similar to: send all outbound traffic through intermediary

On 22/9/20 4:44 pm, Hamish Moffatt wrote: > Is it possible to a configure a tinc (1.0.35) node to only send > outbound through specific nodes, rather than trying to establish > direct connections? > > I have a node which can connect to all the others directly, but some > routes have terrible packet loss, so I'd like to configure it not to try. Anyone? Should I just

On 7/10/20 2:45 pm, Erich Eckner wrote: > Hi, > > On Wed, 7 Oct 2020, Hamish Moffatt wrote: > > > On 22/9/20 4:44 pm, Hamish Moffatt wrote: > >> Is it possible to a configure a tinc (1.0.35) node to only send > outbound through specific nodes, rather than trying to establish > direct connections? > >> > >> I have a node which can connect to all

Hello, I understand that only the destination of a call should do jitter buffering. So, if IAX2/PhoneA calls IAX2/PhoneB through my server (no transfers), PhoneA and PhoneB need to perform their own jitter buffering, and Asterisk will just forward the frames, correct? What happens if the peer does not support jitter buffering, but is close by so there's no need for jitter buffering? My

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Hi, On Wed, 7 Oct 2020, Hamish Moffatt wrote: > On 22/9/20 4:44 pm, Hamish Moffatt wrote: >> Is it possible to a configure a tinc (1.0.35) node to only send outbound >> through specific nodes, rather than trying to establish direct connections? >> >> I have a node which can connect to all the others directly, but some

I'm using rsync, ssh, and cron glued together with Python as a push-based synchronization system. From a single location, I push content out to various offices. I log stdout/stderr on the master server to make sure everything is running smoothly. I would now like for some of our "regional hubs" to take on some of the load (bandwidth-wise), while still retaining my centralized

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, I'm doing some basic operations on large sparse matrices, for example getting a row. it takes close to 30 seconds on a 3Ghz machine, and shots the memory usage up to the sky. I suspect there are dense intermediary steps (which, if true would defeat the purpose of trying to use sparse representaitons). As much as I try understanding the