Hi I reacted mostly on the extremely low throughput in the top post in the twitter thread, it is possible that radio link failure caused a large amount of bloat in this case. The amount of buffer bloat in 5G systems is quite vendor specific. So it is hard to give any good numbers what one can expect. But sure, seconds of bloat is gross overkill and possible evidence that AQM is disabled. Typically the AQM on the network side handles the downlink traffic, while for the uplink traffic the AQM needs to be on the modem side. L4S, which is being standardized in 3GPP can reduce bloat quite considerably, but one need to realize that L4S does perhaps not fit for all traffic. See page 5 in https://datatracker.ietf.org/meeting/115/materials/slides-115-iccrg-5g-and-l4s-congestion-control-considerations-00
/Ingemar From: Dave Taht <[email protected]> Sent: Monday, 8 May 2023 21:17 To: Ingemar Johansson S <[email protected]> Cc: [email protected] Subject: Re: [Bloat] ookla speedtests on 5G On Fri, May 5, 2023 at 6:55 AM Ingemar Johansson S via Bloat <[email protected]<mailto:[email protected]>> wrote: 1.08Mbps down and 0.33Mbps up is hardly representative for 5G.. I wonder if the measurement was done three stories down in a basement :-) the twitter thread itself was a pretty representative example of how bad the bloat can get, even at the highest 5G speeds. /Ingemar > > ---------------------------------------------------------------------- > > Message: 1 > Date: Wed, 3 May 2023 18:32:56 -0700 > From: Dave Taht <[email protected]<mailto:[email protected]>> > To: Rpm <[email protected]<mailto:[email protected]>>, bloat > <[email protected]<mailto:[email protected]>>, > libreqos > <[email protected]<mailto:[email protected]>> > Subject: [Bloat] ookla speedtests on 5G > Message-ID: > > <caa93jw5gm3ujeemmjjv81vxspkigzbjo64g7j6tkbjxxk93...@mail.gmail.com<mailto:caa93jw5gm3ujeemmjjv81vxspkigzbjo64g7j6tkbjxxk93...@mail.gmail.com> > > > Content-Type: text/plain; charset="UTF-8" > > This was quite the long list of speedtest pics, all showing pretty > gnarly bufferbloat across all services. > > https://twitter.com/ianzelbo/status/1653757825972142081 > > Not a single poster caught on to the awfulness in the responsiveness > numbers. > > > -- > Podcast: > https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937 > / > Dave Täht CSO, LibreQos > > > ------------------------------ _______________________________________________ Bloat mailing list [email protected]<mailto:[email protected]> https://lists.bufferbloat.net/listinfo/bloat<https://protect2.fireeye.com/v1/url?k=31323334-501d5122-313273af-454445555731-5f070e024f500e6b&q=1&e=12b802be-c3ee-4da6-bc15-6d9b81605a53&u=https%3A%2F%2Flists.bufferbloat.net%2Flistinfo%2Fbloat> -- Podcast: https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/ Dave Täht CSO, LibreQos
_______________________________________________ Bloat mailing list [email protected] https://lists.bufferbloat.net/listinfo/bloat
