netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Stephen Hemminger <shemminger@linux-foundation.org>
To: Lachlan Andrew <lachlan.andrew@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>,
	Herbert Xu <herbert@gondor.apana.org.au>,
	shaoliu@princeton.edu, Douglas Leith <doug.leith@nuim.ie>,
	Robert Shorten <robert.shorten@nuim.ie>,
	netdev@vger.kernel.org
Subject: Re: [PATCH] tcp-illinois: incorrect beta usage
Date: Wed, 28 Nov 2007 16:43:31 -0800	[thread overview]
Message-ID: <474E0B33.2060108@linux-foundation.org> (raw)
In-Reply-To: <aa7d2c6d0711281625m44bf3ca3me188a7041d3421da@mail.gmail.com>

Lachlan Andrew wrote:
> Thanks Stephen.
>
> A related problem (largely due to the published algorithm itself) is
> that Illinois is very aggressive when it over-estimates the maximum
> RTT.
>
> At high load (say 200Mbps and 200ms RTT), a backlog of packets builds
> up just after a loss, causing the RTT estimate to become large.  This
> makes Illinois think that *all* losses are due to corruption not
> congestion, and so only back off by 1/8 instead of 1/2.
>
> I can't think how to fix this except by better RTT estimation, or
> changes to Illinois itself.  Currently, I ignore RTT measurements when
>    sacked_out != 0    and have a heuristic "RTT aging" mechanism, but
> that's pretty ugly.
>
> Cheers,
> Lachlan
>
>   
Ageing the RTT estimates needs to be done anyway.
Maybe something can be reused from H-TCP. The two are closely related.


  reply	other threads:[~2007-11-29  0:44 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <aa7d2c6d0711261023m3d2dd850o76a8f44aef022f39@mail.gmail.com>
     [not found] ` <001001c83063$9adbc9d0$d5897e82@csp.uiuc.edu>
2007-11-28 23:47   ` [PATCH] tcp-illinois: incorrect beta usage Stephen Hemminger
2007-11-29  0:25     ` Lachlan Andrew
2007-11-29  0:43       ` Stephen Hemminger [this message]
2007-11-29  5:26         ` Shao Liu
2007-12-03 22:52           ` [RFC] TCP illinois max rtt aging Stephen Hemminger
2007-12-03 23:06             ` Lachlan Andrew
2007-12-03 23:59               ` Shao Liu
2007-12-04  0:32                 ` Stephen Hemminger
2007-12-04  1:23                   ` Lachlan Andrew
2007-12-04  8:37                     ` Ilpo Järvinen
2007-12-07  3:27                       ` Lachlan Andrew
2007-12-07 11:05                         ` Ilpo Järvinen
2007-12-07 12:41                           ` David Miller
2007-12-07 13:05                             ` Ilpo Järvinen
2007-12-07 18:27                               ` Ilpo Järvinen
2007-12-08  1:32                               ` David Miller
2007-12-11 11:59                                 ` [RFC PATCH net-2.6.25 uncompilable] [TCP]: Avoid breaking GSOed skbs when SACKed one-by-one (Was: Re: [RFC] TCP illinois max rtt aging) Ilpo Järvinen
2007-12-11 12:32                                   ` [RFC PATCH net-2.6.25 uncompilable] [TCP]: Avoid breaking GSOed skbs when SACKed one-by-one David Miller
2007-12-12  0:14                                     ` Lachlan Andrew
2007-12-12 15:11                                       ` David Miller
2007-12-12 23:35                                         ` Lachlan Andrew
2007-12-12 23:38                                           ` David Miller
2007-12-13  0:00                                           ` Stephen Hemminger
2007-12-15  9:51                                     ` SACK scoreboard (Was: Re: [RFC PATCH net-2.6.25 uncompilable] [TCP]: Avoid breaking GSOed skbs when SACKed one-by-one) Ilpo Järvinen
2008-01-08  7:36                                       ` SACK scoreboard David Miller
2008-01-08 12:12                                         ` Ilpo Järvinen
2008-01-09  7:58                                           ` David Miller
2008-01-08 16:51                                         ` John Heffner
2008-01-08 22:44                                           ` David Miller
2008-01-09  1:34                                             ` Lachlan Andrew
2008-01-09  6:35                                               ` David Miller
2008-01-09  2:25                                             ` Andi Kleen
2008-01-09  4:27                                               ` John Heffner
2008-01-09  6:41                                                 ` David Miller
2008-01-09 14:56                                                   ` John Heffner
2008-01-09 18:14                                                     ` SANGTAE HA
2008-01-09 18:23                                                       ` John Heffner
2008-01-09 12:55                                                 ` Ilpo Järvinen
2008-01-09  6:39                                               ` David Miller
2008-01-09  7:03                                                 ` Andi Kleen
2008-01-09  7:16                                                   ` David Miller
2008-01-09  9:47                                                   ` Evgeniy Polyakov
2008-01-09 14:02                                                     ` Andi Kleen
2007-11-29 14:12     ` [PATCH] tcp-illinois: incorrect beta usage Herbert Xu

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=474E0B33.2060108@linux-foundation.org \
    --to=shemminger@linux-foundation.org \
    --cc=davem@davemloft.net \
    --cc=doug.leith@nuim.ie \
    --cc=herbert@gondor.apana.org.au \
    --cc=lachlan.andrew@gmail.com \
    --cc=netdev@vger.kernel.org \
    --cc=robert.shorten@nuim.ie \
    --cc=shaoliu@princeton.edu \
    /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;
as well as URLs for NNTP newsgroup(s).