<p dir="ltr">Doing throughput tests on a live working pipe seems like a *very bad* idea.</p>
<p dir="ltr">[In case it's not apparent, saturating a working pipe is going to cause packet loss (or other bad things) for anything going over the connection at that point in time. Doing so every 5m seems like an even worse idea.</p>
<p dir="ltr">Further, depending on what's using bandwidth, your throughput test will yield results that don't much mirror the actual capacity of the pipe.]</p>
<p dir="ltr">If you're really interested [in spite of my warnings] in doing something like this, you might be able to adapt jperf.</p>
<p dir="ltr">-Greg<br>
</p>