Re: [MMUSIC] WG adoption of other WebRTC data channel usage drafts?; Re: New Version Notification for draft-ietf-mmusic-msrp-usage-data-channel-02.txt

There are several things one may want to do.

1)	Register the protocol name. We had a previous discussion on this where I think we agreed that we would use the same registry as the websocket. That registry is first come first served, and as such, formally does not even need a reference specification, only a contact. By the reuse, we are also not defining whether the appropriate usage is for websocket, datachannel, or both - one would need to communication with the contact or use any referenced specification to identify that.

2)	Give guidance on how this data channel is used to support this protocol. This does not have to be an RFC, but we have plenty of cases in the IETF where we have a non-IETF protocol with the supporting infrastructure defined by IETF. Take numerous payload documents for example. Part of that would also state that a data channel usage is appropriate, see 1) above.

3)	Give guidance or even normative statements on how various SDP attributes are placed within the dcsa attribute, or even whether they are required or not.

Whether these are done in a RFC, or in a specification from another group, or even proprietary documentation, depends on what interest there is in creating documentation and where. Essentially the basic IETF rule should apply; if there is a group of people interested in working on a specification to become an RFC, and it is in the general spirit of the IETF to do so, then work should proceed.

Regards

Keith

-----Original Message-----
From: mmusic [mailto:mmusic-bounces@ietf.org] On Behalf Of Paul Kyzivat
Sent: 18 October 2015 17:55
To: Christer Holmberg; mmusic@ietf.org
Subject: Re: [MMUSIC] WG adoption of other WebRTC data channel usage drafts?; Re: New Version Notification for draft-ietf-mmusic-msrp-usage-data-channel-02.txt

On 10/18/15 8:44 AM, Christer Holmberg wrote:
> Hi,
>
> ...
>
>> So I think what is needed is to register a protocol name for T.140 
>> (or for RTT). While no spec is required to register the name, one is 
>> needed to specify the semantics of the messages sent in that protocol. If it follows T.140 then it might be a very simple document.
>
> For T.140 a specification makes sense, in order to ensure interoperability.
>
> But, what if company wants to use the data channel for a protocol related to a game? In order to avoid name collisions it would be good for them to register the name, but I don't think they would want to reveal the details of the game protocol.

Well, the registry exists, and is FCFS. Each registry entry contains a reference to a document. So there needs to be *something* to reference. 
I haven't investigated if there is any requirement on what it contains. 
And I don't think there are any police that check up on whether the referenced document continues to exist.

So, if you want to register a secret protocol you can probably do so.

	Thanks,
	Paul

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