Re: [Gen-art] Gen-ART Last Call review of draft-ietf-clue-datachannel-13

[Christer Holmberg <[e-mail]>]

Hi Brian,

...

>>> Minor issues:
>>> -------------
>>>
>>> Mainly for my own education:
>>>
>>> 3.2.6.  SCTP Multihoming
>>>
>>>   SCTP multi-homing is not supported for SCTPoDTLS associations, and
>>>   can therefore not be used for a CLUE data channel.
>>>
>>> What is the advantage of SCTP if you don't get the benefit of multihoming?
>> 
>> There are other SCTP features that are used. The most essential is the SCTP multi stream feature, which 
>> allows multiple data channels using a single SCTP associations: each data channel is implemented using two 
>> unidirectional SCTP streams.
>> 
>> SCTP also provide different options when it comes to data transport reliability and ordering, and data channels can use different combinations.
>
> OK, thanks. I have the impression that this explanation is given nowhere in the CLUE documents 
> (and not in draft-ietf-rtcweb-transports either). I think it would be helpful if it was recorded *somewhere*. 
> It doesn't really belong in clue-datachannel.

I agree - it's not the task of CLUE to justify the decisions made by RTCWEB.

For the details of the data channel mechanism, please take a look at draft-ietf-rtcweb-data-channel.

https://tools.ietf.org/html/draft-ietf-rtcweb-data-channel

Regards,

Christer



> 
> 
>> Nits:
>> -----
>>
>> I expected a reference to draft-ietf-clue-protocol where CLUE is first mentioned in the Introduction.
> 
> I'll fix that.

Thanks
    Brian