Many ISPs need the kinds of quality shaping cake can do
 help / color / mirror / Atom feed
From: Dave Taht <dave.taht@gmail.com>
To: libreqos <libreqos@lists.bufferbloat.net>,
	Andrew McGregor <andrewmcgr@gmail.com>
Subject: [LibreQoS] Fwd: an alternate statistics
Date: Mon, 31 Oct 2022 08:00:39 -0700	[thread overview]
Message-ID: <CAA93jw4x2JQmWSPRkkcuro8o_vJvoFr_q4=1HDC2yDzj9=rmXw@mail.gmail.com> (raw)
In-Reply-To: <CAPRuP3kH7GBUy9Zr1cBrduX=5k02Jtiw4L_6cTevqvap15gKLg@mail.gmail.com>

Now, there was a whole book on tracy widom that I gave away because it
was too hard for me and cannot remember the title of, or maybe it
was on something related in control theory...

There's so much activity around interactive vs bulkier traffic
(fq_codel's sparse flow optimization being one, l4s being another),
as when we try to apply gaussian statistics to networking things it
breaks down. There's other misleading indicators we deal with
like "cpu usage" when the real binding factors are interrupt response
time and running out of cache, that go non-linear, suddenly,
but per these bits of math (like the extreme value theorem), have a
describable shape...

Anyway, BBRv0 came out of princeton as best as I remember, and tracy
widom was popular at the time,
https://www.google.com/search?channel=fs&sxsrf=ALiCzsayJH-LIf3QMgoTx0GInGUeNvzOeQ:1667228286598&source=univ&tbm=isch&q=tracy+widom+princeton

---------- Forwarded message ---------
From: Andrew Mcgregor <andrewmcgr@google.com>
Date: Thu, Sep 13, 2018 at 8:18 PM
Subject: Re: an alternate statistics
To: Dave Taht <dave.taht@gmail.com>


That's one such, yes.

Turns out, the extreme value theorem is incredibly handy in networking
too. Basically, it shows that whenever you are dealing with rates of
events exceeding a threshold, there's only a few possible
distributions for the rate and size of those events, and that it
doesn't matter what the bulk of the distribution is, just whether it
has upper and/or lower bounds or not.

On Fri, 14 Sep 2018 at 00:53, Dave Taht <dave.taht@gmail.com> wrote:
>
> https://en.wikipedia.org/wiki/Tracy%E2%80%93Widom_distribution ?
> On Thu, Sep 13, 2018 at 7:48 AM Dave Taht <dave.taht@gmail.com> wrote:
> >
> > ages ago you pointed me at an alternate algebra/statistics thing that
> > I totally spacing on this morning. It looked better than guassian...
> > ???
> >
> > --
> >
> > Dave Täht
> > CEO, TekLibre, LLC
> > http://www.teklibre.com
> > Tel: 1-669-226-2619
>
>
>
> --
>
> Dave Täht
> CEO, TekLibre, LLC
> http://www.teklibre.com
> Tel: 1-669-226-2619



-- 
Andrew McGregor | SRE | andrewmcgr@google.com | +61 4 1071 2221


-- 
This song goes out to all the folk that thought Stadia would work:
https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-6981366665607352320-FXtz
Dave Täht CEO, TekLibre, LLC

           reply	other threads:[~2022-10-31 15:00 UTC|newest]

Thread overview: expand[flat|nested]  mbox.gz  Atom feed
 [parent not found: <CAPRuP3kH7GBUy9Zr1cBrduX=5k02Jtiw4L_6cTevqvap15gKLg@mail.gmail.com>]

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/libreqos.lists.bufferbloat.net/

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

  git send-email \
    --in-reply-to='CAA93jw4x2JQmWSPRkkcuro8o_vJvoFr_q4=1HDC2yDzj9=rmXw@mail.gmail.com' \
    --to=dave.taht@gmail.com \
    --cc=andrewmcgr@gmail.com \
    --cc=libreqos@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