Re: [Bloat] Quote of the week

[Sebastian Moeller <moeller0@gmx.de>]

Hmmm....

On 23 May 2023 23:05:05 CEST, Dave Taht via Bloat <bloat@lists.bufferbloat.net> wrote:
>from: https://www.ietf.org/blog/banishing-bufferbloat/
>
>"Therefore, innovations like L4S and FQ-CoDel complement each other
>and should be deployed together."

The article is quite positive in its framing which suits its purpose well. It unfortunately does not mention that fq_codel and l4s differ in some arguably important dimensions, like fq_codel has proved its worth (and documented its shortcomings) by over 5 years of in-the-field deployment, while l4s is just about to leave the lab# (with an already impressive list of shortcomings).




Here is a gauntlet for team L4S to demonstrate that 'the market' actually has been waiting for l4s:
Video conferencing streams tend to ada'tively adjust their rates trying to get acceptable quality within the available capacity. However, currently some like teams and discord seem to have a lower bound below which they become quite sluggish in throttling down even further, which is problematic on low capacity or variable capacity links like LTE, as the under-responsive video data will 'clog the pipes' resulting in bad latency for the whole link (assuming no competent AQM on the ISP's to-the-customer link).
So here is the challenge: modify teams/discord to rapidly respond* to l4s signaling even at low rates to improve latency for all, and then convince the maker to actually incorporate those changes at least as user controlled option. Level2: make sure the same method also works trans-oceanicly, that is will allow video conferencing between continents...

Regards

#) Still displeasef that the IETF ratified L4S based essentially on insufficient data about delivery of its promises over the existing internet (testing mostly consisted of repeating the same known working condituons with mild changes). It taught me an important lesson about how the IETFs stated goals and processes differ from what is actually permissable inside the process. So for one I appreciate the flexibility and willingness to make some compromises, but on the other hand that flexibility apparently is easy to game for actors willing to invest time.

*)IMHO this will introduce painful decisions about video resolution and quality, up to dropping video completely and switching to audio only.
>
>-- 
>Podcast: https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
>Dave Täht CSO, LibreQos
>_______________________________________________
>Bloat mailing list
>Bloat@lists.bufferbloat.net
>https://lists.bufferbloat.net/listinfo/bloat

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.