Ah, I was under the (wrong) impression that STUN incorporated UDP hole punching[1] support.
Then, let me rephrase my suggestion; would UDP hole-punching be a viable way of approaching this, and has anybody worked with that? /Stefan [1] http://en.wikipedia.org/wiki/UDP_hole_punching > -----Original Message----- > From: [email protected] [mailto:opensim-dev- > [email protected]] On Behalf Of Infinity Linden > Sent: den 15 juli 2009 00:49 > To: [email protected] > Subject: Re: [Opensim-dev] STUN for OpenSim? > > STUN is used by a client to establish it's routable IP address. in > other words, the address a peer would have to connect to in order to > have a chance at crossing the NATting firewall. > > DynDNS can help with this too, but is used to do name resolution. > > they're effectively used for different scenarios. DynDNS and STUN can > both be used by systems which are given dynamic IP addresses from > their ISPs. depending on the type of fabric you want to use, either > might be useful. if you're talking about software that starts up, > registers it's IP address with a service, that's STUN. if you want to > publish a DNS name and have it route to your NATting firewall even if > you're on the wrong end of a dynamic IP address, thats DynDNS. > > for what it's worth, STUN considered to address a subset of the > problem of maintaining stateful application layer sessions across a > NATted firewall. RFC5389 describes the problem in a little more > detail, along with more generic solutions. ( > http://tools.ietf.org/html/rfc5389 ) > > RFC5389 describes a series of techniques for UDP messages to > successfully traverse a firewall that are not found in DynDNS. > Ultimately, a good read of 5389 and practical experience with your > firewall will do you a load of good. > > -cheers > -meadhbh > > On Tue, Jul 14, 2009 at 3:15 PM, Mojito Sorbet<[email protected]> > wrote: > > If you are running OpenSim behind a NAT router, you are going to have > to > > configure the routing in some way, regardless of whether STUN, or > even > > TCP, is used. That is because the connection is originating outside > of > > the firewall, and the whole purpose of firewalls is to not let that > > happen, except in carefully prescribed circumstances. Same thing > > applies to VOIP. > > > > The problem of people not being able to find your IP address has > already > > been solved by services such as DynDNS. > > > > On Tue, 2009-07-14 at 07:57 +0200, Stefan Andersson wrote: > >> One of the main shortcomings of the linden-legacy model is that > >> OpenSim does not work well (as in simple and consistent) from > behind > >> NATs and several home routers. > >> > >> > >> > >> I was thinking, if something like STUN could help us overcome this? > > > >> > > > > _______________________________________________ > > Opensim-dev mailing list > > [email protected] > > https://lists.berlios.de/mailman/listinfo/opensim-dev > > > _______________________________________________ > Opensim-dev mailing list > [email protected] > https://lists.berlios.de/mailman/listinfo/opensim-dev _______________________________________________ Opensim-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/opensim-dev
