New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
"Serverless client" UDP test #194
Comments
This is definitely worth considering. We do this kind of testing at ESnet sometimes as well. |
I still need to have this feature. |
I would also like this - I can help if this is something thats easy to do. We are trying to use iperf to debug a webrtc connection on demand (in c++) and are having to fall back to using iperf2 for this purpose! Having a TCP out of band connection is problematic in this scenario for us! Thanks! |
I need this feature too. |
Was this feature included ? Its quite annoying to have a server when you only want to do a UDP test. |
The client-server architecture of iperf3 makes this impossible without a major change to the design and implementation. At this point we acknowledge its usefulness, but have no plans to do this. If this is needed, it's probably best to try using a different tool such as iperf2. |
Greetings,
Thanks so much for re-igniting iPerf development.
In iPerf 2, we often use the client mode (bound to a specific source IP) to do a one-way UDP throughput test to routers at remote sites. iPerf 2 in UDP client mode will simply bind locally & connect to any open UDP port on the remote router (in our case we use 162 snmp trap) and send a UDP stream to it. By running this test for one minute & checking the router's interface counters, we can determine the bandwidth of the line without requiring any server on-site.
iPerf 3 seems to have (& require) a separate out-of-band client-server TCP control socket. Any option to forgo a successful client server connection & simply bind locally & send UDP packets at a certain rate would be highly appreciated.
Alternatively, I suppose there's always Ostinato that's more suited for this kind of usage...
Thanks very much for your time & best regards,
Blake Willis
Network Architecture Consultant
The text was updated successfully, but these errors were encountered: