Lethal - Multicasting should be as fast as Unicast yes, but if FOG is sending packets via UDP at 1gpbs, the switch would surely drop a lot of the incoming packets because it can’t send them out of the 100mbps fast enough (and thus lots of retransmissions)? Even with some buffering, whats coming in will be ~10 times faster than whats being sent out and thus it would be SLOWER, due to the constant re-transmissons and dropped packets
Ffor this reason also, I had an idea that a 10 host multicast was the most efficient thing to do and so far its seemed about right - a single host takes 1.5 hours, but 10 take just under an hour to do. My guess was; less retransmissions because of the slower rate to each host.
If one PC was slowing the rest down, that would show in the log (have had this issue on mixed speed rooms). But it seems like they’re all hiccuping from the log
As for the switch things… its tricky because we have no way to access the switches. I work in the academic computing department at a University - normally, all departments are managed via a central ICT department but we have moved away from their systems and recently decided to use FOG. They keep all their switches under lock and key pretty much, so even physical access is limited. However, it seems that the particular switch (note, there are many switches in a stack) has no other hosts on it as far as I can tell. There might be a phone or a printer somewhere but I have no way of knowing. I’ll make a note to ask - communication face-to-face is difficult as they have to manage the infrastructure for the entire University - which involves multiple campuses. So there is a growing list of things to ask!
