[Starlink] Starlink in Northern Europe: A New Look at Stationary and In-motion Performance
Ulrich Speidel
u.speidel at auckland.ac.nz
Thu Feb 27 23:22:03 EST 2025
I noticed this some time ago but it seems to come up as "available" again.
There are two ways in which Starlink can run out of capacity:
1. Lack of beams. This is fixable by having more sats in orbit in your
neighbourhood (or sats with more beams).
2. Lack of spectrum. That's when you have enough satellites with beams
available but you can't use the extra beams because they'd collide
at the receiver with another beam on the same frequency. This isn't
really fixable with more beams or sats if you've used up all
available spectrum. They don't really make any more of that ;-) What
you can in principle do there is make beams narrower (=cells
smaller), but that's a matter of flying appropriate hardware in
order to increase frequency re-use.
Some time late last year, a lot of areas around the globe started
popping up as "sold out" on Starlink's availability. These areas have
since increased, especially in the Americas. For all I could tell, most
of these areas have the following in common: significant population
density and severe lack of terrestrial broadband infrastructure. A lot
of that looks like they're running out of spectrum, especially since
there's been relatively little in terms of areas coming back on-stream
for sales. Note that when you get away from the "crowded" target area
("other parts of the Yukon"), that problem goes away.
Note that given Starlink's roaming plans, they have to throttle back on
selling fixed units in an area well before they hit spectral capacity
there. That's so they can accommodate roaming units that come into the
area. This may have been the case last year with a lot of RVs coming up
to Whitehorse over summer, with pressure now relenting over winter,
offering a breather (and perhaps reworked roaming rules for next
summer's roaming units). However that isn't really a perfect solution:
Locals desperate for a domestic connection can bring in roaming units
and use them in a de-facto fixed location, even if they're more
expensive to run than the fixed unit that their neighbour bought
earlier. This has also been the case in places where Starlink offers
roaming service but isn't locally licensed to offer fixed service
(Kiribati comes to mind).
SpaceX seem to have been addressing this by roaming price increases,
attempting to justify this with the fact that you can now take your RV
with the unit offshore for a few miles and still get service under
maritime coverage. Now I'm not sure that this is what a lot of people do
given that most RVs don't float all that well. SpaceX have also tried to
restrict the amount of time that you can operate a roaming unit in the
same place (and have sent comms to users exceeding that time requesting
them to move the units - or else). I guess there's probably a business
plan in renting out roaming units on a rotating basis for a few weeks at
a time ;-)
On 28/02/2025 4:40 pm, Daniel AJ Sokolov via Starlink wrote:
> I live North of 60, in Whitehorse in the Yukon. With about 35,000
> inhabitants, We are the largest Canadian city in the North.
>
> Here, we could get Starlink for a while. But last year, they
> introduced a waiting list. There is no more capacity to go around.
>
> I assume this will remain so for a couple more years, until we get new
> satellites.
>
> In other parts of the Yukon, I am not aware of waiting lists.
>
> Cheers
> Daniel AJ
>
>
>
> On 2025-02-27 at 14:02, Ulrich Speidel via Starlink wrote:
>> I had a quick look.
>>
>> The most important bit of information I was looking for is on page 7,
>> and it's not explicitly mentioned despite its importance - rather
>> it's delivered on the side of the figures: the latitude of the
>> measurements. Ballpark 65 deg north. That puts the measurements
>> beyond the range of the bulk of the Starlink shells at 43, 53, and
>> 53.2 degrees inclination, leaving only the 70 and 97.6 deg
>> inclination shells within view.
>>
>> Why does this matter? Two reasons:
>>
>> 1. A location at 65 deg north sees on average around 8 qualifying
>> satellites at any time - those are satellites that are at least 25
>> deg above the horizon (so their beams don't get into terrestrial
>> microwave link receivers). That compares to over 40 qualifying
>> satellites should you find yourself luck to live between 40 and 45
>> deg north, and over 20 at the Equator (even keeping GSO protection
>> into account).
>> 2. The qualifying satellites you see north of about 60 deg are still
>> >90% version 1.5's. They have lasers for backhaul but a
>> comparatively small number of Ku band beams for downlink to Dishy.
>> South of 40 degrees, almost half the qualifying satellites you're
>> going to encounter are from the version 2 series, which have a lot
>> more beams. These beams are also higher capacity ones.
>>
>> Why does the number of qualifying satellites and beams matter?
>> Basically, if you add up all beams on all satellites within view, you
>> get the pool of beams that Starlink can pick from to serve your
>> Dishy. More beams in total = more options = bigger cake = bigger
>> slice of capacity for your Dishy.
