[p2p-hackers] NAT hole-punch keepalive/timeouts

[Alex Pankratov]

David Barrett wrote:

> Ok, sounds like 20 seconds might be an upper limit, then.  Are you using 
> "unconfirmed" keepalives, or bidirectional?  Thanks for the info.
> 

Both peers monitor for the activity and send the keepalive out if
there were none with 15(?) seconds. The handshake protocol however
is such that both peers send some traffic when the tunnel goes up,
so I guess it can count as bidirectional.