Re: [clue] Ben Campbell's Discuss on draft-ietf-clue-rtp-mapping-12: (with DISCUSS and COMMENT)

["Ben Campbell" <[e-mail]>]

Hi Roni, thanks for the response. See comments below. I removed sections 
that I think  are resolved.

On 19 Jan 2017, at 0:51, Roni Even wrote:

> Hi Ben,
> See inline
> Roni
>
>> ----------------------------------------------------------------------
>> DISCUSS:
>> ----------------------------------------------------------------------
>>
>> I plan to ballot "yes" for this, but there's an issue in section 9 
>> that I think
>> needs to be fixed first:
>>
>> In the second paragraph, the draft says "CLUE endpoints MUST support 
>> RTP/
>> SAVPF and DTLS-SRTP keying [RFC5764]." But the framework draft goes
>> further by saying that media MUST be secured, and that DTLS-SRTP 
>> SHOULD
>> be used unless the media is secured by some other mechanism. I think 
>> that
>> readers will expect the mapping spec to be authoritative about that 
>> sort of
>> thing. It's likely to be misleading to have it mention the 
>> requirement to
>> support RTP/SAVPF and DTLS-SRTP without also mentioning the MUST be
>> secured, SHOULD be used requirements. This can easily be fixed by
>> mentioning the additional requirements and citing the framework.
> [Roni Even] You are right, I will use the text from the framework 
> which is the correct text
>>

Thanks. I will clear the discuss now, since this is a simple enough fix.

>>
>> ----------------------------------------------------------------------
>> COMMENT:
>> ----------------------------------------------------------------------
>>
>> Substantive:
>>
>> -3: The opening paragraph mentions "Point-to-Point, as well as 
>> Media-Mixing
>>    mixers, Media- Switching mixers, and Selective Forwarding 
>> Middleboxs."
>> The section goes on to discuss the first two, but doesn't mention the 
>> last
>> two.
> [Roni Even] There was some text in previous revision but the feedback 
> from the WG was to remove it since it was just repeating text from the 
> topology draft (RFC7667)

Okay. Something to the effect that "Media-switching mixers and Selective 
Forwarding Middleboxes behave as described in [RFC7677]" towards the end 
of the section would help balance things--but it's not critical one way 
or the other.

[...]

>>
>> -5, paragraph 5:
>> It would be helpful to add a sentence or two clarifying the 
>> difference
>> between an MCC that carries multiple MCs at the same time, and one 
>> that
>> switch between MCs, but only carry one at a time.
> [Roni Even] I think that it is clear, it is not carrying multiple MCs 
> it "sends a composed stream of multiple MCs". This document does not 
> define what is an MCC so if you do not know what is an MCC you need to 
> read the framework document which is a normative reference.

Your call on this one. I am not sure that people will realize that 
"composed stream of multiple MCs" means one at a time.

[...]

>>
>> -9, 2nd paragraph: Please don't use 2119 to describe requirements 
>> from
>> other documents, unless in a direct quote (in quotation marks.)
> [Roni Even] I am not sure what to do, I assume it is about "CLUE 
> endpoints MUST support RTP/
>    SAVPF and DTLS-SRTP keying [RFC5764].". We want to say that they 
> MUST support SAVPF and DTLS-SRTP keying . but I think that this text 
> will change based on Magnus request to provide specific security 
> profile

You can always say that "[RFCXXXX] requires endpoints to..."