Hi!
As far as I understand even the simplest server in Ice is multithreaded. I guess that applies to the simplest (say the "hello world") client as well. Since CPU, thread and process scheduling are OS features, I guess that there is no way to make any real time assumptions about the behaviour of Ice applications, and no way to accurately measure and monitor the time consumption of the different threads running in them.
How close to real time programming can one come with Ice, and which mechanisms/modules should be used (if any

)?
(I have looked at the latency and throughput demos and have seen how you use time constructions there. Is that all there is to "real time"?)
Regards,
Catalin