Re: [rtcweb] Working Group Last Call for draft-uberti-rtcweb-rfc8829bis-01.txt

[Roman Shpount <[e-mail]>]

I am available during US Eastern hours.

I will try to summarize this discussion:

With the latest bundle update, subsequent offers with the bundle (offers
where bundled m= lines which can no longer be unbundled are not marked in
any way except that they share the same communication address) are not
valid initial offers for the bundle (where bundled m= lines which can no
longer be unbundled have the port set to zero and bundle-only attribute).
It was not the case with RFC8843 (all offers and answers had bundle-only m=
line attributes). If, as a result of 3PCC, a subsequent offer becomes an
initial offer, it either needs to be fixed by something (originating
endpoint or 3PCC controller), or RFC8843 should be modified to make
subsequent offers valid initial offers (detect that bundled m= lines share
the same address and cannot be unbundled). The solution that Christer and I
are proposing is that the 3PCC controller should fix the offer if it were
the subsequent offer for one session and was used as the initial offer for
another. Christer came up with the text for RFC8843bis to specify this.
_____________
Roman Shpount


On Thu, Nov 4, 2021 at 11:33 PM Cullen Jennings <[e-mail]> wrote:

>
> Could we have a quick call on this next week during one of the breaks. I
> have tried to follow this whole thread and some it does not make much sense
> to me. I’m a bit lost on what the varios proposed resolutions are.
>
>
> On Nov 3, 2021, at 12:38 PM, Christer Holmberg <
> christer.holmberg=[e-mail]> wrote:
>
> Justin, are you ok with the suggested text? I would probably be a good
> idea to refer it in 8829bis.
>
> Regards,
>
> Christer
> ------------------------------
> *From:* rtcweb <[e-mail]> on behalf of Christer Holmberg <
> christer.holmberg=[e-mail]>
> *Sent:* Tuesday, November 2, 2021 11:13 PM
> *To:* Roman Shpount <[e-mail]>
> *Cc:* Justin Uberti <[e-mail]>; RTCWeb IETF <[e-mail]>
> *Subject:* Re: [rtcweb] Working Group Last Call for
> draft-uberti-rtcweb-rfc8829bis-01.txt
>
> Hi,
>
>
> >This makes it much better. I think it is missing a couple of commas ("In
> some 3PCC scenarios," and "In the rewritten offer,") but works for me.
>
>
> I can add them when I add the text to the draft, but I think the RFC
> editor will most likely notice such thing.
>
>
> >I have changed the section name so it is clear that it applies to JSEP as
> well, not just SIP.
>
>
> Sure, that’s fine.
>
>
> Regards,
>
>
> Christer
>
>
>
>
>
>
> On Tue, Nov 2, 2021 at 1:24 PM Christer Holmberg <
> [e-mail]> wrote:
>
> Hi,
>
>
> Eventhough I would not like to make more changes than necessary, I am fine
> with "3PCC Considerations".
>
>
> However, your suggested text is very difficult to understand in some
> places, so let me give it a try.
>
>
> (The first paragraph is generic, the second SIP specific, and the third
> BUNDLE specific.)
>
>
> ---
>
>
> 3PCC Considerations
>
>
> In some 3PCC scenarios a new session will be established between an
> endpoint that is currently part of an ongoing session and an endpoint that
> is currently not part of an ongoing session. The endpoint that is part of a
> session will generate a subsequent offer that will be forwarded to the
> other endpoint by a 3PCC controller. The endpoint that is not part of a
> session will process the offer as an initial offer.
>
>
> The Session Initiation Protocol (SIP) [RFC3261] allows a User Agent Client
> (UAC) to send a re-INVITE request without an SDP body (sometimes referred
> to as an empty re-INVITE). In such cases, the User Agent Server (UAS) will
> include an SDP offer in the associated 200 OK response. If the UAS is a
> part of an ongoing session, it will include a subsequent offer in the 200
> OK response. The offer will be received by a 3PCC controller (UAC) and then
> forwarded to another User Agent (UA). If the UA is not part of an ongoing
> session, it will process the offer as an initial offer.
>
> When the BUNDLE mechanism is used, an initial BUNDLE offer is constructed
> using different rules than subsequent BUNDLE offers, and it cannot be
> assumed that a UA is able to correctly process a subsequent offer as an
> initial offer. Therefore, the 3PCC controller SHOULD rewrite the subsequent
> offer into a valid initial offer, following the procedures in (Section
> 7.2), before it forwards the offer to a UA. In the rewritten offer the 3PCC
> controller will set the port value to zero (and include an SDP
> 'bundle-only' attribute) for each "m=" section within the BUNDLE group,
> excluding the offerer-tagged "m=" section.
>
>
>
>
>
>
>
>
>
>
>
>
> ------------------------------
> *From:* Roman Shpount <[e-mail]>
> *Sent:* Tuesday, November 2, 2021 6:33 PM
> *To:* Christer Holmberg <[e-mail]>
> *Cc:* Justin Uberti <[e-mail]>; Justin Uberti <
> [e-mail]>; RTCWeb IETF <[e-mail]>
> *Subject:* Re: [rtcweb] Working Group Last Call for
> draft-uberti-rtcweb-rfc8829bis-01.txt
>
>
