[Starlink] Bloat generic discussions on Starlink list
David P. Reed
dpreed at deepplum.com
Mon Aug 2 10:57:56 EDT 2021
It appears that Dave Taht and others have expanded this list to an incredibly diverse collection of folks that are interested in all kinds of things (imagined and real) about Starlink by adding cc's etc.
I have no particular issue either way, but...
The technical issues of AQM and the deep understanding of the "lag under load" or "queueing anomalies" in the Internet are somewhat technically difficult to comprehend, as those who have been working on this issue for nearly 50 years now understand.
The diverse nature of people on this list allows "everyone to have the expertise to have an opinion" largely based on incorrect understandings of the underlying network architectural problems that the Internet has successfully addressed, and those that are still ongoing work.
Sadly, the idea of "back pressure from the network to the sender" is now resurfacing as a "solution", rather than local queue management and signalling to the *packet destination* not the source. There are very good reasons why the Internet is architected the way it is, and why congestion control is done at the queue management level in a forwarding node whose input rate can exceed its output rate.
And few, even most wannabe protocol designers, understand why congestion control is a *global* phenomenon, not a per-flow phenomenon, which requires a specific kind of *fairness* in order to prevent livelock or deadlock of the network. (Not the kind of "fairness" that strives to perfectly balance all users with precisely equal service, but what is technically called fairness in the literature on synchronization and scheduling primitives like semaphores, ...).
So - I'd suggest that it is a waste of time here to propose solutions to the "bloat" problem of the Internet here on this list, or to debate the clueless, or even to invent "explanations" (like "backpressure in the pipes") that are profoundly misleading.
Instead, talk about how cool the "phased array" antennas are (even though they aren't phased arrays, just multi-element MIMO antennas). Or speculate on how Starlink might perhaps replace any need for fiber by using inter-satellite mesh routing dynamics that somehow fix the "congestion" problem by scrambling routing every few minutes so the paths are never stable.
I personally think fixing the Starlink-caused internal bufferbloat issues to deliver on Musk's promise of well under 20 msec. end to end packet latency is pretty damn hard, and this free-for-all won't help anyone get there.
Have fun...
On Sunday, August 1, 2021 12:00pm, starlink-request at lists.bufferbloat.net said:
> Send Starlink mailing list submissions to
> starlink at lists.bufferbloat.net
>
> To subscribe or unsubscribe via the World Wide Web, visit
> https://lists.bufferbloat.net/listinfo/starlink
> or, via email, send a message with subject or body 'help' to
> starlink-request at lists.bufferbloat.net
>
> You can reach the person managing the list at
> starlink-owner at lists.bufferbloat.net
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Starlink digest..."
>
>
> Today's Topics:
>
> 1. Re: [Bloat] Of interest: Comcast AQM Paper (Neal Cardwell)
> 2. Re: [Bloat] Of interest: Comcast AQM Paper (Michael Richardson)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 31 Jul 2021 18:55:53 -0400
> From: Neal Cardwell <ncardwell at google.com>
> To: Aaron Wood <woody77 at gmail.com>
> Cc: Simon Barber <simon at superduper.net>,
> "starlink at lists.bufferbloat.net" <starlink at lists.bufferbloat.net>,
> bloat <bloat at lists.bufferbloat.net>
> Subject: Re: [Starlink] [Bloat] Of interest: Comcast AQM Paper
> Message-ID:
> <CADVnQyk+DnCootU9bOZhX9PS2tAqXDc+wZB4djr6rQTVAcMm-Q at mail.gmail.com>
> Content-Type: text/plain; charset="UTF-8"
>
> On Sat, Jul 31, 2021 at 3:27 PM Aaron Wood <woody77 at gmail.com> wrote:
>>
>> If we see that AQM appears to be not functioning as expected
>> for upstream connections on DOCSIS3.1, what's the right avenue
>> for getting that resolved? (and does that only apply to the
>> Comcast-owned, vs. customer-owned, modems?)
>
> FWIW, from the paper it sounds like not all Comcast cable modems
> had/have PIE, which enabled the A/B experiment:
>
> "10. Latency Measurement Results
> As explained earlier, for two variants of XB6 cable modem gateway,
> upstream DOCSIS-PIE AQM was enabled on the CGM4140COM (experiment)
> variant but was not available on the TG3482G (control) variant during
> the measurement period. The TG3482G variant used a buffer control
> configuration that predated AQM in DOCSIS."
