1

How would one go about objectively benchmarking different x86 server for their potential as x86 routers?

One idea I had is to setup two subnets, 10.0.0.0/24 and 10.0.1.0/24, on opposite sides of a router. Then setup an Iperf client and server on the two subnets. However, this would only test raw throughput, and not how services like NAT and firewall rules act under load.

What is an objective way to test the throughput a router could handle in reality, eg. test NAT, firewall, and thoughput all at the same time.

EDIT: Is there a way to test a routers throughput directly without dealing with the bandwidth limits of the CAT6 cable in between.

Soviero
  • 4,366
  • 8
  • 36
  • 60
  • 1
    I think you're on the right track with iperf. – hookenz Dec 12 '11 at 09:05
  • Without dealing with the cable bandwidth limits? But the traffic generator & receiver has to be plugged to the router. I don't see what you mean... – petrus Dec 12 '11 at 21:32
  • The max bandwidth on Cat6 is 10 Gigabit. What kind of router do you have that you're overrunning the bandwidth for Cat6? – Driftpeasant Dec 12 '11 at 21:35
  • Well, my switch is 100 Mbps. My point was to try and figure out how much throughput the CPU can handle regardless of switch speed. – Soviero Dec 12 '11 at 21:55

1 Answers1

4

iperf would be one way. You could design a bunch of tests and manually and then run iperf under each scenario.

However, have a look here. This is how these guys do it. http://www.smallnetbuilder.com/lanwan/lanwan-howto/26563-how-we-test-hardware-routers-

They use a tool call lxChariot which is meant to simulate real world loads.

hookenz
  • 14,472
  • 23
  • 88
  • 143