> -----Original Message-----
> From: JiangXingFeng [mailto:[EMAIL PROTECTED] 
> Sent: Friday, January 25, 2008 1:17 AM
> To: 'Dan Wing'; 'Bruce Lowekamp'
> Cc: 'P2PSIP Mailing List'
> Subject: RE: [P2PSIP] Choice of STUN peer or TURN peer
> 
> Hi, 
> 
> I'd like to see your proposed method works, because it could 
> make full use
> of decentralized resource in the overlay to serve each other. 

It seems useful to phase this for p2p-sip:  initially, just have
p2p-sip nodes advertise their TURN servers if the p2p-sip node
is not behind a NAT.

Once we know how to have p2p-sip TURN servers qualify their 
NAT's p2p-friendlyness, then can have p2p-sip nodes advertise 
those TURN servers, too.  If we make fast progress on a document
that describes the qualification procedure, and running code
that shows it works, we should be good to go.  No?


> Regards!
> 
> JiangXingFeng
> 
> > 
> > Yep.  That is part of the qualification a p2p-sip TURN server
> > would have to do before it declares itself a TURN server.
> > 
> I'm just worried about that rare end-point independent 
> filtering NAT exists
> according the data in the paper 
> http://saikat.guha.cc/pub/imc05-tcpnat.pdf.
> It said the proportion of the end-point independent filtering 
> NAT is about 5.8%

That paper is for TCP.  I thought the primary concern here, for
TURN, was UDP?

-d


_______________________________________________
P2PSIP mailing list
[EMAIL PROTECTED]
https://www1.ietf.org/mailman/listinfo/p2psip

Reply via email to