> How about we replace the SIP Considerations with:
>
>
> 3PCC Considerations
>
>
> In some 3PCC scenarios, an offer generated during an ongoing session,
> i.e., a subsequent offer, will be used by a 3PCC controller to establish a
> new session and forwarded as an initial offer to another endpoint that is
> currently not part of a session.
>
> For example, the Session Initiation Protocol (SIP) [RFC3261] allows a User
> Agent Client (UAC) to send a re-INVITE request without an SDP body
> (sometimes referred to as an empty re-INVITE). In such cases, the User
> Agent Server (UAS) will include an SDP offer in the associated 200 OK
> response. If UAS is a part of an ongoing session, it will include a
> subsequent offer in the 200 OK response. The offer will be received by a
> 3PCC controller (UAC) and then forwarded to another User Agent (UA) as an
> initial offer.
>
> When the BUNDLE mechanism is used, an initial BUNDLE offer is constructed
> using different rules than subsequent BUNDLE offers. It cannot be assumed
> that a subsequent offer is a valid initial offer and that the endpoint that
> expects an initial offer will properly process such a subsequent offer.
> Therefore, the 3PCC controller SHOULD rewrite the subsequent offer into a
> valid initial offer before it is used to establish a new session. To make
> the subsequent offer a valid initial offer, 3PCC will need to modify all
> the non-tagged m= lines to include the bundle-only attribute and set the m=
> line port to zero.
> _____________
> Roman Shpount
>
>
>
>
> On Tue, Nov 2, 2021 at 6:00 AM Christer Holmberg <
> [e-mail]> wrote:
>
> Hi,
>
>
> What about something like this:
>
>
> ---
>
>
> OLD:
>
>
> “The Session Initiation Protocol (SIP) [RFC3261] allows a User Agent
> Client (UAC) to send a re-INVITE request without an SDP body (sometimes
> referred to as an empty re-INVITE).
> In such cases, the User Agent Server (UAS) will include an SDP Offer in
> the associated 200 OK response. This is typically used for 3rd Party Call
> Control (3PCC) scenarios.
> From a BUNDLE perspective, such SDP Offer SHOULD be generated using the
> procedures defined in Section 7.2.”
>
>
> NEW:
>
>
> “The Session Initiation Protocol (SIP) [RFC3261] allows a User Agent
> Client (UAC) to send a re-INVITE request without an SDP body (sometimes
> referred to as an empty re-INVITE).
> In such cases, the User Agent Server (UAS) will include an SDP offer in
> the associated 200 OK response. This is typically used for 3rd Party Call
> Control (3PCC) scenarios.
>
>
> In some 3PCC scenarios the UAS will be part of an ongoing session, and
> will therefore include a subsequent offer in the 200 OK responses. The
> offer will be
> received by a 3PCC controller (UAC) and then forwarded as an initial offer
> to another User Agent (UA) that is currently not part of a session.
>
>
> When the BUNDLE mechanism is used, as an initial BUNDLE offer look
> different than a subsequent BUNDLE offer, it cannot be assumed that a UA
> that expects an initial offer
> will be able to properly process a subsequent offer. Therefore, the 3PCC
> controller needs to act as a Back-To-Back User Agent (B2BUA), and when it
> receives the subsequent
> offer it needs to rewrite it into an initial offer before it is forwarded
> to such UA.”
>
>
> ----
>
>
> Regards,
>
>
> Christer
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> *From:* Roman Shpount <[e-mail]>
> *Sent:* tiistai 2. marraskuuta 2021 10.41
> *To:* Justin Uberti <[e-mail]>
> *Cc:* Christer Holmberg <[e-mail]>; Justin Uberti <
> [e-mail]>; RTCWeb IETF <[e-mail]>
> *Subject:* Re: [rtcweb] Working Group Last Call for
> draft-uberti-rtcweb-rfc8829bis-01.txt
>
>
> On Mon, Nov 1, 2021 at 2:52 PM Justin Uberti <
> [e-mail]> wrote:
>
> The PROBLEM is that we have two endpoints, where one sends a subsequent
> offer, and the other one expects an initial offer. What do you normally do
> when you have that kind of problem? You use an SBC/B2BUA. In this case that
> SBC/B2BUA would be the 3PCC controller.
>
>
> So, my suggestion would be to remove the SHOULD text from 8843bis, and
> simply add a note somewhere (in 8843bis and/or 8829bis) which describes the
> issue and says that the 3GPP controller needs to modify the offer
> accordingly.
>
>
>
> Roman, thoughts on this? If the 3PCC is going to rewrite the offer SDP
> anyway then maybe adding a=bundle-only isn't the end of the world.
>
>
> I am not opposed to this idea. 3PCC typically knows that the subsequent
> offer is going to be used as initial, and should be able to rewrite the
> offer to make it valid. We can change SIP Considerations section in 8843bis
> (
> https://www.ietf.org/archive/id/draft-ietf-mmusic-rfc8843bis-05.html#name-sip-consideration),
> remove the SHOULD, and specify that 3PCC controller should fix the offer.
> We can then reference this note from 8829bis or restate the same guidance.
> _____________
> Roman Shpount
>
>
>
> _______________________________________________
> rtcweb mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtcweb
>
>
>