Starlink has bufferbloat. Bad.
 help / color / mirror / Atom feed
From: Sebastian Moeller <moeller0@gmx.de>
To: Ulrich Speidel <u.speidel@auckland.ac.nz>
Cc: starlink@lists.bufferbloat.net
Subject: [Starlink] Re: Starlink Standby Mode
Date: Wed, 10 Sep 2025 08:00:04 +0200	[thread overview]
Message-ID: <53818AC2-2651-49AC-80A7-A940FE4BCD54@gmx.de> (raw)
In-Reply-To: <c7731f13-37bf-446b-a891-cf0dcbc0eace@auckland.ac.nz>

Hi Ulrich,

thanks for the nice description. I think there is one more (engineering-wise challenging) option, reduce the number of users per cell by shrinking the cells... This likely would require better antennas on the satellites and maybe better dishies, no? Same as with any other segment-limited system, with a given user density, reduce the size and you end up with fewer users each having a higher likelihood of getting a larger share of the pie.
Now here is my question (we might have discussed this before, but my memory is dim), how hard would it be to say reduce the area per cell by a factor or 2? And what is the realistic lower boundary for cell sizes that can be gotten vom the 500 km+ orbits (and respective longer paths when the satellite is not exactly above a dishy)?

Regards
	Sebastian

P.S.: Your scenario of a thundering herd of users rushing to their backup link when something with the primary goes haywire seems to haaapen over here for LTE/5G backups, fully acceptable most of the time except those times when you really need it ;)