>>
>> Now how big a slice of the cake you can get depends not only on the
>> satellite mix in view, but also on how many other user terminals in
>> your immediate (cell) and wider (nearby cells) in your neighbourhood
>> want to access that capacity cake. This depends a lot on population
>> density and on what the competing terrestrial connectivity options
>> are. In a place with low population density, fibre to almost
>> everywhere and a good 4G and 5G coverage, all at good prices, there
>> won't be a lot of competing users for the cake. The Oulu area in
>> Finland, where they took the measurements, appears to be in that
>> category, mostly. The paper doesn't discuss these determinants of
>> performance, however.
>>
>> On 28/02/2025 4:04 am, Hesham ElBakoury via Starlink wrote:
>>> Hi Craig,
>>> No it is not my paper.
>>> It has interesting results that I would like others to see and
>>> provide feedback on.
>>>
>>> Hesham
>>>
>>> On Thu, Feb 27, 2025, 6:36 AM Craig Polk <c.polk at comsoc.org> wrote:
>>>
>>> Hesham,
>>>
>>> Is this your paper? Are you submitting it for the WG to review as
>>> a possible INGR Topic article?
>>>
>>> Best regards,
>>> Craig
>>>
>>> ----
>>> Craig Polk, MSEE, MBA
>>> Program Manager
>>> Future Networks Tech Community | futurenetworks.ieee.org
>>> <http://futurenetworks.ieee.org>
>>> 3 Park Avenue, 17th Floor, New York, NY 10016
>>> <https://www.google.com/maps/search/3+Park+Avenue,+17th+Floor,
>>> +New+York,+NY+10016%C2%A0++Office:+%2B1?entry=gmail&source=g>
>>> Office: +1
>>> <https://www.google.com/maps/search/3+Park+Avenue,+17th+Floor,
>>> +New+York,+NY+10016%C2%A0++Office:+%2B1?entry=gmail&source=g>
>>> 212-705-8926 | Mobile: +1 908-255-6568
>>> Email: c.polk at comsoc.org
>>> Future Networks World Forum | https://fnwf.ieee.org/
>>> Connecting the Unconnected | https://ctu.ieee.org/
>>>
>>> On Thu, Feb 27, 2025, 12:01 AM Hesham ElBakoury
>>> <helbakoury at gmail.com> wrote:
>>>
>>> This paper [1] This paper evaluates the Flat High Performance
>>> (FHP) terminal's performance in Finland, Northern Europe.
>>>
>>> *_Abstract_*
>>> "Starlink has introduced the Flat High Performance (FHP)
>>> terminal, specifically designed to support the vehicles and
>>> the vessels in motion as well as the high-demand stationary
>>> users. The research on FHP terminal throughput analysis
>>> remains limited, only a few existing studies evaluate FHP,
>>> focusing on the limited parameters and scenarios. This paper
>>> evaluates the FHP terminal's performance in Finland, Northern
>>> Europe. We examine round-trip time (RTT), uplink, and downlink
>>> throughput for both stationary and in-motion use. We measure
>>> network efficiency across six geographically diverse servers
>>> and get insights of network routing strategies. Our results
>>> show that Starlink provides high-speed, low-RTT connectivity,
>>> however, the throughput experiences fluctuations with slight
>>> degradation when in motion. Additionally, we compare Starlink
>>> and terrestrial network RTT and possible routing paths."
>>>
>>> Hesham
>>> [1] https://arxiv.org/abs/2502.15552
>>> ------------------------------------------------------------------------
>>>
>>>
>>> To unsubscribe from the 5GRM-SATELLITE list, click the
>>> following link:
>>> https://listserv.ieee.org/cgi-bin/wa?SUBED1=5GRM-SATELLITE&A=1
>>> <https://listserv.ieee.org/cgi-bin/wa?SUBED1=5GRM-SATELLITE&A=1>
>>>
>>>
>>> _______________________________________________
>>> Starlink mailing list
>>> Starlink at lists.bufferbloat.net
>>> https://lists.bufferbloat.net/listinfo/starlink
>>
>>
>> _______________________________________________
>> Starlink mailing list
>> Starlink at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/starlink
>
> _______________________________________________
> Starlink mailing list
> Starlink at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
--
****************************************************************
Dr. Ulrich Speidel
School of Computer Science
Room 303S.594 (City Campus)
The University of Auckland
u.speidel at auckland.ac.nz
http://www.cs.auckland.ac.nz/~ulrich/
****************************************************************
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/starlink/attachments/20250228/1deb3f28/attachment-0001.html>
More information about the Starlink
mailing list