This is quite interesting... and wrong.

I would add that checking at the beginning of connection you are likely to get a lower value for bandwidth due to window size or congestion algorithms.
The ping has the issue to be not working with proxies.
We should account for bandwidth using all channels.
In the streaming code there are similar computation but one issue is that they don't take into consideration possible packet already queued resulting in reduced bandwidth and higher latency.
There is already a protocol for ping/pong in the channels, we could use that.
Without proxies the estimated RTT can be read using TCP_INFO socket option in Linux (see tcp(7)).

-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/spice/spice/issues/1.