Re: Question about test results

[Dan W. <dw...@g4...>]

On Mon, 2004-04-05 at 10:30, Jim Ferguson wrote:
> Dan,
> At first blush, I don't know why you can't get the dualmode test to work in 
> this configuration.  I'm fairly certain we've never tested Iperf using a 
> crossover cable, but I don't see why that should be different.
> 
> I'm certain we can think of some excuses for non-symmetrical results you're 
> seeing, but you may eliminate several of them if you have more info about 
> you end workstations besides the fact they are both running FreeBSD 
> 4.9.  Are they identical in their hardware configuration?  How much 
> different are the two directions?

Jim,

The hardware is identical with identical configurations. The difference
is quite extreme, almost as if the links were running in half duplex but
I know it is running full duplex.

A theory I have is that one side starts transmitting and fills the
outbound TCP/IP queue so that it takes a significant amount of time for
the TCP/IP connection establishment for the second connection to
complete.  This would explain why the UDP test shows a very nice
symmetrical result but the TCP/IP test does not.

I've thought about modifying the code so that the transmit test delays
for a few seconds before starting but I think I may still run into the
same issue with ACK packets getting stuck behind so many transmitted
packets.

Dan

> -Jim Ferguson
> 
>   At 05:21 PM 4/2/2004 -0500, you wrote:
> >Hello,
> >
> >   I'm a new user of Iperf and a customer of mine is using Iperf to try
> >and validate their available bandwidth.
> >
> >Using command lines of:
> >
> >   server: iperf -s
> >   client: iperf -c #.#.#.# -d
> >
> >we see results that are not symmetrical. If we use UDP:
> >
> >   server: iperf -s -u
> >   client: iperf -c #.#.#.# -d -u -b 1500k
> >
> >we get symmetrical results. These tests were run on the equivalent of a
> >point-to-point T1 data circuit.
> >
> >If we connect two computers together (FreeBSD 4.9) via a switch, we see
> >similar results. Oddly, when the two FreeBSD systems are connected
> >together with a crossover cable we are unable to get the dualmode test
> >to run. It always run in tradeoff mode. We've yet to dig into the source
> >code to see what would trigger this condition. We built our executables
> >from the source that was available at:
> >
> >   http://dast.nlanr.net/Projects/Iperf/
> >
> >My question is whether this is normal. Should we see a non-symmetrical
> >test result using TCP when trying to perform a test that would use all
> >available bandwidth?
> >
> >Dan
> >--
> >Dan Watts
> >G4 Communications
> >One Sundial Avenue
> >Manchester, NH 03103
> >(603) 296-4422 voice direct
> >(603) 625-0555 voice
> >(603) 647-7576 fax
> >www.G4communications.com
> >
> >"High-Speed Intelligence... Business Internet & Voice Solutions."
> 
> J. W. Ferguson (fer...@nl...)
> Distributed Applications Support Team
> National Laboratory for Applied Network Research (NLANR)
> NCSA - University of Illinois
> Ph. (217) 244-1974
-- 
Dan Watts
G4 Communications
One Sundial Avenue
Manchester, NH 03103
(603) 296-4422 voice direct
(603) 625-0555 voice
(603) 647-7576 fax
www.G4communications.com

"High-Speed Intelligence... Business Internet & Voice Solutions."