> On 10. Sep 2025, at 07:36, Ulrich Speidel via Starlink <starlink@lists.bufferbloat.net> wrote:
> 
> What gets me here is that this is yet more evidence of SpaceX trying to do user density and capacity management.
> 
> Recap: Spectral capacity per cell is limited - to quite how much depends a good bit on how much service neighbouring cells need. They can divide that capacity between users via channel division schemes such as time division, frequency division or code division, but the bottom line is that if you divide capacity, you end up with slices that you can assign to users much like slices from a pie (or pizza). The more users in a cell, the more competition for the slices.
> 
> Note also that there are two ways for Starlink to run out of capacity in a cell:
> 
> 1. They can run out of frequencies to use. This is impossible to fix
>   without someone having to compromise on their slice.
> 2. They can run out of beams - meaning there would be a frequency to
>   communicate on but there's just no satellite in view that has a beam
>   that it can currently point at the cell. This is possible to fix
>   with more satellites / beams per satellite.
> 
> Now when you are starting to run out of capacity, you have a bunch of options at your disposal:
> 
> * You can launch more satellites with more beams to reduce beam
>   capacity bottlenecks. SpaceX clearly do this as their constellation
>   is growing in sat numbers and beams per sat.
> * You can try and deter new users from signing up by charging a higher
>   price in the area you're running out of capacity in, or by charging
>   a congestion fee. We've seen SpaceX do this when they ostensibly ran
>   out of beams a couple of years ago in some areas & didn't extend
>   discount schemes to cells with high user densities. We're seeing
>   them do it now in many areas where new subscribers are tapped for
>   congestion fees.
> * You can growl at existing users to try and make them go away. We've
>   seen SpaceX do this in an island location where local users whose
>   dishys were on a roaming plan but had taken up permanent residence
>   there were told to pack them up and continue operation elsewhere.
> * You can shrink the size of the slices for existing users. If you
>   look at the Starlink speeds map, that's what seems to have happened
>   in a few of the places where they were "sold out" for a while and
>   are now available again. You'll be looking for the minimum download
>   speeds here and want to compare those with places where Starlink
>   user density is low. Of course, this slice shrinking isn't ideal
>   because folk love their bandwidth and all it takes is a cranky
>   influencer who isn't happy with what they see. So it's best avoided
>   if you can.
> 
> Last but not least, you can try and put the squeeze on the amount of spare capacity you need to retain in case not-so-active users become active. That's you guys with the "I have my Starlink only for backup in case my fibre gets cut" or those of you with the "I only use my Starlink at my summer house during the holidays". While you're not using your Starlink, you don't need any slices, so SpaceX can sell that capacity in the area to new users. But woe betide them should all those inactive users suddenly activate - be it because the holidays have started, or because that backhoe or natural disaster has hit.
> 
> So what you can do is sell smaller slices of the capacity pie and at a cheaper price, but make that "official" so nobody complains. Similarly, you can try and ease those out that put their plans on hold for extended periods of time - you want to make it unattractive for them to hold on to their entitlement to a big slice at the drop of a hat.
> 
> TLDR: I'd be a little skeptical that you'll be able to switch from a backup plan or pause to a full one as easily as you might think.
> 
> On 10/09/2025 9:20 am, J Pan via Starlink wrote:
>> nothing better than free ;-) but it may cost them more than $5/mo to
>> maintain an active dish. don't they just want to keep as many
>> revenue-generating customers for an upcoming ipo?
>> --
>> J Pan, UVic CSc, ECS566, 250-472-5796 (NO VM),Pan@UVic.CA, Web.UVic.CA/~pan
>> 
>> On Tue, Sep 9, 2025 at 2:01 PM Oleg Kutkov via Starlink
>> <starlink@lists.bufferbloat.net> wrote:
>>> Well. I prefer a good old free pause.
>>> I have seven Starlink terminals on my account, and I mostly don't use
>>> them, except for some experiments, occasional tests, firmware dumps, and
>>> similar purposes.
>>> Now they will charge me $35 each month for basically nothing.
>>> 
>>> On 9/9/25 23:14, Luis A. Cornejo via Starlink wrote:
>>>> There is another electric cooperative in my county that did just that.
>>>> Strung fiber along the posts, I was not lucky enough to be part of their
>>>> territory. But from people that I know it’s similar that’s it’s fast but
>>>> not always reliable, so storms can take some down, often around here it’s
>>>> also a backhoe, and when it does down, it’s down for a while.
>>>> 
>>>> But I agree, it’s a great backup. Although it probably costs almost as much
>>>> to run in electricity as the service itself! =o)
>>>> 
>>>> -Luis
>>>> 
>>>> On Tue, Sep 9, 2025 at 1:41 PM Colin_Higbie via Starlink <
>>>> starlink@lists.bufferbloat.net> wrote:
>>>> 
>>>>> I find the Standby mode to be a great backup option. $5/mo for unlimited
>>>>> low-bandwidth usage. This is something I've been urging them to offer for
>>>>> about a year now.
>>>>> 
>>>>> About 18 months after I originally purchased and subscribed to Starlink,
>>>>> our power company (not phone company as you'd expect, really our electric
>>>>> power company) rolled out fiber to rural communities in northern NH. As
>>>>> great as Starlink was compared to what we had before, fiber is even better
>>>>> now that it's available (1Gbps for $79/mo, consistently tests as A on
>>>>> Bufferbloat). But it's not 100% reliable. For example, when there are
>>>>> widespread power outages, it goes down. The Starlink standby option is
>>>>> perfect for those situations.
>>>>> 
>>>>> Latency remains decent, just limited bandwidth. And if the fiber outage
>>>>> remains in effect for too long, we could always activate Starlink at full
>>>>> bandwidth for that month, where it appears they're taking that option away
>>>>> from people with inactive accounts not already on Standby mode (they may be
>>>>> bluffing on that – you'd think they would want to make it easy for anyone
>>>>> to give them money and resubscribe, standby customer or not).
>>>>> 
>>>>> Cheers,
>>>>> Colin
>>>>> _______________________________________________
>>>>> Starlink mailing list --starlink@lists.bufferbloat.net
>>>>> To unsubscribe send an email tostarlink-leave@lists.bufferbloat.net
>>>>> 
>>>> _______________________________________________
>>>> Starlink mailing list --starlink@lists.bufferbloat.net
>>>> To unsubscribe send an email tostarlink-leave@lists.bufferbloat.net
>>> --
>>> Best regards,
>>> Oleg Kutkov
>>> 
>>> _______________________________________________
>>> Starlink mailing list --starlink@lists.bufferbloat.net
>>> To unsubscribe send an email tostarlink-leave@lists.bufferbloat.net
>> _______________________________________________
>> Starlink mailing list --starlink@lists.bufferbloat.net
>> To unsubscribe send an email tostarlink-leave@lists.bufferbloat.net
> 
> -- 
> ****************************************************************
> Dr. Ulrich Speidel
> 
> School of Computer Science
> 
> Room 303S.594 (City Campus)
> 
> The University of Auckland
> u.speidel@auckland.ac.nz http://www.cs.auckland.ac.nz/~ulrich/
> ****************************************************************
> 
> 
> _______________________________________________
> Starlink mailing list -- starlink@lists.bufferbloat.net
> To unsubscribe send an email to starlink-leave@lists.bufferbloat.net


  reply	other threads:[~2025-09-10  6:00 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <175739769285.1561.9299932820839760282@gauss>
2025-09-09 18:41 ` Colin_Higbie
2025-09-09 20:14   ` Luis A. Cornejo
2025-09-09 21:01     ` Oleg Kutkov
2025-09-09 21:20       ` J Pan
2025-09-10  5:36         ` Ulrich Speidel
2025-09-10  6:00           ` Sebastian Moeller [this message]
2025-09-10  7:53           ` David Lang
2025-09-10 10:24             ` Ulrich Speidel
2025-09-10 16:01               ` J Pan
2025-09-08 21:34 [Starlink] " Luis A. Cornejo
2025-09-09  5:35 ` [Starlink] " J Pan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://lists.bufferbloat.net/postorius/lists/starlink.lists.bufferbloat.net/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=53818AC2-2651-49AC-80A7-A940FE4BCD54@gmx.de \
    --to=moeller0@gmx.de \
    --cc=starlink@lists.bufferbloat.net \
    --cc=u.speidel@auckland.ac.nz \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox