Re: [MMUSIC] Data channel dcsa attribute usage and IANA

[Paul Kyzivat <[e-mail]>]

On 8/28/19 10:45 AM, Christer Holmberg wrote:
> Hi,
> 
>>> Section 7.2 of draft-ietf-mmusic-msrp-usage-data-channel updates the
>>> IANA registry for the SDP ‘setup’ attribute, by indicating that it can
>>> be used inside an SDP ‘dcsa’ attribute.
>>>
>>> Do we really need to update the IANA registry for all SDP attributes
>>> that can be included in an ‘dcsa’ attribute???
>>
>> I think so. Look at RFC4566bis for the revised attribute registry.
>>
>> The way to avoid this would be to remove "Usage Level" from the registry and thus force
>> people to consult the defining documents for the attribute in order to learn this. But it is
>> seriously late in the cycle to update 4566bis.
> 
> I think it can be seen in two different ways, that have no impact on 4566bis.
> 
> 1) If the usage level is "media", that doesn't change inside a dcsa attribute. From an SDP perspective it is still "media", as it is part of an m= line. If you place a "session" level attribute inside a dcsa attribute that changes, but I don't think we do that; or
> 
> 2) From an SDP perspective, only the dcsa attribute is used. What's inside is "application specific", and has nothing to do with SDP.

I think you are mixing two things here:

- what currently is required based on the specs as written (including 
4566bis);

- what would be the ideal way for it to work.

Based on the current specs I think the registry must be updated. And I 
think that was always expected during the development of the dcsa attribute.

What is most important is that people have a well defined way to 
determine whether an attribute can be used in dcsa in a particular case. 
If they are expected to find this in a document defining the attribute, 
then the document had better provide an answer.

If you feel a change is needed to 4566bis then we will have to figure 
out if and how to accomplish that. It is already at the editor.

	Thanks,
	Paul