|
Hi, I need help with a 802.11g WLAN test using multiple client
bridges connected to individual interfaces on a quad (100Mb Ethernet) NIC card
running under Linux. It appears that all ports of the quad NIC may not
always keep the clients equally busy (or there’s an internal problem in
the client bridges). This should not be a problem because each interface/bridge
can only carry about 20Mb over the air. Can anyone help by: 1. Pointing out any reports of a similar phenomenon
when using a quad NIC card with iperf (or believed absence of same). 2. Advising how to configure to avoid any differential
packet turnaround delays between interfaces. 3. Suggesting a test strategy to differentiate between quad
NIC, TCP, and 802.11g root causes for unbalanced throughput measurements. 4. Any other help. Thanks to all, P.S. I’ve got some interesting spectrograms of
the over-the-air behavior in this case. If you’re interested in
that sort of thing, ask for and I’ll post a sample. |