[p2p-hackers] UDP Hole Punching through Symmetric NATs

[Saikat Guha]

On Wed, 2005-06-15 at 15:28 -0700, David Barrett wrote:
> Regarding the data, I see it 
> dtail TCP port prediction; do those results closely correlate with UDP 
> prediction?

I do not know for sure, but I believe they will be similar. Just as
competing HTTP traffic etc contents with port pred. for TCP, DNS traffic
will contend with UDP. At the same time more NATs are cone for UDP so
the overall picture may be more rosy, but unfortunately, I don't have
the numbers to substantiate that. Alex mentioned stellar success rates
so perhaps the situation is much more rosy that I anticipate. 

> And I think the relevant column is "TCP Binding/Delta", correct?  If so, 
> it appears that only two of the tested NATs used random port 
> assignments, which is encouraging.

Correct. And of those 2, one is an old lab box chosen for testing
because of that particular property and the other was submitted by an
anonymous user who didn't leave any NAT or contact information.

-- 
Saikat
(who thinks fuzzy words like 'rosy' are great for use in technical
discussions)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://zgp.org/pipermail/p2p-hackers/attachments/20050615/2f82891f/attachment.pgp