Okay, basic Flent support for irtt implemented in 2503caba7d32f847b242eede225638bd64d88d53 - there's an rrul_be_irtt test that will substitute the netperf udp_rr test for irtt.

However, I don't really want to duplicate all tests for irtt, so I think a better way to integrate it is needed. I.e., a test should specify that it just wants a UDP RTT measurement, and Flent will pick the right tool depending on availability. Something similar might be useful for TCP (someone wanted to substitute iperf for netperf in #115).

Picking a tool depending on what is available locally is straight-forward, but checking for server-side support is trickier. Is it OK to punt on that, or should we fire off a one-shot test (something like `irtt client -d 1ms -wait 1ms <hostname>`) to check?

-- 
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/tohojo/flent/issues/106#issuecomment-344873086