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

[Sebastian Moeller <moeller0@gmx.de>]

Hi David,


> On 10. Nov 2025, at 12:41, David Fernández via Starlink <starlink@lists.bufferbloat.net> wrote:
> 
> In Spain ISDN was replaced by DSL and now both, HFC and DSL and the POTS
> have been replaced by FTTH.

I would guess by DSL you mostly mean ADSL? Over here ADSL is only used on long rural lines everywhere else it is mostly VDSL2@FTTC with 100/40 or even 250/40 Mbps plans available on most links. The incumbent still spent time upgrading to FTTC while Spain was already focussing on FTTH. The results are twofold:
a) most users can get access rates via DSL that make most application work well enough
b) we did not start the FTTH roll-out while interest rates were low ;)

Also, to add insult to injury, I have seen estimates for the average home connected for Spain and Francs (unsure why these were lumped together) of around 200-300 Euro, while the cost in Germany is more in the 2000-4000 Euro range... (there are almost no utility poles in use in Germany and putting cables underground is slower and more expensive).

> 
> Minimum you can get is 100 Mbit/s and max 10 Gbit/s (symmetrical),

No doubt that stiff gets advertized as 10 Gbps, but I am willing to bet that this mostly is XGS-PON, likely with FEC and that simply does not have a gross rate of 10 Gbps.... (XGSPON has a gross rate of 9.95328 and with the 216/248 stadard FEC this gets us to a shared 9.95328 * (216/248) = 8.66899 Gbps, even less of 10 Gbps, but I understand marketing loves round numbers)

> but for
> households max 1 Gbit/s.

Interesting, Switzerland and France offer >> 1 Gbps even for private households (not that typical houlsehold would notice the difference between 1 and say 2 Gbps in normal use).

> 
> Prices start from ~20 euros/month.

Sweet, proces over here rather start at 35 Euros, or 45 for fiber...

> 
> The quality of the service that ISPs with revenues higher than 20 millions
> are obliged to report follows ETSI standards (mainly speed tests), but will
> include latency in 2026 ( as measured by IETF RFC 2681), jitter and packet
> losses:
> 
> https://www.cnmc.es/sites/default/files/5753856.pdf

I occasionally am puzzled by the regulators... nothing bad about that RFC, except is seems rather complicated for a simple use case, and more to the point the regulator failedf to require latency under load numbers, which tend to be quite informative how well a link will do under sustained loads around saturation.


Thanks for the information
	Sebastian

> 
> Regards
> 
> David
> 
> 
> Date: Mon, 10 Nov 2025 07:27:37 +0100
>> From: Sebastian Moeller <moeller0@gmx.de>
>> Subject: [Starlink] Re: [LibreQoS] Re: Re: Keynote: QoE/QoS -
>>        Bandwidth Is A Lie!  at WISPAPALOOZA 2025 (October 16)
>> To: Jim Forster <jim@connectivitycap.com>
>> Cc: J Pan <Pan@uvic.ca>, dan <dandenson@gmail.com>, Frantisek Borsik
>>        <frantisek.borsik@gmail.com>, Cake List <
>> cake@lists.bufferbloat.net>,
>>        bloat <bloat@lists.bufferbloat.net>, codel@lists.bufferbloat.net,
>>        libreqos <libreqos@lists.bufferbloat.net>, l4s-discuss@ietf.org,
>>        starlink@lists.bufferbloat.net
>> Message-ID: <CDF65ACE-06AD-4C4A-8F74-14B704B89637@gmx.de>
>> Content-Type: text/plain; charset=utf-8
>> 
>> 
>> 
>> On 10 November 2025 05:48:38 CET, Jim Forster <jim@connectivitycap.com>
>> wrote:
>>>> On Nov 8, 2025, at 1:11 PM, Sebastian Moeller <moeller0@gmx.de> wrote:
>>>> 
>>>> 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.
>>> 
>>> 
>>> Yeah, I also don’t think there’s an efficient, fair, market here that
>> gets us what we w. In some ways, the Digital Divide is an expected outcome
>> of capital allocation decisions by deregulated companies in a sector that
>> has economies of scale and network effects.
>> 
>> Indeed... I just note that the POTS network was much more comprehensive in
>> its reach due to stricter regulation...
>> 
>>> 
>>> At the same time, a "competently managed/regulated monopoly” may be as
>> uncommon as Homo Economicus sitings are.
>> 
>> Na, only if we put our aim for competence too high ;) . Over here
>> electricity, water and street "networks" are dece
>> ntly regulated infrastructure.
>> 
>>> Which example can you cite? NZ? UK? SE? And have they transitioned
>> smoothly to new technology that would diminish the value of their existing
>> infrastructure?
>> 
>> Tricky... for infrastructure in general I believe there are loads of
>> examples in Europe, for internet access networks it gets a bit trickier,
>> but there are some examples of combining a single network with operator
>> competition. (And that is my preferred model, monopoly network with
>> regulated and fair access for operators, and then have as many operators as
>> possibke offer their services over that network). But partial examples
>> exist, e.g. the fiber network built in Amsterdam, or the point to point
>> fiber network in switzerland where the incumbent built most of the ftth
>> network and is regulated to physically unbundle individual lines to end
>> customers, resultung in surprising competition of ISPs operating different
>> technology over the same fibers (swisscom uses xgspon, salt.ch uses their
>> own xgspin OLTs, init7 uses AON up to 25 Gbps). Sweden also seems to have a
>> decent (albeit not regulated) separation between network operators and ISPs
>> that offer services over these networks.
>> 
>> 
>>> 
>>> I recall that in the US prior to the .com boom, the telco’s idea of
>> broadband was ISDN or maybe DSL or SMDS. They wrote many papers, had lots
>> of trials, but did not aggressively do broadband,
>> 
>> Yes, I agree that the old model of a vertically integrated full service
>> telco breed complacency and was not ideal either (even though the POTS
>> network had better reach than the HFC networks).
>> 
>> : 'Everyone knew’ that the cablecos’ HFC would never work, and that they
>> could not do digital and certainly not voice,  HFC worked, and DOCSIS was a
>> big success. That pressured the telcos to start actually deploying DSL, but
>> it was too late, and the cablecos have dominated US broadband for a couple
>> of decades.
>> 
>> The outcome in Germany was different... hfc networks only ever reached
>> around 75% of households and never exceeded 10 of estimated 45 million
>> access sites for broadband services, while DSL still serves almost 23
>> million (and reaches almost all 45 million).But yes on the technology side
>> it likely was hfc's pressure that sped up dsl development.
>> Now, the german market is a bit odd, as customers are neither terribly
>> hungry for high capacity nor terribly price sensitive (the old ex-monopoly
>> telco still serves most dsl customers in spite of being more expensive due
>> to valid regulatory interventions).
>> 
>> Regards
>>        Sebastian
>> 
>> P.S.: I understand that in this question there are of course multiple
>> equally valid and justifyable positions one could take, this just happens
>> to be mine. A couple of friendly ISPs for example reject this idea as they
>> consider access networks to be a field where ISPs can differentiate and
>> compete (some of them however proposed a regulated middle mile to be able
>> to economically reach IXs and peering points to even the playing field).
>> 
>>> 
>>> Jim
>>> 
>>> 
>>>> 
>>> 
> _______________________________________________
> Starlink mailing list -- starlink@lists.bufferbloat.net
> To unsubscribe send an email to starlink-leave@lists.bufferbloat.net