Re: [webrtc-pc] Why is RTCIceCandidatePairStats.totalRoundTripTime mandatory to implement, but not responsesReceived? (#2819)

> Note that we don't ask for totalRoundTripTime on remote-inbound-rtp as MTI either.

Yes we seem to have discrepancies in both approach (preference/redundancy), naming, and MTI for RTT:

| stats type | naming | divisor naming | [MTI](https://w3c.github.io/webrtc-pc/#mandatory-to-implement-stats) |
|--|--|--|--|
| "remote-inbound-rtp" | [roundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteinboundrtpstreamstats-roundtriptime) | | Yes |
|  | [totalRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteinboundrtpstreamstats-totalroundtriptime) | [roundTripTimeMeasurements](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteinboundrtpstreamstats-roundtriptimemeasurements) | No |
| "remote-outbound-rtp" | [roundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteoutboundrtpstreamstats-roundtriptime) | | No |
|  | [totalRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteoutboundrtpstreamstats-totalroundtriptime) | [roundTripTimeMeasurements](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteoutboundrtpstreamstats-roundtriptimemeasurements) | No |
| "candidate-pair" | [currentRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcicecandidatepairstats-currentroundtriptime) | | Yes |
| | [totalRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcicecandidatepairstats-totalroundtriptime) | [responsesReceived](https://www.w3.org/TR/webrtc-stats/#dom-rtcicecandidatepairstats-responsesreceived) | Yes / No |
| "sctp-transport" | [smoothedRoundTripTime](https://w3c.github.io/webrtc-provisional-stats/#dom-rtcsctptransportstats-smoothedroundtriptime) | | No |

Is it too late to make naming consistent (x vs currentX vs smoothedX, and XMeasurements vs responsesReceived)?
Is it too late to make MTI consistent (x vs totalX/divisor vs both as MTI)?

That being said, [responsesReceived](https://www.w3.org/TR/webrtc-stats/#dom-rtcicecandidatepairstats-responsesreceived) not being MTI seems like a bug, so a PR to match it to [totalRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcicecandidatepairstats-totalroundtriptime) seems editorial, which would close out this issue. OTOH, looking at the table above, mandating totalX/divisor looks inconsistent.

@docfaraday correct me if I'm wrong, but no answer to this issue would help Firefox though, unless we also remove MTI from "candidate-pair"'s [currentRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcicecandidatepairstats-currentroundtriptime), isn't that so?

> All that said - the users have already had to deal with the lack of these numbers because of Mozilla not implementing them, so I guess they can live without them.

Yes, I think Mozilla's preference would be to remove [currentRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcicecandidatepairstats-currentroundtriptime), [totalRoundTripTime](https://www.w3.org/TR/webrtc-stats/#dom-rtcremoteoutboundrtpstreamstats-totalroundtriptime) and [smoothedRoundTripTime](https://w3c.github.io/webrtc-provisional-stats/#dom-rtcsctptransportstats-smoothedroundtriptime) from MTI for this reason. But I'm not sure we have consensus to do so.

-- 
GitHub Notification of comment by jan-ivar
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/2819#issuecomment-1439150054 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Tuesday, 21 February 2023 22:05:02 UTC