[Standards] Do we need STUN?

Kevin Smith kevin at kismith.co.uk
Thu Mar 8 20:02:11 UTC 2007


On 8 Mar 2007, at 19:43, Matt Tucker wrote:
> 2) As an alternative to STUN, try to get network address info via XMPP
> first (since you already have a connection to your XMPP server).  
> Imagine
> the following (very simplfied and using fake packet extension values,
> etc):

I don't remember the discussions we had exactly, but there are two  
things niggling away at the back of my mind.

The first is that the address the xmpp server sees may well not be  
the address which the public internet sees, which makes getting the  
ip from the server harder.

The second is that the subsequent p2pconnections are made, I seem to  
recall, through the holes in the NAT (which is the case we're  
typically trying to address) by the stun - and it has to be this way  
from what I remember, as on subsequent attempts the properties of the  
nat may have changed.

It all got terribly confusing at this point, but I remember being  
convinced by what Rob was saying about 'stun is good, mkay'.

/K

-- 
Kevin Smith
Psi XMPP client project leader - http://psi-im.org






More information about the Standards mailing list