Re: [Fwd: I-D ACTION:draft-jennings-midcom-stun-results-00.txt

Here at AOL we have found that the WRT54G is port-restricted cone in its 
normal mode, which is what is documented by Cullen. Our other test 
results are also mostly consistent with the document, with the exception 
of some Linksys models, where the difference may be explained by 
different firmware versions. We were not able to find any symmetric NATs 
in our admittedly small sample.

Cisco PIIX       Port
DLink DI-704P    Cone
Linksys BEFSR41  Addr
Linksys BEFSX41  Addr
Linksys BEFW11S4 Port
Linksys WRT54G   Port
Netgear RP614    Cone
Netgear WGR614   Cone

FWIW, we have found that some Linksys routers will rewrite the IP 
address in the STUN reply to be the nonroutable (e.g. 192.168.x.x) 
address rather than the NAT address, causing the STUN detection to 
classify the connectivity as "Open". For our STUN-like protocol we have 
the server XOR the IP address prior to returning it to prevent such 
rewriting.

Also, we have observed from our customers this approximate breakdown in 
market share: (obviously heavily skewed toward a US population)
Linksys     59%
Netgear     14%
DLink       13%
Belkin       5%
Microsoft    3%
SMC          2%
Siemens      2%
Zyxel        1%
ActionTec    1%
3Com        <1%
Comcast     <1%
Dell        <1%

--justin

Jiri Kuthan wrote on 2/13/2004, 7:08 AM:

 > At 06:17 PM 2/12/2004, Jiri Kuthan wrote:
 > >At 08:57 PM 2/10/2004, Jonathan Rosenberg wrote:
 > >>This is a great piece of work. Thanks to Cullen for putting this
 > together.
 > >
 > >Cullen,
 > >
 > >thanks, that's indeed a very helpful piece of information. In particular
 > >the observation that modern NATs are rarely symmetric makes me feel
 > better.
 > >(The current situation is AFAIK roughly up to 20% of user population is
 > >behind symmetric NATs.)
 > >
 > >Can you send me STUN traces (PCAP at best) for some of the
 > non-deterministic
 > >NATs? Particularly according to some tests we did long time ago,
 > Linksys WRT54G
 > >was symmetric all the time. Unfortunately, I no longer have one
 > around to retest
 > >and I did not record firmware version either.
 >
 > WRT to this particular case, linksys web tells me the box is Linux based
 > (http://www.linksys.com/support/gpl.asp) in which case chances are high
 > the box is symmetric all the times. I agree it would be good to verify
 > your
 > list -- what a pity we just missed an opportunity at sipit.
 >
 > -jiri
 >
 >
 > _______________________________________________
 > midcom mailing list
 > [email protected]
 > https://www1.ietf.org/mailman/listinfo/midcom
 >