4 ms·

I dont understand how what you propose successfully traverses NAT. Just curious: How do you punch through, and then keep your NAT pipe open? There is no guara

I dont understand how what you propose successfully traverses NAT. Just curious: How do you punch through, and then keep your NAT pipe open? There is no guarantee that this part: "our VPN assigns each user a public, routable, IP address, on the outside world, such as 72.10.x.y.z" will always lead back to a client behind layer(s) of NAT, unless the listening port is manually forwarded in the router itself. Protocols such as STUN have developed methods to attempt to solve this, but its not easy because of issues such as there being no standard for the amount of time that NAT must keep pipes open after they are finished being used.

e1ven 19y ago
That's what openvpn is for. Basically, it deals with the NAT-traversal elements for me, once I've twisted it's arm into doing so.. It works, if I recall correctly, using the UDP trick to keep a port open by sending traffic to a host repeatedly. The nice part about the hack that I'm using is that I don't need to deal with the specifics of the UDP punching, or how often to poll, etc.. I can use industry tested and standard software for that. Think of it this way- Have you ever worked at a large corporation, that has a VPN? Once connected to that VPN, you could access all the servers internally, and they could access you, right? We're basically creating a VPN, and giving everyone an internal IP address.. And then using linux, iptables, and NATing to map all those internal ip addresses to external ones.