Starlink has bufferbloat. Bad.
 help / color / mirror / Atom feed
From: David Lang <david@lang.hm>
To: Michael Richardson <mcr@sandelman.ca>
Cc: "David Fernández" <davidfdzp@gmail.com>, starlink@lists.bufferbloat.net
Subject: Re: [Starlink] Starlink Digest, Vol 25, Issue 28
Date: Thu, 20 Apr 2023 09:20:25 -0700 (PDT)	[thread overview]
Message-ID: <3p3n43p6-633p-7776-02r4-n7r4or6436qq@ynat.uz> (raw)
In-Reply-To: <11217.1682007269@localhost>

[-- Attachment #1: Type: text/plain, Size: 2105 bytes --]

anycast doesn't require that you use BGP, it just means that you have the same 
IP address used in multiple places, and routes that send you to the closest one.

With starlink, they could put an address on every satellite (or every gateway) 
and it would handle the traffic would be handled by the first one that it ran 
across.

David Lang

  On Thu, 20 Apr 2023, Michael Richardson via Starlink wrote:

> David Fernández via Starlink wrote:
>    > As we were discussing recently, maybe starting with anycast DNS
>    > servers on satellites is a first step to consider, before embarking
>    > any other type of cloud servers.
>
> For Starlink customers, having a DNS recursive cache on the satellite you are
> talking to certainly saves half a hop, if the cache is big enough.
> If the gateway part of the bent pipe suffers from bufferbloat,
> then having DNS queries from the client avoid that part seems like a win.
> Also, the recursive queries from the DNS recursive cache can be switched on a
> different circuit to the gateway, or marked better to avoid the bufferbloat.
>
> {or, you know, you could fix the bloat. In the satellite to satellite laser
> circuits, I suspect that the bandwidth available will fluctuate wildly.
> Engineers will be reluctant to throw away delay tolerant packets in that case}
>
> (I'm not sure we should call it an anycast DNS service, because even though
> it looks similiar to the clients to the things we do with BGP, its not really
> the same.  Also, squatting on/intersepting 8.8.8.8, etc. would be a very bad
> thing to do.  It would have to be a new address. But, because we wouldn't be
> using BGP, it does not need to occupy an entire /24. )
>
> --
> ]               Never tell me the odds!                 | ipv6 mesh networks [
> ]   Michael Richardson, Sandelman Software Works        |    IoT architect   [
> ]     mcr@sandelman.ca  http://www.sandelman.ca/        |   ruby on rails    [
>
>
>
>
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink

  reply	other threads:[~2023-04-20 16:20 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <mailman.751.1681965188.1222.starlink@lists.bufferbloat.net>
2023-04-20 12:24 ` David Fernández
2023-04-20 16:14   ` Michael Richardson
2023-04-20 16:20     ` David Lang [this message]
2023-04-20 21:34   ` Eugene Chang
2023-04-20 22:07     ` Spencer Sevilla
2023-04-20 22:11     ` David Lang
2023-04-21  0:03       ` tom

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=3p3n43p6-633p-7776-02r4-n7r4or6436qq@ynat.uz \
    --to=david@lang.hm \
    --cc=davidfdzp@gmail.com \
    --cc=mcr@sandelman.ca \
    --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