Re: [MMUSIC] draft-uberti-mmusic-nombis and (D)TLS

Hi,

Even without forking, all candidate pairs associated with the virtual connection have to be handled at the same DTLS stack/physical device.

For example, two m- lines (non-bundled) could be associated with separate physical devices, so you would need separate DTLS connections.

Regards,

Christer

From: Jonathan Lennox [mailto:jonathan@vidyo.com]
Sent: 23 March 2015 06:08
To: Justin Uberti
Cc: Roman Shpount; Christer Holmberg; Ari Keränen; draft-uberti-mmusic-nombis@tools.ietf.org; mmusic
Subject: Re: [MMUSIC] draft-uberti-mmusic-nombis and (D)TLS

The virtual connection is, I think, a single component of what RFC 5245 and draft-ietf-mmusic-rfc5245bis-04 call a "media stream".  (This will be clearer, I think, if you think about non-muxed RTP/RTCP.)

Unfortunately, the "ICE media stream" terminology in 5245 conflicts rather badly with the RTP Taxonomy, so we may want to come up with something better.

In SDP, the virtual connection is identified by a BUNDLE group if you're using BUNDLE, or an m-line if not; and in either case sharing the same component ID, and considering both sides of the offer/answer exchange (to handle forking cases).

On Mar 22, 2015, at 10:56 PM, Justin Uberti <juberti@google.com<mailto:juberti@google.com>> wrote:

Alas, an ICE ufrag doesn't uniquely identify an ICE virtual connection; with ICE forking you can have N virtual connections from the same ufrag.

A lfrag:rfrag tuple is closer, but even this does not work, because a) an ICE restart changes lfrag/rfrag without invalidating the connection, and b) because a ufrag can be shared across multiple m= lines (and thereby ICE connections).

The closest thing is m-line, keeping in mind that bundled m-lines use the virtual connection of the m-line onto which they are bundled.

On Sun, Mar 22, 2015 at 6:10 PM, Roman Shpount <roman@telurix.com<mailto:roman@telurix.com>> wrote:
On Sun, Mar 22, 2015 at 8:08 PM, Christer Holmberg <christer.holmberg@ericsson.com<mailto:christer.holmberg@ericsson.com>> wrote:

What was discussed in RTCWEB was to, instead of binding a DTLS connection to a 5-tuple, bind it to a "virtual connection". A "virtual connection" would be the set of all candidate pairs associated with a.... something.

Whether "something" is an m- line, a BUNDLE group, a complete SDP, or something else, hasn't been discussed - as far as I remember (please correct me if I'm wrong).

I was proposing that "something" is ICE ufrag, i.e. all 5-tuples for which a ICE/STUN bind request with the particular ufrag is received gets associated with the same "virtual" connection.
_____________
Roman Shpount

_______________________________________________
mmusic mailing list
mmusic@ietf.org<mailto:mmusic@ietf.org>
https://www.ietf.org/mailman/listinfo/mmusic