[tsvwg] Re: WGLC for A NQB PHB for Differentiated Services (draft-ietf-tsvwg-nqb) to end 27th May 2024.

[Sebastian Moeller <[e-mail]>]

Hi Jason,


> On 21. May 2024, at 15:53, Livingood, Jason <jason_livingood=[e-mail]> wrote:
> 
> On 5/21/24, 02:30, "Sebastian Moeller" wrote:
> 
>> [SM] Respectfully, how did you assess the potential harm done by using AC_VI to other traffic (e.g. other APs/SSIDs sharing the same channel)? If you relied on traffic shaping ingress NQB traffic at the CMTS/CPE level how did you assert that the NQB traffic did not exceed its acceptable share of the (variable) capacity/airtime of the WiFi link?
> 
> Comcast operates one of the world's largest Wi-Fi networks and we have extensive Wi-Fi labs and field testing programs. Separate from our low latency program, we have for some time tested the effects of user-directed per-device and app-based per-flow WMM priority. We certainly do focus on adjacent AP, as this is especially critical in the many multi-dwelling units (apartment buildings and the like) in which we operate. We have not observed any performance problems in such testing.

[SM] So you can remotely monitor the fraction of airtime in a channel used up by the different UP/ACs for the own SSID as well as SSID/Ps in the neighbourhood? Impressive!

> 
> In addition, many commercially available consumer routers and AP systems already offer similar prioritization for many years and I have not heard complaints in the marketplace about this (e.g., 'my neighbor is using a gaming AP and now my WiFi network doesn't work').  

[SM] That sounds like  the "but dad, the other kids do it too" defence members of my family sometimes try to employ... Not an acceptable excuse IMHO, e.g. when D-Link routers abused/overload public NTP servers, that was something worth fixing and not something for everybody else to emulate.

Actually diagnosing that "bad WiFi" is caused by some neighbour/station hogging inappropriate amounts of airtime/transmit opportunities due to using AC_VO or AC_VO is hard, let alone for normal users. At least that is my experience as a normal user. E.g. how would I even know that my neighbour uses a gaming router? 
Not getting such reports is not the strongest evidence for the absence of the underlaying issue.


Regards
	Sebastian


> 
> JL
>