On Sat, 2019-03-30 at 13:15 -0400, Greg Troxel wrote: > > It's different in that it has known characteristics and a system > design > can take those into account.
Sure, but we aren't there yet. > I suspect if SIP/UDP had retransmission rules adjusted to be "TCP > friendly" in terms of congestion avoidance, it wouldn't work as well > for > you. You are probably right about that. > In situations with congestion avoidance, local greedy behavior > almost always wins, even if everyone doing that leads to collapse. True. > I think there's something bad going on in terms of packet loss and > perhaps otherwise in your case. I'm not sure it's "bad". It's expected packet loss, but only at the phone where the phone simply isn't awake to response. No different than the computer that you just turn-off without shutting down. Or maybe the better analogy is suspend, only to resume later. That's all expected packet loss. Cheers, b.
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Linphone-users mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/linphone-users
