Re: [tsvwg] Alissa Cooper's Yes on draft-ietf-tsvwg-diffserv-intercon-12: (with COMMENT)

[Alissa Cooper <[e-mail]>]

Hi David,

> On Dec 1, 2016, at 10:32 AM, Black, David <[e-mail]> wrote:
> 
> Alissa,
> 
> Thanks for the review and comments:
> 
>> One thing I didn't get is why not have a fifth aggregate that CS1 could
>> be mapped into. Is it just because in other standards that have specified
>> aggregates like this, they've gone with four and not specified one for
>> less-than-best-effort?
> 
> Available MPLS label space considerations (i.e., lack of space) are a major
> reason to limit the number of aggregates to four, and RFC 5127 did
> likewise for similar reasons.
> 
> Beyond that, there is no longer IETF consensus that CS1 should be used for
> the traffic class variously known as Lower Effort, less-than-best-effort or
> scavenger, see: https://datatracker.ietf.org/doc/draft-ietf-tsvwg-le-phb/ .

Got it, thanks for the explanation.

> 
>> I'm also wondering about the choice to reserve AF42 and AF43. For WebRTC
>> and real-time applications (see table at
>> https://tools.ietf.org/html/draft-ietf-tsvwg-rtcweb-qos-18#section-5)
>> traffic marked as "Medium" in the table would be treated the same as
>> "Low," which I fear would give incentives for applications to mark their
>> traffic as "High" instead. Unless there is a strong need to reserve AF42
>> and AF43, it might be nice to just include them.
> 
> Well, that rtcweb-qos draft uses 11 DSCPs, so there are a number of distinctions
> that have to be lost in order to get to 4 traffic aggregates.  In contrast to AF33,
> which is reserved by this draft, the text on AF42 and AF43 is weaker:
> 
>          The AF42 and AF43 PHBs could be added if there
>           is a need for three-color marked Multimedia.
> 
> Are you asking for these two to be added now because WebRTC is expected to
> make extensive use of two- and three-color marking (e.g., of video streams)?

Well, the expectation is that WebRTC implementations will make use of AF42 and AF43. I can’t predict how extensively.

Alissa

> 
> Thanks, --David
> 
>> -----Original Message-----
>> From: Alissa Cooper [mailto:[e-mail]]
>> Sent: Thursday, December 01, 2016 9:27 AM
>> To: The IESG
>> Cc: [e-mail]; Gorry Fairhurst; tsvwg-
>> [e-mail]; [e-mail]; [e-mail]
>> Subject: Alissa Cooper's Yes on draft-ietf-tsvwg-diffserv-intercon-12: (with
>> COMMENT)
>> 
>> Alissa Cooper has entered the following ballot position for
>> draft-ietf-tsvwg-diffserv-intercon-12: Yes
>> 
>> When responding, please keep the subject line intact and reply to all
>> email addresses included in the To and CC lines. (Feel free to cut this
>> introductory paragraph, however.)
>> 
>> 
>> Please refer to https://www.ietf.org/iesg/statement/discuss-criteria.html
>> for more information about IESG DISCUSS and COMMENT positions.
>> 
>> 
>> The document, along with other ballot positions, can be found here:
>> https://datatracker.ietf.org/doc/draft-ietf-tsvwg-diffserv-intercon/
>> 
>> 
>> 
>> ----------------------------------------------------------------------
>> COMMENT:
>> ----------------------------------------------------------------------
>> 
>> Thanks for writing this document.
>> 
>> One thing I didn't get is why not have a fifth aggregate that CS1 could
>> be mapped into. Is it just because in other standards that have specified
>> aggregates like this, they've gone with four and not specified one for
>> less-than-best-effort?
>> 
>> I'm also wondering about the choice to reserve AF42 and AF43. For WebRTC
>> and real-time applications (see table at
>> https://tools.ietf.org/html/draft-ietf-tsvwg-rtcweb-qos-18#section-5)
>> traffic marked as "Medium" in the table would be treated the same as
>> "Low," which I fear would give incentives for applications to mark their
>> traffic as "High" instead. Unless there is a strong need to reserve AF42
>> and AF43, it might be nice to just include them.
>> 
>