Talk:Simple traversal of UDP over NATs

From Wikipedia, the free encyclopedia

Contents 1 Commercial Implementations 2 Request for diagram changes/explanation 3 This needs a disambiguation page 4 List requested 5 Just want to say thanks 6 Proposed UDP Hole Punching Article Merge

[edit] Commercial Implementations

Is there a possibility to add links to a commercial implementation for a general purpose STUN client and server in the implementations section? (Tsahi Levent-Levi 06:17, 30 March 2007 (UTC))

[edit] Request for diagram changes/explanation

I would like to request the diagram be changed or have further explanation. In particular for the boxes labeled "Request echo from same/different address, same/different port" it would be nice to have an explanation on what this means. Does this mean that endpoint A requests and echo from endpoint B? What is same address, same port and different address, different port mean?

—The preceding unsigned comment was added by Spendleton (talk • contribs) 20:27, 5 February 2007 (UTC).

[edit] This needs a disambiguation page

most users searching for stun are not looking for information about serial tunneling of UDP protocol or whatever... I believe the more common meaning is the dictionary definition. Granted there doesnt need to be an article about stun, but perhaps a link to shock or something. The current page would be of interest to only a techie. see WP:bias 129.64.169.38 07:47, 26 January 2007 (UTC)

STUN can also stand for Serial TUNnel. It's a form of IP tunnelling (like GRE or L2TP), not to be confused with Simple Traversal of UDP over NAT). It appears to be a Cisco-specific protocol.

---

Another vote for a disambiguation page. Information on Cisco's STUN/BSTUN can be found Cisco STUN. This technology also plays in with SDLC. Ahbanks 21:24, 28 January 2007 (UTC)

--

Another vote from me. --Treekids 19:20, 3 July 2007 (UTC)

[edit] List requested

It would be nice to provide example of STUN-friendly implementations or, even better, list what kind of NAT each implementation is. (eg. is linux >= 2.4 a simmetric NAT?)

not had time to fully experiment but i get the feeling linux 2.4.x is somewhere between a restricted port code and a symetric. Afaict it will look like a restricted port cone normally but will start behaving in ways more like a symetric under certain conditions such as two apps behind the NAT using the same local port. Never got round to testing this though its just suspicion based on other things i've noticed about linux as a nat. Plugwash 20:04, 22 September 2005 (UTC)

[edit] Just want to say thanks

Been trying to comprehend how gtalk does p2p, and now it finally is beginning to make sense. Wikipedia rawks.--Waxmop 20:43, 4 February 2006 (UTC)

[edit] Proposed UDP Hole Punching Article Merge

It seemed like the whole article on UDP Hole Punching was a somewhat weak recapitulation of the information in this article. From my perspective, it doesn't make sense to have two articles on the same concept -- even if the name is slightly different, but it may be nice to keep some of the information. So, I'll start off this proposal to merge the UDP Hole Punching article into this article...

Pg8p 10:24, 22 January 2007 (UTC)

- Absolutely not. One is a general technique and the other is a specific algorithm. Or more eplicitly: there exist hole-punching algorithms besides the specific one published as STUN. --Treekids 19:23, 3 July 2007 (UTC)