Starlink has bufferbloat. Bad.
 help / color / mirror / Atom feed
From: Frantisek Borsik <frantisek.borsik@gmail.com>
To: Nick Matthews <matthnick@gmail.com>
Cc: Inemesit Affia <inemesitaffia@gmail.com>,
	Dave Taht via Starlink <starlink@lists.bufferbloat.net>
Subject: [Starlink] Re: Amazon Leo announcing up to 1 Gbps (DOWN) and 400 Mbps (UP)
Date: Wed, 10 Dec 2025 02:24:24 +0100	[thread overview]
Message-ID: <CAJUtOOjNfvSOYWX1QDOuNfEuzwatwK-JyofO=+JbPrG1AmFzhg@mail.gmail.com> (raw)
In-Reply-To: <CAM-K-Nr4+G+2nZDY3f=oMR-VNZnp8Vkok0jOWbabSCKxj2_H0Q@mail.gmail.com>

Hey Nick,

Great to see you here. Thanks for the info.

Nice to hear that you guys are aware of bufferbloat.

There are people willing to help here, I’m sure.

All the best,

Frank
Frantisek (Frank) Borsik

In loving memory of Dave Täht: 1965-2025
https://libreqos.io/2025/04/01/in-loving-memory-of-dave/


https://www.linkedin.com/in/frantisekborsik
Signal, Telegram, WhatsApp: +421919416714
iMessage, mobile: +420775230885
Skype: casioa5302ca
frantisek.borsik@gmail.com


On Wed, 10 Dec 2025 at 2:06 AM, Nick Matthews <matthnick@gmail.com> wrote:

> I work at Leo, and am the speaker in that video. I mostly lurk out here.
>
> A few notes from above:
>
>    - The 50ms is based on modeling, yes. We'll know more as we get
>    through preview - production code working, production gateways operating,
>    etc.
>    - The top speeds we've proven, using a single antenna. We've posted on
>    linkedin a video of a unit doing 1.2 gbps. I've seen higher, so the antenna
>    itself is definitely capable. Though as you all know, the magic is in mass
>    traffic/beam management, which is in development and we'll see as we get
>    closer to availability.
>    - Point to point video would be cool. There are some latency
>    advantages, but the vast majority can handle the latency
>    antenna-pop-antenna. Right now everything is encrypted between
>    Antenna<->POP so it makes any Antenna<->Antenna challenging from a
>    security/technical aspect. But, we get a lot of requests for this kind of
>    thing.
>    - There's folks open to various QoS/QoE/queuing mechanisms. We need to
>    finish primary development first, and then it comes down to practicality of
>    implementation at scale. The main factor is getting gateways close to your
>    location and tweaking the various L1/L2/L3 and terminal software first. A
>    number of us are aware of buffer bloat.
>
> -nick
>
> On Tue, Dec 9, 2025 at 2:53 PM Frantisek Borsik via Starlink <
> starlink@lists.bufferbloat.net> wrote:
>
>> Chatted with an ISP, LibreQoS customer, using eero since forever, this is
>> what he see:
>>
>> *"Depends on the models. The eero 5's added CAKE, then the eero 6 and 6+
>> was FQ-CoDel, I think performance issues, and then the eero7 line is back
>> to CAKE. I've not heard anything about them removing it."*
>
>
>>
>> It would be really crazy if they would do it. People are using them
>> because
>> of great mesh AND fixed bufferbloat, for the most part.
>>
>> All the best,
>>
>> Frank
>>
>> Frantisek (Frank) Borsik
>>
>>
>> *In loving memory of Dave Täht: *1965-2025
>>
>> https://libreqos.io/2025/04/01/in-loving-memory-of-dave/
>>
>>
>> https://www.linkedin.com/in/frantisekborsik
>>
>> Signal, Telegram, WhatsApp: +421919416714
>>
>> iMessage, mobile: +420775230885
>>
>> Skype: casioa5302ca
>>
>> frantisek.borsik@gmail.com
>>
>>
>> On Tue, Dec 9, 2025 at 9:29 AM Inemesit Affia <inemesitaffia@gmail.com>
>> wrote:
>>
>> > https://news.ycombinator.com/item?id=30982654
>> >
>> >
>> >
>> https://forums.thinkbroadband.com/fibre/4728289-needing-clarification-on-eero-models-for-fttc-and-fttp.html
>> >
>> >
>> > https://lwn.net/Articles/873848/
>> >
>> >
>> > Here are some links to clarify/muddy the water as appropriate. Like I
>> said
>> > earlier, I don't know the current reality, but there are people who know
>> > it's a thing at Amazon
>> >
>> >
>> > Dec 9, 2025 9:21:15 AM Inemesit Affia <inemesitaffia@gmail.com>:
>> >
>> > I don't know anything non public, I got the information by following
>> Dave
>> > Täht's work online.
>> >
>> > Dec 9, 2025 9:13:32 AM Frantisek Borsik <frantisek.borsik@gmail.com>:
>> >
>> > Yeah, eero has (or had?!) a great CAKE implementation. So now they got
>> rid
>> > of it? WTH
>> >
>> > Can you share more about it? Even if only privately…
>> >
>> > All the best,
>> >
>> > Frank
>> > Frantisek (Frank) Borsik
>> >
>> > In loving memory of Dave Täht: 1965-2025
>> > https://libreqos.io/2025/04/01/in-loving-memory-of-dave/
>> >
>> >
>> > https://www.linkedin.com/in/frantisekborsik
>> > Signal, Telegram, WhatsApp: +421919416714
>> > iMessage, mobile: +420775230885
>> > Skype: casioa5302ca
>> > frantisek.borsik@gmail.com
>> >
>> >
>> > On Tue, 9 Dec 2025 at 9:11 AM, Inemesit Affia <inemesitaffia@gmail.com>
>> > wrote:
>> >
>> >> Amazon also makes residential routers in the EERO brand. I heard they
>> had
>> >> then removed their AQM (in a new product).
>> >>
>> >> Don't know the current state.
>> >>
>> >> But there's institutional knowledge.
>> >>
>> >> Dec 9, 2025 8:48:53 AM Frantisek Borsik via Starlink <
>> >> starlink@lists.bufferbloat.net>:
>> >>
>> >> I can't imagine a situation in which Amazon Leo - without using CAKE or
>> >> FQ-CoDel - would get good results, as Starlink didn't without FQ-CoDel
>> >> Hope that they will talk to us, and get help. Sadly, there is no Dave
>> >> Taht
>> >> here anymore, to chase them for the good of their users :(
>> >>
>> >> Improvements to CAKE are happening; MQ-CAKE (from Toke), some other
>> good
>> >> options are raising, algorithms like FQ-PIE and PURPE CAKE,
>> improvements
>> >> like xTCP2, XDP2: I will be talking about it at
>> >> https://understandinglatency.com - register, if you didn't, yet.
>> >>
>> >> All the best,
>> >>
>> >> Frank
>> >>
>> >> Frantisek (Frank) Borsik
>> >>
>> >>
>> >> *In loving memory of Dave Täht: *1965-2025
>> >>
>> >>
>> >> https://libreqos.io/2025/04/01/in-loving-memory-of-dave/
>> >>
>> >>
>> >> https://www.linkedin.com/in/frantisekborsik
>> >>
>> >> Signal, Telegram, WhatsApp: +421919416714
>> >>
>> >> iMessage, mobile: +420775230885
>> >>
>> >> Skype: casioa5302ca
>> >>
>> >> frantisek.borsik@gmail.com
>> >>
>> >>
>> >> On Tue, Dec 9, 2025 at 2:26 AM David Collier-Brown via Starlink <
>> >> starlink@lists.bufferbloat.net> wrote:
>> >>
>> >> Michael Richardson wrote:
>> >>
>> >> …
>> >>
>> >>
>> >> On 12/8/25 15:57, David Lang via Starlink wrote:
>> >>
>> >> …
>> >>
>> >>
>> >> MS is really /really/ bad at handling anything bigger/more-laggy than
>> >> the original token-ring and ethernet LANs. In the case of SMB, people
>> >> built local caching/middleware boxes using Samba to get the RTT down.
>> >>
>> >> They're the first folks who taught me that it's latency, not bandwidth.
>> >>
>> >> --dave
>> >>
>> >> --
>> >> David Collier-Brown,         | Always do right. This will gratify
>> >> System Programmer and Author | some people and astonish the rest
>> >> davecb@spamcop.net           |              -- Mark Twain
>> >> _______________________________________________
>> >> Starlink mailing list -- starlink@lists.bufferbloat.net
>> >> To unsubscribe send an email to starlink-leave@lists.bufferbloat.net
>> >>
>> >> _______________________________________________
>> >> Starlink mailing list -- starlink@lists.bufferbloat.net
>> >> To unsubscribe send an email to starlink-leave@lists.bufferbloat.net
>> >>
>> >>
>> _______________________________________________
>> Starlink mailing list -- starlink@lists.bufferbloat.net
>> To unsubscribe send an email to starlink-leave@lists.bufferbloat.net
>>
>

