Network Performance Testing

In troubleshooting our backup issue, it seems as though the issue might not be the backup agent, but rather networking related. This particular server that was having problems is virtualized and running in a vSphere cluster. Moving the VM to another host in the cluster improved the backup performance, though perhaps coincidentally. I’ve been running network bandwidth tests to try and confirm or refute this hypothesis. In the process I’m familiarizing myself with iperf and wanted to jot some notes down for myself.

  • iperf operates in a client/server method and therefor needs to be installed on two systems
    • The server will discard traffic “iperf -s”
    • The client with generate traffic “iperf -c server-hostname

The above is handy as I was able to compare/contrast not only bandwidth between various systems, but systems on different hosts, systems on the same host, etc to get a better picture

My tests so far have not backed up my hypothesis, but they did show some interesting figures, some more surprising than others

  • Bandwidth between guests on the same hosts is 14X faster than guests on separate hosts (not so much a surprise as it is eye-opening when actually seeing the numbers)
  • More surprising, performance of client running on guest A to server on guest B != performance of client on guest B to server on guest A. This is what I’m trying to figure out now