From: "Richard Scheffenegger" <rscheff@gmx.at>
To: <bloat@lists.bufferbloat.net>
Subject: [Bloat] Goodput fraction w/ AQM vs bufferbloat
Date: Sat, 30 Apr 2011 21:18:51 +0200 [thread overview]
Message-ID: <D02B19AE0CC44AFCBA30F6CD0B10C56C@srichardlxp2> (raw)
In-Reply-To: <4DB70FDA.6000507@mti-systems.com>
I'm curious, has anyone done some simulations to check if the following
qualitative statement holds true, and if, what the quantitative effect is:
With bufferbloat, the TCP congestion control reaction is unduely delayed.
When it finally happens, the tcp stream is likely facing a "burst loss"
event - multiple consecutive packets get dropped. Worse yet, the sender with
the lowest RTT across the bottleneck will likely start to retransmit while
the (tail-drop) queue is still overflowing.
And a lost retransmission means a major setback in bandwidth (except for
Linux with bulk transfers and SACK enabled), as the standard (RFC
documented) behaviour asks for a RTO (1sec nominally, 200-500 ms typically)
to recover such a lost retransmission...
The second part (more important as an incentive to the ISPs actually), how
does the fraction of goodput vs. throughput change, when AQM schemes are
deployed, and TCP CC reacts in a timely manner? Small ISPs have to pay for
their upstream volume, regardless if that is "real" work (goodput) or
unneccessary retransmissions.
When I was at a small cable ISP in switzerland last week, surely enough
bufferbloat was readily observable (17ms -> 220ms after 30 sec of a bulk
transfer), but at first they had the "not our problem" view, until I started
discussing burst loss / retransmissions / goodput vs throughput - with the
latest point being a real commercial incentive to them. (They promised to
check if AQM would be available in the CPE / CMTS, and put latency bounds in
their tenders going forward).
Best regards,
Richard
next prev parent reply other threads:[~2011-04-30 19:23 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-04-26 17:05 [Bloat] Network computing article on bloat Dave Taht
2011-04-26 18:13 ` Dave Hart
2011-04-26 18:17 ` Dave Taht
2011-04-26 18:28 ` dave greenfield
2011-04-26 18:32 ` Wesley Eddy
2011-04-26 19:37 ` Dave Taht
2011-04-26 20:21 ` Wesley Eddy
2011-04-26 20:30 ` Constantine Dovrolis
2011-04-26 21:16 ` Dave Taht
2011-04-27 17:10 ` Bill Sommerfeld
2011-04-27 17:40 ` Wesley Eddy
2011-04-27 7:43 ` Jonathan Morton
2011-04-30 15:56 ` Henrique de Moraes Holschuh
2011-04-30 19:18 ` Richard Scheffenegger [this message]
2011-05-05 16:01 ` [Bloat] Goodput fraction w/ AQM vs bufferbloat Jim Gettys
2011-05-05 16:10 ` Stephen Hemminger
2011-05-05 16:30 ` Jim Gettys
2011-05-05 16:49 ` [Bloat] Burst Loss Neil Davies
2011-05-05 18:34 ` Jim Gettys
2011-05-06 11:40 ` Sam Stickland
2011-05-06 11:53 ` Neil Davies
2011-05-08 12:42 ` Richard Scheffenegger
2011-05-09 18:06 ` Rick Jones
2011-05-11 8:53 ` Richard Scheffenegger
2011-05-11 9:53 ` Eric Dumazet
2011-05-12 14:16 ` [Bloat] Publications Richard Scheffenegger
2011-05-12 16:31 ` [Bloat] Burst Loss Fred Baker
2011-05-12 16:41 ` Rick Jones
2011-05-12 17:11 ` Fred Baker
2011-05-13 5:00 ` Kevin Gross
2011-05-13 14:35 ` Rick Jones
2011-05-13 14:54 ` Dave Taht
2011-05-13 20:03 ` [Bloat] Jumbo frames and LAN buffers (was: RE: Burst Loss) Kevin Gross
2011-05-14 20:48 ` Fred Baker
2011-05-15 18:28 ` Jonathan Morton
2011-05-15 20:49 ` Fred Baker
2011-05-16 0:31 ` Jonathan Morton
2011-05-16 7:51 ` Richard Scheffenegger
2011-05-16 9:49 ` Fred Baker
2011-05-16 11:23 ` [Bloat] Jumbo frames and LAN buffers Jim Gettys
2011-05-16 13:15 ` Kevin Gross
2011-05-16 13:22 ` Jim Gettys
2011-05-16 13:42 ` Kevin Gross
2011-05-16 15:23 ` Jim Gettys
[not found] ` <-854731558634984958@unknownmsgid>
2011-05-16 13:45 ` Dave Taht
2011-05-16 18:36 ` Richard Scheffenegger
2011-05-16 18:11 ` [Bloat] Jumbo frames and LAN buffers (was: RE: Burst Loss) Richard Scheffenegger
2011-05-17 7:49 ` BeckW
2011-05-17 14:16 ` Dave Taht
[not found] ` <-4629065256951087821@unknownmsgid>
2011-05-13 20:21 ` Dave Taht
2011-05-13 22:36 ` Kevin Gross
2011-05-13 22:08 ` [Bloat] Burst Loss david
2011-05-13 19:32 ` Denton Gentry
2011-05-13 20:47 ` Rick Jones
2011-05-06 4:18 ` [Bloat] Goodput fraction w/ AQM vs bufferbloat Fred Baker
2011-05-06 15:14 ` richard
2011-05-06 21:56 ` Fred Baker
2011-05-06 22:10 ` Stephen Hemminger
2011-05-07 16:39 ` Jonathan Morton
2011-05-08 0:15 ` Stephen Hemminger
2011-05-08 3:04 ` Constantine Dovrolis
2011-05-08 13:00 ` Richard Scheffenegger
2011-05-08 12:53 ` Richard Scheffenegger
2011-05-08 12:34 ` Richard Scheffenegger
2011-05-09 3:07 ` Fred Baker
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/bloat.lists.bufferbloat.net/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=D02B19AE0CC44AFCBA30F6CD0B10C56C@srichardlxp2 \
--to=rscheff@gmx.at \
--cc=bloat@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