On 25/05/13 16:01, Ricardo Lanziano wrote: > On Sat, May 25, 2013 at 9:24 AM, Yussi <[email protected] > <mailto:[email protected]>> wrote: > > > > > There is a C++ port of I2P at http://projects.i2p but it is not fully > > functional yet. > > > > If you want, we can try to set up some tunnels and hook up some > services, > > I have a machine I can spare in I2P. > > > > I rather use I2P than TOR (CJDNS is out of the question). > LOL, is this a matter of pride? There are a lot of good people on cjdns, > and it's not a bad idea. I don't think we should position ourselves > against it, and when possible we should try and work together. > > > No, of course not. My main problem with CJDNS is that is *not* > a secure darknet because it is not anonymous, in contrast with I2P.
But it is very similar to ntk, ntk has no anonymity mechanism behind encryption, and all we need is a tunnel, which we can build over CJD, and should be much faster than I2P. > > I2P is an anonymous mixnet with networking layer *on top*. It > vastly differs from CJDNS. > > In fact, you can't compare CJDNS to I2P. I2P builts unidirectional > tunnels of n lenght that lasts 10 mins or so, and you need to build > a tunnel for both receiving and sending data, that are always changing. > You get a totally different return path for the same connection, ie: IRC, > web, XMPP... I2P provides *anonymity*. I know all this, but in my opinion to say they have nothing in common is dishonest, I2P has much (much) better anonymity than CJD, this is because, all CJD does for anonymity is encryption, that means that Eve can figure our some things about your actions by analysing the encrypted traffic. For example, a get request has a certain size, a post request another, if she know those she can figure out some things. she can also know your next node in line, and so can knock out a large part of the network she knows you're not connecting to. For I2P, you are routing from like 10 tunnels, and you pass around packages which are not intended to you, noise in essence, which make it very difficult to establish what you are doing. > > CJDNS offers you a static path for your packets, since tunnels crosses > for the same machines. That is almost identical to what NTK does. In an ideal world we could implement some sort of mixnet, noise, regular route switching,different routing algorithms based on graph theory etc on top, but at the moment, by design, ntk holds a route and use it, and will only change that route if a node in the middle goes down (or if it finds a better route), this is almost identical to cjd. > > The main idea of Netsukuku is that the network grows itselfs, free > of censorship, control, and single-point of failure. So, yes, I think > it's a bad idea if we want Netsukuku to grow in features and in > research. The main differences i can find between ntk and cjd is that cjd is best described by a tree diagram (Nodes and leafs but in general no cycles), and ntk by a graph diagram(inside groups) and another graph diagram connecting those groups. and even this analogy fails in certain situations depending on the border nodes. > > I am not against CJDNS nor the people that runs in Hyperboria (an > invite-only network iirc? ;) ) but in terms of features, related work > and anonymity and censorship resistant mindset, I rather look at > I2P. Definitely, i2p provides better anonymity than cjd, but cjd provides a better infrastructure for the day the net goes down. > > For example, taking real content authentication for the network: > http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.73.2699 This looks awesome, I'll spend time reading it later. This is all theoretical, what we need is some way to start a network, once we have a few base nodes connected those could do the routing for us, and so i don't think it's massively important how we implement those, and I would go with the fastest rather than most anon method at this stage. I too have some unrealistic dreams about what we can do with ntk, but right now just having a network will be a very good start. suppose we set up 3 ntk nodes over a normal vpn, each one of those could have more nodes connected to it over i2p or tor or wireless or whatever, the only reason i would go with cjd over i2p is that it's a faster way of establishing a tunnel that's not requiring a static ip address. Ideally we just want new users to set up a tunnel to netsukuku.org (and either join the main group, or receive some hints about closer groups) as this is by far the simplest and quickest way of doing this. > > -- > Ricardo Lanziano > To iterate is human, to recurse, divine. > > > _______________________________________________ > Netsukuku mailing list > [email protected] > https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/netsukuku _______________________________________________ Netsukuku mailing list [email protected] https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/netsukuku
