[Rpm] VI vs BE on wifi

[Rpm] VI vs BE on wifi

[Dave Taht]

https://forum.openwrt.org/t/reducing-multiplexing-latencies-still-further-in-wifi/133605/32

If you squint, you can see what's left of the BE defaults when competing with VI


-- 
FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
Dave Täht CEO, TekLibre, LLC

Re: [Rpm] [Make-wifi-fast] VI vs BE on wifi

[Sebastian Moeller]

Hi Dave,

in all fairness it depends a bit on the station AP combination how much AC_BE traffic gets hit. But essentially this graph (or rather similar graphs fro my own testing a few years back) convinced me that selecting the "L4S" DSCP to ensure mapping into AC_VI is massively misguided, to be friendly. (This also nicely illustrates the folly of https://datatracker.ietf.org/doc/html/rfc8325 Mapping Diffserv to IEEE 802.11, which does not seem to cite even a single reference showing data that the proposed scheme actually survives a contact with the existing WiFi gear in the field)...


> On Aug 8, 2022, at 19:03, Dave Taht via Make-wifi-fast <make-wifi-fast@lists.bufferbloat.net> wrote:
> 
> https://forum.openwrt.org/t/reducing-multiplexing-latencies-still-further-in-wifi/133605/32
> 
> If you squint, you can see what's left of the BE defaults when competing with VI
> 
> 
> -- 
> FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
> Dave Täht CEO, TekLibre, LLC
> _______________________________________________
> Make-wifi-fast mailing list
> Make-wifi-fast@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/make-wifi-fast