[Cake] Re: [Starlink] [LibreQoS] Re: Re: Keynote: QoE/QoS - Bandwidth Is A Lie! at WISPAPALOOZA 2025 (October 16)

[Sebastian Moeller <moeller0@gmx.de>]

Hi J,


> On 8. Nov 2025, at 18:03, J Pan via Starlink <starlink@lists.bufferbloat.net> wrote:
> 
> yes, availability (at least two competing network providers with
> reliable services),

As David already mentioned that gets us a duopoly, but going mildly higher still results in an oligopoly... As a market realist (that is someone who accepts efficient market when he sees them, but does not naive believe in the fairy tales of the invisible hand of the market) I think that we would be often much better off with a competently managed/regulated monopoly than with duo- to oligopolies that are treated as if they were efficient markets... Infrastructure (and at least access networks are at least infrastructure-ish IMHO) is not something where the free market typically excels at.

> affordability (so the competition to bring the
> price and cost down)

I agree, but that is really at odds with your first point, to get that from a market we clearly need to grow the supply side to get out of oligopoly territory, and I am not sure that that is actually feasible.

> and applicability to modern internet applications
> (video streaming, conferencing and gaming in addition to email and web
> browsing) shall be the user-centric metrics in addition to throughput,
> latency/jitter, packet loss, etc

I am 100% behind this. I will mention though that I believe that latency increase under load is a decent proxy for the utility of a given access link for the usability with interactive applications.

Regards
	Sebastian

> --
> J Pan, UVic CSc, ECS566, 250-472-5796 (NO VM), Pan@UVic.CA, Web.UVic.CA/~pan
> 
> On Sat, Nov 8, 2025 at 8:00 AM dan <dandenson@gmail.com> wrote:
>> 
>> I'm starting to see the signs that raw bandwidth is starting to lose
>> it's dominance for marketing.  It's still the clear #1 ask but price
>> is rapidly overtaking speed for our customer requests.
>> 
>> I believe we've hit this era's threshold on throughput needs and
>> people have started to notice that 'more' doesn't feel like a faster
>> service.
>> 
>> one common scenario that we are using to win customers, in combination
>> with facebook testimonials, is that people have bad experiences with
>> wifi and they order a faster service from cable/fiber company and the
>> wifi just gets worse.  This scenario I think is incredibly common and
>> seems to be a catalyst for 'speed isn't everything'.  We come in with
>> 50-500Mbps of service and solid whole-home wifi and they are
>> converted.
>> 
>> I hope we're not to far off from having 'speed' be just a feature, not
>> the entire story.
>> 
>> and yes, we QoE or service with cake via libreqos which is the
>> difference between great service and inadequate service IMO.
>> 
>> On Fri, Nov 7, 2025 at 12:50 PM J Pan via LibreQoS
>> <libreqos@lists.bufferbloat.net> wrote:
>>> 
>>> marketing is even worse. some claim 200mbps because 150mbps down and
>>> 50mbps up at peak data rate. of course, this is not the only problem
>>> in telecom, but likely the worst
>>> 
>>> nevertheless, there are stats such as 10% inflation for food and 20%
>>> for gas, so in total 30% ;-) at this rate, any numbers can be floating
>>> around but none are telling the truth ;-)
>>> --
>>> J Pan, UVic CSc, ECS566, 250-472-5796 (NO VM), Pan@UVic.CA, Web.UVic.CA/~pan
>>> 
>>> On Fri, Nov 7, 2025 at 10:55 AM Jim Forster <jim@connectivitycap.com> wrote:
>>>> 
>>>> Exactly so.
>>>> 
>>>> Consumer expectations and service provider marketing may be influenced by memories of experience when transmission delay did matter.  At one time I was very happy with my home ISDN connection, and even shared it with my neighbor.  At about 128kbs, it was three orders of magnitude slower than my home fiber link.  I’ve not run the numbers but I’m pretty sure transimission speed mattered for video, even for crummy quality video,  So then when I learned a bit about digital video, and cable’s 64 QAM 27mbps channels, I got excited and thought, “wow, they could deliver 1mbps service!  And wouldn’t it be cool to have 1M home online at 10x the speed of ISDN?”.  It was cool!  And two more orders of magnitude later, here we are.
>>>> 
>>>>  — Jim
>>>> 
>>>> On Nov 7, 2025, at 12:52 PM, J Pan <Pan@uvic.ca> wrote:
>>>> 
>>>> latency is based on round-trip time, and one-way delay includes
>>>> transmission delay, propagation delay, queuing delay and processing
>>>> delay. bandwidth does affect transmission delay (or serialization
>>>> delay), propagation delay is determined by the link length and the
>>>> "travel" speed of the signal, queuing delay is the hardest part and
>>>> affected by the buffer bloat a lot, and processing delay is another
>>>> variable. of course, transmission delay takes less and less portion of
>>>> the end-to-end delay now due to higher and higher "speed" links
>>>> 
>>>> consumers may mistaken the speed of the link (the "width" of their
>>>> pipe) as how fast their internet is (the "length" of the pipe), due to
>>>> the poor terminology we have been using ;-)
>>>> 
>>>> 
>>> _______________________________________________
>>> LibreQoS mailing list -- libreqos@lists.bufferbloat.net
>>> To unsubscribe send an email to libreqos-leave@lists.bufferbloat.net
> _______________________________________________
> Starlink mailing list -- starlink@lists.bufferbloat.net
> To unsubscribe send an email to starlink-leave@lists.bufferbloat.net