Changes

The In this test bed the two interfaces are connected on the same Host host machine (a PC running linux) via a gigabit Gigabit switchbut are associated to two different subnets:* ETH0: 192.168.0.xxx* ETH1: 192.168.12.xxx

The two Ethernet interfaces are associated on two differnet subnets:* ETH0 : 192.168.0.xxx* ETH1 : 192.168.12.xxx On the Linux Host linux host machine run <code>iperf </code> application in server mode, with pretty the same performanceTBD???:

Here is are the results of the test on the two Ethernet interfaces launched sequentially:

Here is the Iperf test executed In case <code>iperf</code> clients are run simultaneously on the two Ethernet interfaces. Here the bandwidth on target, each interface instance's bandwitdh is half the bandwidth roughly halved because on the host side there is a single physical Ethernet interface with two different subnets associated.

On another test setup the In this case ETH0 port is connected as before to a Linux Host machine via Gigabit switch. The ETH1 is point-to-point on connected to a second Linux Host host machineinstead.

In this case we can This configuration allows to achieve better performance on the single ETH1 Iperf <code>iperf</code> test of 780-820 Mb/s . But in the case of simultaneous Iperf test <code>iperf</code> instances on ETH0 the bandwidth drops down to 330-350 Mb/s.TBD ???