I05354e57cc4cdda3fa6d3cd23f46462b69cc3bee
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/355900
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42561}
Ensure OnNetworkStateEstimate behaves the same way as internal networks state updates.
Also, ignore OnNetworkStateEstimate if an internal estimator exist.
Bug: webrtc:10742
Change-Id: I7967d202381250c406824fb2d0574bb95d2cd592
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/354102
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@google.com>
Cr-Commit-Position: refs/heads/main@{#42456}
With this cl, sending can be forced with field trial "WebRTC-RFC8888CongestionControlFeedback/force_send:true/"
In the future, ReceiveSideCongestionController::EnablSendCongestionControlFeedbackAccordingToRfc8888 if RFC 8888 has been negotiated.
Bug: webrtc:42225697
Change-Id: Ib09066aa89ca7b3fffc551da541090c69ab8d75f
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/352720
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42413}
CongestionControlFeedbackGenerator collect receive time information about received
packets and sends feedback according to RFC8888
Bug: webrtc:42225697
Change-Id: I70b7f7322fd262f99f45fd56b6eb8630a11b30c9
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/351543
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Auto-Submit: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42404}
BWE logging has as far as I know know been used for a long time. RTC event logs are the prefered method of logging.
Removed since it causes some BUILD pain.
For debugging the metrics API https://source.chromium.org/chromium/chromium/src/+/main:third_party/webrtc/api/test/metrics/ can be used instead.
Bug: webrtc:343347276
Change-Id: I046b58d880faabfadbc22269b0392fdd644155fc
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/352602
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Auto-Submit: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Jeremy Leconte <jleconte@webrtc.org>
Commit-Queue: Jeremy Leconte <jleconte@webrtc.org>
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42402}
This is done to better reflect the responsibility of the class.
The implementation implement a new interface FeedbackGeneratorInterface. The purpose of the interface is to allow a new implementation that supports RFC 8888.
Bug: webrtc:42225697
Change-Id: Id087dd7422abbcd6016693c076a65f4c4efd5712
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/351280
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42366}
The test tests that a route change does not cause BWE do drop unless the adapter is changed.
Bug: webrtc:42221538
Change-Id: I49be55172aff285c55d2564ec3389f3fc7c01d62
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/350820
Reviewed-by: Jonas Oreland <jonaso@webrtc.org>
Reviewed-by: Jonas Oreland <jonaso@google.com>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42347}
Repeated initial probes are sent every second until
ProbeController::OnMaxAllocatedBitrate is invoked (Media is beeing sent) or 5s has passed.
Each probe has a duration of 100ms, sent in packet bursts every 20ms.
ProbeController::waiting_for_initial_probe_result_ is no longer needed
and is removed.
Setting field trial for duration between probe packets bursts are moved
from BitrateProber to ProbeController.
Bug: webrtc:14928
Change-Id: I3170533f679fc2cc2aa5402e248fa1f6996d3ccd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/350640
Reviewed-by: Diep Bui <diepbp@google.com>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42323}
This makes Environment and thus field trials a required parameter, thus
RemoteBitrateEstimators member no longer need to fallback to the global field trial string.
Bug: webrtc:42220378
Change-Id: Ieb6ff442d5fde5fa9715573c758a7e078f0ceea4
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/349922
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42314}
Environment guarantees field trials are provided, thus GoogCcNetworkController doesn't need to fallback to the global field trials.
Bug: webrtc:42220378
Change-Id: Iff8e00504b43b074dc41b5ac9908fd0e2be18959
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/350540
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42300}
Replace factory that takes optional FieldTrialView with a constructor that takes non-optional reference to the same interface - all callers already guarantee it is not nullptr
Replace several local IsEnabled/IsDisabled helpers with the same helpers in FieldTrialView
In CongestionWindowPushbackController tests pass field trials bypassing global field trial string
Bug: webrtc:42220378
Change-Id: Ic49ad78919d834a5e3b9b69545d3b39088023a75
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/349900
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42270}
This would allow network controllers, GoogCcNetworkController in particular to have access to Environment at construction and thus it can rely on propagated field trials and won't need to fallback to the global field trial string
Bug: webrtc:42220378
Change-Id: I36099522e3866a89a8c8d6303da03f7d5b1cad8e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/350201
Reviewed-by: Björn Terelius <terelius@webrtc.org>
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42260}
This adds a new mode to the ProbeController that sends probes every
second the first 5seconds. Intented usage is before normal media has
started flowing. If enabled, the max allocated bitrat is ignored during
these initial probes.
The purpose is to have a more accurate estimate at the beginning of a
call.
The cl also removes ProbeController::SetFirstProbeToMaxBitrate.
Bug: webrtc:14928
Change-Id: I04feefb2f1b82ff38b35ee2be810f9119c53536a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/349924
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42252}
Environment includes propagated field trials that can be later passed to
RemoteBitrateEstimators member, and would allow not to rely on the global field trial string
Bug: webrtc:42220378
Change-Id: Icf75a433c20352b2c22829c2148c92f69a2517aa
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/349645
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42242}
If pacing rate, (current loss based bwe * pacing factor) is larger than the current upper link capacity estimate, reduce pacing factor to max of current bwe and upper link capacity.
Bug: webrtc:42220543
Change-Id: I5246da1f38530f8d411e7314adaa8651fc848f48
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/349601
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42210}
This is to avoid the case where the initial probe fail and the BWE is not updated, which can lead to a long period of low bandwidth.
Bug: webrtc:14928
Change-Id: Ie8f84270507b59995d57e4ab6e2a984570191529
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/349580
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42208}
This reverts commit 501c4f37bfee47b26999ee291c5355ad64554df7.
Patch set 1 contains pure reland.
The reason why we want to do this is because audio can allocate a needed bitrate before video when starting a call, which may lead to a race between the first probe result and updating the allocated bitrate.
That is the, initial probe will try to probe up to the max configured bitrate.
Bug: webrtc:14928
Change-Id: I6a8660da20ac54237f04a29461e03b31bd988bb0
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/347643
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Reviewed-by: Erik Språng <sprang@webrtc.org>
Reviewed-by: Erik Språng <sprang@google.com>
Owners-Override: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42086}
This reverts commit 33cc83595ae7dd144c57c614fb62d286d9d7bf68.
Reason for revert: Perf bots showed that this cl cause a change in metrics. It looks like it is for the better, but we want this to be behind a field trial.
Original change's description:
> Ignore allocated bitrate during initial exponential BWE.
>
> The reason why we want to do this is because audio can allocate a needed bitrate before video when starting a call, which may lead to a race between the first probe result and updating the allocated bitrate.
> That is the, initial probe will try to probe up to the max configured bitrate.
>
> ProbeController::SetFirstProbeToMaxBitrate will allow the first probe to
> continue up to the max configured bitrate, regardless of of the max
> allocated bitrate.
>
> Bug: webrtc:14928
> Change-Id: I6e0ae90e21a78466527f3464951e6033dc846470
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/346760
> Reviewed-by: Diep Bui <diepbp@webrtc.org>
> Commit-Queue: Per Kjellander <perkj@webrtc.org>
> Reviewed-by: Erik Språng <sprang@webrtc.org>
> Reviewed-by: Per Kjellander <perkj@webrtc.org>
> Cr-Commit-Position: refs/heads/main@{#42049}
Bug: webrtc:14928
Change-Id: I56ba58560b6857b6069552c02df822691f7af64d
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/347622
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Per Kjellander <perkj@webrtc.org>
Reviewed-by: Diep Bui <diepbp@webrtc.org>
Owners-Override: Per Kjellander <perkj@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#42081}
