How do I test the increased bandwidth after setting up LACP?

13,189

Solution 1

  1. Failover: Unplug each cable and ensure traffic continues to flow between the switches.

  2. Bandwidth: iperf seems to be the standard took for testing these kinds of things. If you can get a setup like this:

    A ----- |--------|===L1===|---------| ------ C
            | Switch |        | Switch2 | 
    B ----- |________|===L2===|_________|
    

    Then setup Iperf on all 3 boxes and try various client/server setups or you can also just SCP over some files if your on linux. Simultaneously transfer files form A to C and B to C. Try unplugging L1 or L2, or L1 and then L2 and observe if the bandwidth drops. The reason you need to do 2 simultaneous transfers is becasue each transfer(TCP) can only use one line(L1 or L2 in this case). Depending on your switches, you might be able to transfer between 2 boxes(for example A and C) if the transfers are on different ports, but thats only newer switches, most hash based on src and dst IP addresses.

Solution 2

You can transfer a file and watch it go faster... but it might not. Some cheap LACP implementations do per-MAC address hashing, so that any 1 transaction between any 2 stations is only on one actual link to prevent out of order packets. Generally speaking, if all sides say that the LACP is active, it is active.

Solution 3

Aside from actual throughput testing you ought to be able to clear then view traffic stats on the interfaces themselves (not sure with netgear) or monitor via SNMP with a network management tool to verify utilization on the link.

Share:
13,189
dannymcc
Author by

dannymcc

Updated on September 18, 2022

Comments

  • dannymcc
    dannymcc almost 2 years

    I have setup LACP, and LAG on our Netgear Switches (GS724T). I have bound(?) two ports together to create one 2Gbps connection between two of the switches.

    I was wondering if there is a way of testing the increased bandwidth? I know it's setup because the switch no longer creates a routing loop, and the switch management (on both switches) show the two ports as members of the LAG.

    Obviously if I ping the two switch IP's they succeed at the same speed as before because its bandwidth not transfer that I am trying to view.