[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: Tor speed
-----BEGIN PGP SIGNED MESSAGE-----
> This is why I wrote that Tor should make a new circuit for every top
> TCP-based protocol. So Tor would map TCP connection to one circuit so
> other TCP connections would be independent from underlying
> transmission errors.
> Sorry, I do not know how I would describe this idea better. As is was
> only one simple thought: use already implemented things.
That would be helpful, but not perfect. Ideally, TCP wouldn't be used in
Tor's operations at all; any TCP-based protocol using Tor would do its
own housekeeping. In other words - just as it does in things like
Hamachi and OpenVPN - Tor acts (from the program's point of view) just
like the underlying packet network.
In short, tunneling TCP over TCP is something that should be avoided
when possible, since only one set of TCP sequence numbers should be in
play at any given time, for any given socket.
Owner of Tor node "kitsune"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----