Bidirectional testing

[Craig Artley]

I note that iperf3 lacks the -d (--dualtest) option of the older iperf. This page indicates that is deliberate:
    http://software.es.net/iperf/dev.html?highlight=dualtest

What is the preferred method for testing bidirectional/duplex communication with iperf3?

  -craig

[Bruce A. Mah]

If memory serves me right, Craig Artley wrote:
> I note that iperf3 lacks the -d (--dualtest) option of the older iperf.
> This page indicates that is deliberate:
> http://software.es.net/iperf/dev.html?highlight=dualtest

I'm not sure to what degree this was a deliberate "we're not going to do
this at all" versus a more "this isn't quite as important to us as other
things". (For ESnet, we don't use this functionality at all.)

If someone were to provide an implementation of bidirectional testing,
we'd certainly take a look at it, particularly if it could be
implemented in a way that doesn't have a huge impact on the existing code.

> What is the preferred method for testing bidirectional/duplex
> communication with iperf3?

I guess it depends on what you want to measure / find out; if you just
need to get data flowing in two directions, you could set up a client
and server in each direction? I admit that's a bit of extra work to set up.

Bruce.

PS. This is issue #201 in the iperf3 issue tracker.