Re: [tsvwg] Your NQB presentation

["Black, David" <[e-mail]>]

> The proposal (David, feel free to jump in if I've misstated anything) is to select
> the 0x2A code point, and add two requirements to address the WiFi safety
> concern:
> 
> An ISP node that supports the NQB PHB and serves residential users MUST have
> the ability to re-mark NQB traffic as default, and MUST do so by default.
>
> (precise wording to be smithed)
> 
> I think that this achieves the same end result as selecting a low-valued code
> point - it ensures that operators will need to actively decide to pass the 0x2A
> codepoint through to their residential customers.

That’s a reasonable summary, and indeed, the word "residential" will almost certainly need some wordsmithing attention.

The basic idea is that a default configuration that bleaches NQB traffic places the responsibility for the effects of not bleaching on whomever changed that default configuration - we can't prevent people who are determined to shoot themselves in the foot from accomplishing that goal ;-).   With careful wordsmithing, this will also cover end-user-purchased/provided ISP interface equipment because it'll come with that same default configuration if NQB is supported.

Thanks, --David

> -----Original Message-----
> From: tsvwg <[e-mail]> On Behalf Of Greg White
> Sent: Thursday, November 21, 2019 8:30 PM
> To: Sebastian Moeller; tsvwg IETF list
> Subject: Re: [tsvwg] Your NQB presentation
> 
> 
> [EXTERNAL EMAIL]
> 
> Sebastian, all,
> 
> I had a discussion with David Black on the subject of DSCP and he proposed a
> resolution that perhaps addresses the main concerns that have been expressed.
> 
> The proposal (David, feel free to jump in if I've misstated anything) is to select
> the 0x2A code point, and add two requirements to address the WiFi safety
> concern:
> 
> An ISP node that supports the NQB PHB and serves residential users MUST have
> the ability to re-mark NQB traffic as default, and MUST do so by default.
> 
> (precise wording to be smithed)
> 
> I think that this achieves the same end result as selecting a low-valued code
> point - it ensures that operators will need to actively decide to pass the 0x2A
> codepoint through to their residential customers.
> 
> In addition to these two MUSTs, we need some clear and concise informative
> text that discusses the WMM safety concern.
> 
> I think we share the belief that the vast majority (all?) residential ISPs currently
> bleach ALL DSCPs to default for their residential customers, and so the 0x2A
> value will already be handled in the same manner, but these requirements and
> the (to be written) informative text will help ensure operators only deviate from
> that default state consciously.
> 
> -Greg
> 
> 
> On 11/22/19, 5:13 AM, "Sebastian Moeller" <[e-mail]> wrote:
> 
>     Dear Greg, dear list,
> 
> 
>     I just had the opportunity to watch the video of your presentation. Just a quick
> note that I very much appreciated the balanced, concise and unbiased
> presentation of the PHB and the open issues.
> 
>     Regarding code points, I had proposed 0x2E since it is the "sum" of 0x6 and
> CS5, so I saw a cute "symmetry" there, but since it is is taken already (and I
> agree the goal should be for NQB to be ideally unambiguous end to end) I am
> fine with 0x2A for the case when AC_VI is known to be safe to use.
> 
> 
>     Best Regards
>     	Sebastian