Hi Raphael,

> if i understood well your issue, you are using SEMS behind a NAT,  
> and the callers may also be behind another (or same) NAT. Is that  
> right?

Yes. (Though NAT on the client side shouldn't really be a problem  
right now (my own calls will be going through some NAT fixing stuff).)

> The reason why it does not work at the moment is that it never has  
> been a covered scenario.

That makes sense, then :)

> As i understood, it would be enough for your scenario to be able to  
> configure the external IP (i guess you have some port forwarding  
> mechanism) (and maybe the external port also), is that right?

That's my opinion, yes. So far as I can tell from my first look,  
putting the public IP in the Contact header and SDP body should make  
everything work; this should "just" mean splitting the bound listen  
address and the defined "public" address.

There could be some hidden complexity here: e.g., what happens when  
SEMS is listening on two interfaces? Right now I expect it replies  
from the interface that received the message...

> I would like to avoid to implement to much NAT traversal stuff like  
> STUN & Co.,

Agreed and understood! ICE and STUN are best avoided.

> but i think that a possibility to configure the external IP and port  
> would be a pretty easy thing to do.

Sweet.

Is this something you'd rather do? I'm happy to work on it, and I  
promise to try not to break anything, but I can't promise that I'll  
succeed :D

-R
_______________________________________________
Semsdev mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/semsdev

Reply via email to