1

What Percentage of NAT Port Translation Schemes are Suitable for NAT Traversal

view story
linux-howto

http://serverfault.com – If I understand correctly, the basic scenario for STUN (RFC5389) is such that a host behind a NAT sends a request to the STUN server on a particular private UDP endpoint (address/port pair, say A:B) and discovers its public UDP endpoint via the corresponding response (say C:D). Using this new information, the host can now expect to send data to any arbitrary public endpoint (on the other side of the NAT) using the same private endpoint as before (A:B) and expect the NAT to translate it to the public source endpoint C:D. Assuming the common case in which the NAT is using only one public addre (HowTos)