Re: [tsvwg] feedback and thoughts L4S / SCE

[Pete Heist <[e-mail]>]

> On Nov 28, 2020, at 5:24 PM, Mikael Abrahamsson <swmike=[e-mail]> wrote:
> 
> On Sat, 28 Nov 2020, Jonathan Morton wrote:
> 
>> Take a random sampling of CPE devices, and see how many of them use the above - or it might be easier to count the ones that don't.  I'll wait.
> 
> My ubiquiti APs come with it, but it's default turned off.
> 
> It's not only important of the devices have the capability, it also needs to be default turned on so it's actually used (considering most people don't change config much).

Hi Mikael,

I just checked some Open Mesh devices at a site I have access to. Airtime fairness is enabled by default, with no flag exposed in the admin web interface for turning it off:

root@xxx:/sys/kernel/debug/ieee80211/phy0/netdev:mesh0/stations/ac:86:74:xx:xx:xx# cat airtime 
RX: 880257297 us
TX: 71233699 us
Deficit: VO: 102 us VI: 300 us BE: 2 us BK: 300 us 
root@xxx:/sys/kernel/debug/ieee80211/phy0/netdev:mesh0/stations/ac:86:74:xx:xx:xx# cat /sys/kernel/debug/ieee80211/phy0/ath9k/airtime_flags 
7

Their product is more closely tied to OpenWrt than Ubiquiti’s UniFi is, so they appear to have inherited OpenWrt’s default.

Meanwhile, I provided some stateless counter data in a separate thread showing apparent AQM activity in the form of CE marks and ECE flags coming from both inside and outside an ISP. I’m working on improving the data collection method using conntrack and ipsets, but that will take time to complete, and I’m not sure yet if I’ll be able to run it in a production environment. I would point out though that since the severity of problems in RFC3168 queues is high, the tolerance for the likelihood of it occurring should be low.

Regards,
Pete