Decisions reached at the Aug 23 virtual interim

**

*These are the decisions reached at the August 23 interim.*

*

These are posted on the mailing list for review; if there are no
comments raising further issues after a week, the decisions will be
taken as final.


685/759: Receipt of multiple RTP encodings

We will not mandate implementing either reception or transmission of
simulcast.

We will refer to JSEP language once it’s merged.

We will not add any API for controlling whether an UA offers to send
simulcast; we will not say that it’s forbidden to include that in the
offer, we’ll just say that there’s no control surface.


714/740: STUN/TURN OAuth Token Parameter

The meeting concluded that we believe the UA only needs the access-token
and the key ID; these fit within the existing fields fo the parameter
structure.

We need text to say “here’s where OAuth parameters go”.

NOTE: After the meeting, doubt has been raised about whether or not the
analysis was complete, so this decision is already “not final”.


720/738: Fingerprint of an RTCCertificatei

After discussion, the conclusion was that a certificate only has one
fingerprint.

Hash agility is achieved by generating multiple certificates, which can
be based on the same key material if desired.

Thus, no change is needed.

NOTE: After the meeting, people have pointed to
draft-ietf-rfc4572-update as challenging this assumption. This needs to
be resolved on the IETF side. So issue remains open.


726/757: ufrag attribute to RTCIceCandidate

Discussion concluded that we have two separate problems: Indicating
ufrag in an ICE candidate and indicating the end of candidates for a
given transport.

Decision: We will add ufrag to RTCIceCandidate, and raise a new issue
for the end-of-candidates problem.


732/758: replaceTrack on an ended track?

Decision: The text in PR 758 is acceptable.


678: Support recipient identity:

There was no change on this from the last interim. Martin has the token
to suggest text.


692: Meaning of “Liveness checks”

Decision: The text on the slide, which says that the “disconnected”
state is entered whenever the UA thinks the calll’s disconnected, is OK
and will be adopted.

Taylor will work on a PR for it.


700: Event for when circuit breaker is triggered

There was no consensus after discussion. No decision was reached.


729: RTCStats timestamp source

The proposed solution on the slide (stats are timestamped on arrival at
UA, remote NTP time is stored in an additional field if known) is
acceptable.

Harald and Varun will work on text.*


-- 
Surveillance is pervasive. Go Dark.