-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > > No[1], and I doubt we will [2]. We use existing tunnels for > > building new ones. Turtles, all the way down.
> But if each node in C only knows one node in B then how does a > node in C build its first tunnel? Without telescopic tunnel > construction the initiator has to contact every node in the > tunnel directly, doesn't it? No [1]. "After building all of the request messages, they are garlic wrapped for the target router and sent out an exploratory tunnel" [1] http://dev.i2p.net/cgi-bin/cvsweb.cgi/i2p/router/doc/ tunnel-alt.html?rev=HEAD#tunnel.request The outbound exploratory tunnel's endpoint is the one who talks to peers that are being asked to participate in the tunnel, and that peer answers to the inbound exploratory's gateway. The participant (or peer rejecting participation) can be any peer, and the message between it and the endpoint/gateway may be direct, or tunnel routed itself. Telescoping tunnel construction, ala Tor, sends a message from Alice to Bob, then Alice sends a message through Bob to Charlie, then Alice sends a message through Bob and Charlie to Dave. We send a message from Alice through Zack, Yan, and Xena targetting Bob, and Bob replies through Wade (who forwards it to Veronica, who forwards it to Uma, who forwards it to Alice). This prevents trivial tunnel path analysis by a peer inside the tunnel. (The bootstrap scenario uses 0 hop tunnels, then 1 hop, etc) =jr -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFDT4ZXWYfZ3rPnHH0RAlFoAJ94/7Tz3S661SoaQeYLrGTOdv12OACdGrgT XEORZ+FdsDCM9p3AhbiF0p0= =mmJh -----END PGP SIGNATURE-----
