wireguard-tools/contrib/nat-hole-punching
Jason A. Donenfeld cef7ac9ef9 global: put SPDX identifier on its own line
The kernel has very specific rules correlating file type with comment
type, and also SPDX identifiers can't be merged with other comments.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
2018-09-20 19:41:22 +02:00
..
nat-punch-client.c global: put SPDX identifier on its own line 2018-09-20 19:41:22 +02:00
nat-punch-server.c global: put SPDX identifier on its own line 2018-09-20 19:41:22 +02:00
README examples: add nat-hole-punching 2016-08-24 15:47:31 +02:00

== NAT Hole Punching Example ==

This code should never be used, ever. But, it's a nice demonstration of how
to punch holes and have two NAT'd peers talk to each other.

Compile with:
    $ gcc nat-punch-client.c -o client -lresolv
    $ gcc nat-punch-server.c -o server


Server is 1.2.3.4 and is on the public internet accepting UDP:49918.
Client A is NAT'd and doesnt't know its IP address.
Client B is NAT'd and doesnt't know its IP address.


Server runs:
   $ ./server

Client A runs:
   # ip link add wg0 type wireguard
   # ip addr add 10.200.200.1 peer 10.200.200.2 dev wg0
   # wg set wg0 private-key ... peer ... allowed-ips 10.200.200.2/32
   # ./client 1.2.3.4 wg0
   # ping 10.200.200.2

Client B runs:
   # ip link add wg0 type wireguard
   # ip addr add 10.200.200.2 peer 10.200.200.1 dev wg0
   # wg set wg0 private-key ... peer ... allowed-ips 10.200.200.1/32
   # ./client 1.2.3.4 wg0
   # ping 10.200.200.1

And voila! Client A and Client B can speak from behind NAT.



-----
Keep in mind that this is proof-of-concept example code. It is not code that
should be used in production, ever. It is woefully insecure, and is unsuitable
for any real usage. With that said, this is useful as a learning example of
how NAT hole punching might work within a more developed solution.