lots of folks are starting to explore multichannel, and some of them are coming up with setups that require lots of channels into or out of a location. i'm bumping into this puzzler as i work on setting them up.
more than 8 channels to a hub server with 512 frames/period fails, "waiting for peer" on the client end, "not receiving datagrams (timeout)" on the server end. changing to 256 frames/period yields a max of 18 connections and so forth (see below).
- is this due to a packet-size limitation, and thus normal or something else that just spoofed me into thinking that?
- if my hypothesis is correct, is there a way to calculate this and trap it in an error message, with suggestions on what to do?
- i just did a little brute-force testing that might yield some targets for such an error message:
512 frames/period - 8 outbound connections work
256 frames/period - 18 outbound connections work
128 frames/period - 35 outbound connections work
64 frames/period - 65 outbound connections work
the preceding were all done on an M1 Mac Mini. the 2014 Intel Mac could only manage 63 connections
PS: is there a request in the queue to make the GUI scroll the (VERY cool) VU meters rather than making a window that's so tall that i can't get to the "disconnect" button. ;-)
mike o'c