Now that 1 Gb connections to your clients are becoming standard and Virtual Machines (VM) are being deployed for all sorts of good reasons, I believe this excercise is quite helpful.
Many times I ask customers if they think a VM will perform the same as their native PC. The answer is usually 'no way' or 'of course not'.
Then I ask what kind of speed do you think your VM will get compared to their native PC?
When their answer starts with 'it should be' or 'I think it is...' I translate that to mean, 'I DONT KNOW'.
I then ask them if they ever performed a simple 'iperf' test within their native PC and then within their PC VM, or within their servers? The look on their face is a dead give away that they haven't.
I'm not implying or saying you will get the same performance results as I did, I'm just asking you to measure yours.
Here's an example of me using iperf to measure performance with my Virtual machine.