[Battlemesh] Testing performance to multiple end nodes under fq_codel
Dave Taht
dave.taht at gmail.com
Mon Apr 15 18:21:30 CEST 2013
the "rtt_fair" test from netperf-wrappers is suitable for testing to/from
up to 4 endpoint sinks from the test source.
http://snapon.lab.bufferbloat.net/~d/rtt_fair_long_efq_codel_quantum_500.svg
I'll put up the other slides from my too-rapid-talk later today or tomorrow.
Some other links:
BQL: http://lwn.net/Articles/454390/
Codel and fq_codel to some extent:
http://www.bufferbloat.net/projects/codel/wiki
We have a draft on fq_codel intended for lwn due... well... months ago.
The ietf ICCRG reports:
http://www.ietf.org/proceedings/86/slides/slides-86-iccrg-0.pdf
http://www.ietf.org/proceedings/86/slides/slides-86-iccrg-3.pdf
The net result of that was that a standardization effort started in the
ietf, and there will be interesting meetings at the next one, in Berlin, in
late July - (if you've never been to an ietf, there are plenty of other
interesting things that happen there)...
There are only a few billion machines that need an upgrade...
--
Dave Täht
Fixing bufferbloat with cerowrt:
http://www.teklibre.com/cerowrt/subscribe.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ml.ninux.org/pipermail/battlemesh/attachments/20130415/06ee0438/attachment-0001.html>
More information about the Battlemesh
mailing list