Re: [MMUSIC] T140: Usage of "-"fmt value in dcsa encapsulated fmtp attribute

[Gunnar Hellström <[e-mail]>]

Den 2019-12-13 kl. 21:05, skrev Flemming Andreasen:


On 12/13/19 6:50 AM, Christer Holmberg wrote:

Hi,

It seems to me that the solution for the "fmt" value definition might be found in sdpneg



In sdpneg section 5.2.1
https://tools.ietf.org/html/draft-ietf-mmusic-data-channel-sdpneg-28#section-5.2.1
It is said:
"There may be cases, where the usage of a subprotocol related media
  level attribute depends on the subprotocol's transport protocol.  In
  such cases the subprotocol related usage of the attribute is expected
  to be described for the data channel transport.  A data channel
  specific usage of a subprotocol attribute is expected to be specified
  in the same document that registers the subprotocol's identifier for
  data channel usage as described in https://tools.ietf.org/html/draft-ietf-mmusic-data-channel-sdpneg-28#section-9.1. "

In t140-usage, section4.2.1,
https://tools.ietf.org/html/draft-ietf-mmusic-t140-usage-data-channel-10#section-4.2.1
we have this sentence:
  'If the 'fmtp' attribute is included, the 'format' attribute parameter MUST be set to "-". '


The text above is what triggered the discussion :)



It seems to me that the t140-usage draft by that fulfills the requirements of sdpneg and we are fine with fmt=-
Maybe we could even say that for the t.140-data-channel, "fmt" MUST be ommitted, in order to get a simpler syntax (?).

However,

This issue may also relate to the definition of "fmt" in rfc4566bis
 https://tools.ietf.org/html/draft-ietf-mmusic-rfc4566bis-37#section-8.2.3
where we have the following sentences that seem to apply:
"For other protocols, formats MAY be registered according to the rules
  of the associated "proto" specification.

  Registrations of new formats MUST specify which transport protocols
  they apply to."

I have not traced if we are bound to that requirement for dcsa definitions, and how it would map from protocol to sub-protocol, so I hope we can obey sdpneg.

Summary: I would suggest to omit "fmt", or keep fmt=- and claim that the use of fmt is defined in t140-usage as required by sdpneg.


We DO define the usage as required by sdpneg,  so that is not the issue.

The question is whether it in such definition is allowed to "override" the 4566bis statement saying that the format must be one of the formats specified for the media. As far as SDP is concerned, the only format is 'webrtc-datachannel'.

Again, I personally think we should keep "-" (I don't like to omit information elements), and add whatever text needed to indicate that the 4566bis text does not apply to sub-protocol specific fmtp attributes.


Syntactical consistency and strict adherence to the wording in 4566bis would argue for "webrtc-datachannel", whereas conceptual consistency would argue more for "t140". I'm not in favor of "-", since it doesn't seem to add anything.
I tend to lean towards "t140" now, even if it does not add any information.
We could, if needed, draw a map of how different statements in rfc4566bis, sdpneg and t140-usage relate to this topic, and probably find that there is a bit of inconsistency.
But,
RFC4566bis, section 8.2.3
https://tools.ietf.org/html/draft-ietf-mmusic-rfc4566bis-32#section-8.2.3
says for protocols other than rtp and udp

"   For other protocols, formats MAY be registered according to the rules
   of the associated "proto" specification.

   Registrations of new formats MUST specify which transport protocols
   they apply to."

sdpneg intends to map dcmap and dcsa into a kind of second level media section,


and in turn hands over to the specific subprotocol specifications to tell how attributes are defined and used.



In t140-usage, section 9.2, the use of the fmtp attribute is registered and a reference to section 4.2.1 made. And in 4.2.1, the rule for fmt is defined.
https://tools.ietf.org/html/draft-ietf-mmusic-t140-usage-data-channel-10#section-9.2


A question is if the fmt parameter value ( "-" or "t140" ) would need to be registered since RFC4566bis says that new formats MAY be registered, or if that requirement is satisfied by the description of the use of "-" in the fmtp attribute in t140-usage section 4.2.1.

If we think that this is too complex, we could get around the problem by instead defining a new value "cps", specified in a dcsa attribute.

/Gunnar

Thanks

-- Flemming



Regards,

Christer







Den 2019-12-12 kl. 20:12, skrev Christer Holmberg:
Hi,

When the chairs reviewed the T.140 data channel usage draft, the following issue was raised:

Currently, when we send an fmtp attribute encapsulated in a dcsa attribute, we use "-" as the fmt value.

     Example: a=dcsa:2 fmtp:- cps=20

However, 4566bis says the following about the fmtp attribute:

"The format must be one of the formats specified for the media."

Now, the format is 'webrtc-datachannel', so one could claim that should be used.

     Example: a=dcsa:2 fmtp:webrtc-datachannel cps=20


 But, that applies to the whole SCTP association, and the dcsa attribute is associated with a specific data channel usage. In addition, the usage and syntax of the fmtp attribute may vary depending on data channel usage.

Another option could be to include the sub-protocol:

     Example: a=dcsa:2 fmtp:t140 cps=20


But, the dcsa attribute already maps the fmtp attribute to the sub-protocol.


So, what to do?


My personal suggestion would be to keep "-", and perhaps add a note somewhere. Not sure we would need to update 4566bis, but if people want to do that we could for sure do it.

Anyone having a different opinion?


Regards,


Christer




_______________________________________________
mmusic mailing list
mailto:[e-mail]
https://www.ietf.org/mailman/listinfo/mmusic



--

+ + + + + + + + + + + + + +

Gunnar Hellström
Omnitor
[e-mail]<mailto:[e-mail]>
[phone]