bloated vs debloated wireless test (iwllagn <-> wndr5700)

Dave Täht d at taht.net
Thu Feb 3 10:57:55 EST 2011


I did a little testing of my quick n' dirty wireless patch for the ath9k
on what I have.

The two devices are:

a netgear wndr 5700 running openwrt
a laptop with the iwllagn driver

Connected at 36Mbit, going from the (De-bloated)wndr 5700, using iperf
to saturate the link, ping times ranged from ~2ms to ~4ms.

In the opposite (bloated) direction from the lagn, they peaked at about 140ms.

Under contention (10 streams), I experienced about 4% packet loss on ping.
(I've never been so happy to see packet loss in my life)

The ath9k patch for openwrt - while still far from ideal - is here:

https://github.com/dtaht/Cruft/raw/master/bloat/558-ath9k_bufferbloat.patch

My test rig - which uses the babel protocol to make it easier to move
between routers and routed devices - is described (partially) here:

https://www.bufferbloat.net/projects/bloat/wiki/Dtaht_test_rig

-- 
Dave Taht
http://nex-6.taht.net



More information about the Bloat-devel mailing list