Re: [dispatch] Questions regarding draft-petithuguenin-dispatch-rtp-pmtud-00

[Marc Petit-Huguenin <[e-mail]>]

On 03/21/2017 03:29 AM, Marc Petit-Huguenin wrote:
> On 03/20/2017 08:20 PM, Martin Thomson wrote:
>> This draft would appear to be firmly in the MMUSIC working group
>> remit.  I note that the ICE working group defines mechanisms for ICE,
>> and MMUSIC defines SDP.  Since this really only uses a mechanism that
>> is defined in TRAM, the only new protocol pieces here are some SDP,
>> which MMUSIC is good at.
> 
> Yes.

I withdraw that "yes".  It is also about finding the right bytes in all the packets exchanged (RTP, RTCP, STUN, even DTLS?) to be able to match the list of packets sent with the list of identifiers that will be sent back in the Report response.

> 
>>
>> I see that this is a new -00 draft, so maybe it's just too new to have
>> been discussed.  Was it rejected from those groups?
> 
> No, this draft comes from some bits that we removed from draft-ietf-tram-stun-pmtud when we found out that the usage for RTP (or ICE or whatever) was too under-defined to be useful.  I also see that document as a blueprint on how to specify a PMTUD extension to an existing protocol.  This is the first appearance of the idea of doing PMTUD for RTP sessions in this form.
> 
>>
>> On the details, ice-options is a better place to signal this.  Drop
>> the abominable x- prefix at the same time.
>>
> 
> I am not sure about the ice-options (I agree with the x- prefix), as an ice-options would restricts the signaling to ICE.  It's probably safe to say that WebRTC will be the main user of that draft, but may be SIP would like to use that too without ICE.  Hence Dispatch.
> 


-- 
Marc Petit-Huguenin
Email: [e-mail]
Blog: https://marc.petit-huguenin.org
Profile: https://www.linkedin.com/in/petithug