On Monday 24 April 2006 21:18, Matthew Toseland wrote: > On Fri, Apr 21, 2006 at 09:43:37PM +0100, Michael Rogers wrote: > Right. We do need ARKs or something similar so that your other peers can > learn your new IP address of course; that is well up the todo list. But if > all your peers are on the same ISP and are NATted and the ISP recycles > your IP every night, you're in trouble. Okay, but this is independent of STUN, isn't it?
> > > UPnP, on the other hand, sounds useful - LimeWire contains a Java > > implementation so that might be a good starting point. > > I believe there are UP&P implementations out there... the question is, > is UP&P widely used and widely available? If so we should certainly > support it. I think a large share of the users are already using UPnP in their home network to easily configure their NATs. However, I do not have a study to give you exact figures. I checked a few applications and at least Gaim, Ekiga, Windows-Messenger, Emulemorph and PacPhone support UPnP (in addition, I think almost all P2P tools support UPnP, e.g. BitTornado and Azureus). I agree, UPnP may be a security risk in a few scenarios. However, I believe it would highly improve the usability of the software because a lot of users experience problems using freenet behind a NAT. What do you think about this approach: STUN determines the type of internet connection used by the user. If the user is behind a NAT she will be asked if she is at her home network or if she uses an untrusted network. If the users tells us that she is at home we use UPnP to configure her router, otherwise (and this will be default after a short timeout) UPnP will not be used. > > > Cheers, > > Michael -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 481 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/tech/attachments/20060425/a83b5cce/attachment.pgp>