>
> neal
>
>>
>> On Sat, Jul 31, 2021 at 10:50 AM Simon Barber <simon at superduper.net> wrote:
>>>
>>> Awesome to hear that you are turning this on both upstream and downstream. Do
>>> you know if the wifi stacks in your home routers also have AQM?
>>>
>>> Simon
>>>
>>>
>>> > On Jul 30, 2021, at 10:28 PM, Livingood, Jason via Bloat
>>> <bloat at lists.bufferbloat.net> wrote:
>>> >
>>> > FYI that I will be presenting a lightning talk at the IRTF MAPRG meeting today
>>> at 17:30 ET
>>> (https://datatracker.ietf.org/meeting/111/materials/agenda-111-maprg). The
>>> talk links to a just-published paper at https://arxiv.org/abs/2107.13968
>>> (click PDF link in upper right of page) that will likely be of interest to
>>> these two lists.
>>> >
>>> > High-level: turning on AQM in the cable modem (upstream queue) took working
>>> latency from around 250 ms to between 15-30 ms, which is actually kind of
>>> cool. ;-) AQM is turned on in all of our CMTSes (downstream queue) and in
>>> DOCSIS 3.1 modems (upstream queue).
>>> >
>>> > Have a nice weekend,
>>> > Jason
>>> >
>>> > _______________________________________________
>>> > Bloat mailing list
>>> > Bloat at lists.bufferbloat.net
>>> > https://lists.bufferbloat.net/listinfo/bloat
>>>
>>> _______________________________________________
>>> Bloat mailing list
>>> Bloat at lists.bufferbloat.net
>>> https://lists.bufferbloat.net/listinfo/bloat
>>
>> _______________________________________________
>> Bloat mailing list
>> Bloat at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/bloat
>
>
> ------------------------------
>
> Message: 2
> Date: Sun, 01 Aug 2021 09:13:25 -0400
> From: Michael Richardson <mcr+ietf at sandelman.ca>
> To: Neal Cardwell <ncardwell at google.com>, Aaron Wood
> <woody77 at gmail.com>, "starlink\@lists.bufferbloat.net"
> <starlink at lists.bufferbloat.net>, bloat <bloat at lists.bufferbloat.net>
> Subject: Re: [Starlink] [Bloat] Of interest: Comcast AQM Paper
> Message-ID: <27719.1627823605 at localhost>
> Content-Type: text/plain; charset="utf-8"
>
>
> Neal Cardwell via Bloat <bloat at lists.bufferbloat.net> wrote:
> > On Sat, Jul 31, 2021 at 3:27 PM Aaron Wood <woody77 at gmail.com> wrote:
> >>
> >> If we see that AQM appears to be not functioning as expected
> >> for upstream connections on DOCSIS3.1, what's the right avenue
> >> for getting that resolved? (and does that only apply to the
> >> Comcast-owned, vs. customer-owned, modems?)
>
> > FWIW, from the paper it sounds like not all Comcast cable modems
> > had/have PIE, which enabled the A/B experiment:
>
> This does bring up an interesting discovery question:
>
> the presences of AQM (whether PIE or FQ_CODEL), and what the settings might
> be (for some things need to be tuned), would be something that might be
> interested to emit via LLDP.
> And/or at a /.well-known URL (accessible via IPv6-LL only).
>
> --
> Michael Richardson <mcr+IETF at sandelman.ca> . o O ( IPv6 IøT consulting )
> Sandelman Software Works Inc, Ottawa and Worldwide
>
>
>
>
> -------------- next part --------------
> A non-text attachment was scrubbed...
> Name: signature.asc
> Type: application/pgp-signature
> Size: 487 bytes
> Desc: not available
> URL:
> <https://lists.bufferbloat.net/pipermail/starlink/attachments/20210801/52089257/attachment-0001.sig>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> Starlink mailing list
> Starlink at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>
>
> ------------------------------
>
> End of Starlink Digest, Vol 5, Issue 1
> **************************************
>
More information about the Starlink
mailing list