Starlink has bufferbloat. Bad.
 help / color / mirror / Atom feed
From: Dave Taht <dave.taht@gmail.com>
To: Darrell Budic <budic@onholyground.com>
Cc: starlink@lists.bufferbloat.net
Subject: Re: [Starlink] Starlink throughput / bufferbloat change?
Date: Sun, 27 Jun 2021 18:38:00 -0700	[thread overview]
Message-ID: <CAA93jw6je-E+70aaW1PdGPTaZLpE=i0nOuSFnt_6e4varpLEsQ@mail.gmail.com> (raw)
In-Reply-To: <CE32F49F-909E-44F7-8BAC-051DB4967750@onholyground.com>

Dear Darrell:

Toke ended up writing a new test that we haven't deployed yet

 https://github.com/tohojo/flent/commit/ec5dd227cec7f0b8316a7e5c8551fd44c4f9a985

which works for the flent tests (like rrul) that use irtt. Install
flent from git... Try a rrul test from any location in the cloud, or
at home, or over wifi or 5g to fremont.a.taht.net (one of my
measurement servers in the cloud)... or dallas.starlink.taht.net

A dishy is available without sqm on between 1AM and 6AM PDT at
dishy.taht.net (through an ipv6 wireguard tunnel.  (you are
d2.taht.net)

Moving forward, simplifying things (single flows) and since irtt was
not used by our tcp_nup and tcp_ndown tests, a new one "nup"  is under
development which has a few new things like SCE support.

I do have some data on you from today from my cloud, but hey, if you
have energy... please try out something like

flent -H dishy.taht.net --socket-stats --step-size=.05 -t
'whateverthetestcondidtions" --irtt-sampling-interval 3ms -l 100 rrul

and let us know what you find? (the flent-gui can generate plots, you
can also output plots from the command line) The tcp RTT plot is my
favorite.

On Sun, Jun 27, 2021 at 6:26 PM Darrell Budic <budic@onholyground.com> wrote:
>
> Hey Dave, how’d those tests look yesterday (sqm enabled though) vs today (disabled)? Looks like my dishy upgraded to
>
> 4990ce8d-5028-4e51-a015-e9ab1b1ebe1a.uterm.release
>
> around 4:30AM CDT this morning.
>
> Nathan, did you get a chance to check versions or uptime? Probably worth tracking if we’re looking for performance changes.
>
> On Jun 26, 2021, at 9:24 AM, Dave Taht <dave.taht@gmail.com> wrote:
>
>
> I have fired off a new string of tests. should be done in an hour.
>
> I don't know if budic has sqm on or off at the moment.
>
> On Sat, Jun 26, 2021 at 7:10 AM Jared Mauch <jared@puck.nether.net> wrote:
>
>
> I will be able to test perhaps in 6-10 hours as I'll be getting the system online hopefully.
>
> Sent from my TI-99/4a
>
> On Jun 26, 2021, at 9:38 AM, Nathan Owens <nathan@nathan.io> wrote:
>
> 
> I'm on vacation, so I can't do a ton of testing, but it seems like in the last ~2 days, my DL speeds are lower (<160Mbps, generally 90-110Mbps), but download-induced bufferbloat is also greatly reduced (Latency generally <60ms, with small spikes up to maybe 150ms more rarely  -- much better than to up to 400ms I was seeing before).
>
> I confirmed the same behavior with one other list member, has anyone else seen similar?
>
> --Nathan
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>
>
>
>
> --
> Latest Podcast:
> https://www.linkedin.com/feed/update/urn:li:activity:6791014284936785920/
>
> Dave Täht CTO, TekLibre, LLC
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>
>


-- 
Latest Podcast:
https://www.linkedin.com/feed/update/urn:li:activity:6791014284936785920/

Dave Täht CTO, TekLibre, LLC

  reply	other threads:[~2021-06-28  1:38 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-26 13:38 Nathan Owens
2021-06-26 14:09 ` Jared Mauch
2021-06-26 14:24   ` Dave Taht
2021-06-28  1:26     ` Darrell Budic
2021-06-28  1:38       ` Dave Taht [this message]
2021-06-28  2:20       ` Nathan Owens

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='CAA93jw6je-E+70aaW1PdGPTaZLpE=i0nOuSFnt_6e4varpLEsQ@mail.gmail.com' \
    --to=dave.taht@gmail.com \
    --cc=budic@onholyground.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