Re: Web-of-trust questions

On Fri, May 19, 2006 at 10:11:57PM +0200, Helge Preuss wrote:
> >
> >> You're somewhat right about the trouble of getting connected,
> >> while the net is small it'll be hard but as it grow chances are
> >> better that you know someone who also use it.
> Ah, but that's a fundamental problem: As long as you're small, it's
> hard to grow. If you're large, growing is easier (until you reach a
> saturation point, I guess).
> Still, even if freenet grows, there will be many people who don't know
> anyone with access to it (there are still many people around without a
> GMail invite, too. And I dare to predict that freenet never will
> achieve Google's market penetration).
> At least, there should be some central servers to get newbies started.

There is, in effect. Most people get connections from #freenet-refs on
irc. :( But the hope is that the network will grow organically once we
have a bootstrapping core.

> >> About port scanning you're wrong, freenet use random ports.
> >
> >> And it is planned that it will be possible to use stegonography
> >> later so the trafic would look like a game, VoIP or video
> >> streaming so it'd be harder to automatically block it.

> freenet may use random ports, but there still is a protocol behind it
> which can be detected.

> Using steganography is a nice idea and I'm sure it can protect against
> traffic analysis, but I don't see how it can protect against a
> connection request. Will freenet only accept connections from trusted
> IPs? But then, what about dynamically assigned IPs?

Protecting from a connection attempt is actually very easy with UDP. It
is not possible to get a Freenet node to say *anything* if you don't
have its node reference already. At present, it also has to have your
node reference for connection setup to start; with the eventual opennet
version (yes there will be an opennet version; opennet ~= freenet 0.5;
peers are discovered automatically once you're on the network), you will
only need its noderef. It's harder to do this on TCP, but still possible
if we proxy a legitimate TCP service such as a web server.

It may well be possible to detect freenet traffic at a router level, but
this is not the same thing as portscanning; it is FAR more expensive.
And at that point we can indeed have stego. And no, it can't perfectly
protect against traffic analysis. But we can make a start, and make
things difficult for our adversary.
-- 
Matthew J Toseland - toad-EI5O+8PHWbJeeLb3ft/[email protected]
Freenet Project Official Codemonkey - http://freenetproject.org/
ICTHUS - Nothing is impossible. Our Boss says so.

_______________________________________________
chat mailing list
[email protected]
Archived: http://news.gmane.org/gmane.network.freenet.general
Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/chat
Or mailto:[email protected]?subject=unsubscribe

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFEbkEwHzsuOmVUoi0RAlKKAJ0WjKbOZDM4DU0ieHkhdAUgKPPUeACdFFkh
Qh9JYUFQdM8cZE46OM9SPwY=
=TxI9
-----END PGP SIGNATURE-----