  reply	other threads:[~2025-12-10  1:14 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-12-06 15:36 [Starlink] Amazon Leo announcing up to 1 Gbps (DOWN) and 400 Mbps (UP) Frantisek Borsik
2025-12-06 16:36 ` [Starlink] " J Pan
2025-12-06 23:58 ` David Lang
2025-12-07  0:02   ` Inemesit Affia
2025-12-07  0:08     ` David Lang
2025-12-08  4:31       ` Inemesit Affia
     [not found]         ` <5098.1765221714@obiwan.sandelman.ca>
2025-12-08 20:57           ` David Lang
     [not found]             ` <2609.1765228007@obiwan.sandelman.ca>
2025-12-08 21:15               ` David Lang
2025-12-08 21:22                 ` Michael Kan
2025-12-08 23:19                   ` Inemesit Affia
2025-12-08 23:20                     ` Inemesit Affia
2025-12-08 23:33                   ` David Lang
2025-12-09  1:26             ` David Collier-Brown
2025-12-09  7:49               ` Frantisek Borsik
2025-12-09  8:11                 ` Inemesit Affia
2025-12-09  8:23                   ` Frantisek Borsik
2025-12-09  8:21                     ` Inemesit Affia
2025-12-09  8:29                       ` Inemesit Affia
2025-12-09 21:53                         ` Frantisek Borsik
2025-12-10  1:06                           ` Nick Matthews
2025-12-10  1:24                             ` Frantisek Borsik [this message]
2025-12-10  1:45                               ` Joe Hamelin
2025-12-09 13:23       ` Livingood, Jason
2025-12-09 21:33         ` David Lang

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='CAJUtOOjNfvSOYWX1QDOuNfEuzwatwK-JyofO=+JbPrG1AmFzhg@mail.gmail.com' \
    --to=frantisek.borsik@gmail.com \
    --cc=inemesitaffia@gmail.com \
    --cc=matthnick@gmail.com \
    --cc=starlink@lists.bufferbloat.net \
    /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