[Bloat] [Starlink] Instant Messaging Meets Video Conferencing: Studying the Performance of IM Video Calls
Rodney W. Grimes
starlink at gndrsh.dnsmgr.net
Sun Jul 2 02:27:51 EDT 2023
> On Sat, Jul 1, 2023 at 12:23?AM Rodney W. Grimes
> <starlink at gndrsh.dnsmgr.net> wrote:
> >
> > > This paper does a really good job of measuring the impacts of tcp
> > > cross traffic, including BBRv2, against the videoconferencing
> > > subsystems in signal, telegram, and whatsapp.
> >
> > Did I miss something? The paper only shows a dumb bell topology with
> > instream traffic, I did not see any cross flow traffic testing.
>
> Do we share the same definition of cross traffic? The testbed consists
> of fifo, codel, and fq_codel queues, against 1-2 tcp flows, against a
> p2p videoconferencing flow.
What they did was measure competing flows on a COMMON PATH.
Using a COMMON path, that is not cross flow, at least by my understanding
of the definition of cross flow. The fact that they located
the competing flows ON THE SAME NODES it means only one flow or the
other flow can transmit a packet at a single instance in time, this
insuring that the transmission of packets is serialized by the node
between the competing flows.
>
> I am sometimes just as subject to confirmation bias as the next guy,
> and in re-reading this this morning I would have liked it if they had
> cited rfc8290, and extended their tests to use a bigger BDP for the
> fifo as well as tested web PLT. a shorter RTT would also be
> interesting in the case of a ptp app such as this.
>
> An exciting part to me remains in exploring the videoconferencing QoE
> technique they discuss (Brisque) which I haven?t found sources for
> yet...
>
>
> >
> > >
> > > https://tma.ifip.org/2023/wp-content/uploads/sites/12/2023/06/tma2023-final37.pdf
> > >
> > > Naturally fq_codel shines. The paper also introduces a new means of
> > > measuring videoconferenceQoE, called BRISQUE.
> > >
> > > About my only kvetch is they didn't try cake, which has a different
> > > shaper, drops less packets, has less collisions, as well as diffserv
> > > support.
> > >
> > > Seeing BBRv2 stats and impact was very interesting also.
> > >
> > >
> > >
> > >
> > > --
> > > Podcast: https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
> > > Dave T?ht CSO, LibreQos
> > > _______________________________________________
> > > Starlink mailing list
> > > Starlink at lists.bufferbloat.net
> > > https://lists.bufferbloat.net/listinfo/starlink
> > >
> > >
>
>
>
> --
> Podcast: https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
> Dave T?ht CSO, LibreQos
>
>
More information about the Bloat
mailing list