From: Andrea Arcangeli <andrea@suse.de>
To: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Larry McVoy <lm@work.bitmover.com>, Larry McVoy <lm@bitmover.com>,
Pascal Schmidt <der.eremit@email.de>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: bandwidth for bkbits.net (good news)
Date: Mon, 1 Sep 2003 18:38:47 +0200 [thread overview]
Message-ID: <20030901163847.GI11503@dualathlon.random> (raw)
In-Reply-To: <1062433710.14254.7.camel@dhcp23.swansea.linux.org.uk>
On Mon, Sep 01, 2003 at 05:28:31PM +0100, Alan Cox wrote:
> On Llu, 2003-09-01 at 17:13, Andrea Arcangeli wrote:
> > > Each ACK that has caused previous delays generally opens up a 64K window
> > > so you get bursts of data incoming. A sequence of acks can cause the
> >
> > the congestion avoidance shouldn't allow what you say. It sends a few
> > packets immediatly (cwnd starts > 1 recently), and that's why
> > non-keepalive connections are bad, but after that the congestion window
> > will remain low if we drop the packets.
>
> You may trigger fast retransmit patterns. Thats why you have to bend the
> window.
fast restransmit are good to trigger, they don't arrive in a flood like
if we had a huge cwnd, if we can send packets out of the network,
they're right to be sent. During congestions the outgoing acks will be
rejected too (it's limiting both ways).
there is no difference between artificial congestion generated by a
shaper and a true congestion. If tcp is correct it has to slowdown
immediatly when it notices congestion.
I also don't see anything specific to a busy http server non support
keepalive in this ack-fast-retransmit matter.
Andrea
next prev parent reply other threads:[~2003-09-01 16:38 UTC|newest]
Thread overview: 71+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <q2SH.AA.3@gated-at.bofh.it>
[not found] ` <qfwI.15D.27@gated-at.bofh.it>
[not found] ` <qgCn.2y4.11@gated-at.bofh.it>
2003-08-30 22:58 ` bandwidth for bkbits.net (good news) Pascal Schmidt
2003-08-30 23:07 ` Larry McVoy
2003-08-31 1:05 ` Pascal Schmidt
2003-08-31 1:39 ` Andrea Arcangeli
2003-08-31 2:18 ` Jeff Sipek
2003-08-31 2:42 ` Andrea Arcangeli
2003-08-31 2:56 ` Larry McVoy
2003-08-31 13:15 ` Alan Cox
2003-08-31 14:45 ` Andrea Arcangeli
2003-08-31 15:31 ` Alan Cox
2003-08-31 15:43 ` Jörn Engel
2003-08-31 15:50 ` Andrea Arcangeli
2003-08-31 17:06 ` Florian Weimer
2003-08-31 21:21 ` Larry McVoy
2003-09-02 7:06 ` Henning P. Schmiedehausen
2003-09-05 8:10 ` Florian Weimer
2003-09-05 15:35 ` Henning Schmiedehausen
2003-09-05 15:50 ` Florian Weimer
2003-09-05 16:10 ` P
2003-09-05 16:43 ` Ricky Beam
2003-09-07 11:55 ` Henning Schmiedehausen
2003-09-07 18:41 ` Ricky Beam
2003-09-08 8:57 ` [COMPLETELY OFF-TOPIC] " Henning Schmiedehausen
2003-09-02 21:52 ` Ricky Beam
2003-09-05 8:16 ` Florian Weimer
2003-08-31 15:44 ` Andrea Arcangeli
2003-08-31 16:22 ` Larry McVoy
2003-08-31 16:33 ` Andrea Arcangeli
2003-08-31 16:48 ` Larry McVoy
2003-08-31 17:06 ` Andrea Arcangeli
2003-08-31 21:18 ` Larry McVoy
2003-08-31 22:49 ` Andrea Arcangeli
2003-08-31 22:52 ` Alan Cox
2003-08-31 22:58 ` Larry McVoy
2003-08-31 23:02 ` Andrea Arcangeli
2003-08-31 23:07 ` Larry McVoy
2003-08-31 23:22 ` Andrea Arcangeli
2003-09-01 0:12 ` Larry McVoy
2003-09-01 0:19 ` Jamie Lokier
2003-09-01 0:33 ` Larry McVoy
2003-09-01 11:43 ` Alan Cox
2003-09-01 16:13 ` Andrea Arcangeli
2003-09-01 16:24 ` Larry McVoy
2003-09-01 16:28 ` Alan Cox
2003-09-01 16:38 ` Andrea Arcangeli [this message]
2003-08-31 23:39 ` Roman Zippel
2003-09-01 0:09 ` Larry McVoy
2003-09-01 0:20 ` Larry McVoy
2003-08-31 22:56 ` Larry McVoy
2003-08-31 23:13 ` Andrea Arcangeli
2003-09-01 0:18 ` Jamie Lokier
2003-09-01 0:25 ` Larry McVoy
2003-09-01 0:28 ` Andrea Arcangeli
2003-09-01 0:50 ` Jamie Lokier
2003-09-01 1:10 ` Andrea Arcangeli
2003-09-01 1:33 ` Larry McVoy
2003-09-02 7:11 ` Henning P. Schmiedehausen
2003-09-02 7:01 ` Henning P. Schmiedehausen
2003-08-31 16:23 ` Larry McVoy
2003-08-31 16:36 ` Andrea Arcangeli
2003-08-31 15:38 ` Jörn Engel
2003-09-02 6:55 ` Henning P. Schmiedehausen
2003-08-31 13:47 ` Pascal Schmidt
2003-08-31 14:51 ` Andrea Arcangeli
2003-09-02 6:53 ` Henning P. Schmiedehausen
2003-08-31 6:30 ` Jörn Engel
[not found] <qn1b.2Pr.29@gated-at.bofh.it>
[not found] ` <qoTh.5mt.11@gated-at.bofh.it>
[not found] ` <rdje.1sH.11@gated-at.bofh.it>
2003-09-02 17:49 ` Pascal Schmidt
2003-08-30 1:29 Larry McVoy
2003-08-30 15:03 ` Larry McVoy
2003-08-30 16:10 ` Jörn Engel
2003-08-31 10:13 ` Toon van der Pas
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
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20030901163847.GI11503@dualathlon.random \
--to=andrea@suse.de \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=der.eremit@email.de \
--cc=linux-kernel@vger.kernel.org \
--cc=lm@bitmover.com \
--cc=lm@work.bitmover.com \
/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