It's a question of WHAT could we usefully share. We certainly don't want to build Freenet 0.7 on top of I2P as it is, whole. Because I2P requires a harvestable DHT. Now, perhaps we could do something with restricted routes; I don't know, I need a little more information to make that assessmnet. Also, I don't agree with the general assumption that the best way to do distributed content is to run a classic DHT on top of I2P, with nodes anonymized. That is one option, it has certain advantages, but there are others, with other advantages i.e. freenet.
The *other* major difference with I2P is that I2P has pretty much any to any connectivity - nodes in A will connect to other nodes in A, constantly. You don't really have a limited routing table. Well I suppose you have a list of nodes which are preferred in new tunnels, onions etc. I am skeptical that this is safe as far as traffic analysis goes. However you tell me there's a lot of published work on it, so maybe it is workable in an opennet environment. Certainly we have plenty of issues with premix routing etc! As I've explained, we won't do premix routing in 0.7 because of difficulty... Perhaps we should meet up on IRC? I'm free for the next few hours... On Fri, Oct 14, 2005 at 12:23:21PM -0400, jrandom at i2p.net wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > > We have had NAT traversal for a while now. It was described in > > > the SSU specs [1] that I sent to toad for comment sometime in > > > the early summer, and deployed in 0.6.0.6 [2]. > > > Well, I was basing this on: http://www.i2p.net/faq#ports, which says: > [snip] > > Is that information out of date? > > Yes - as you can see, it also doesn't mention the UDP transport. I > haven't had time to work on the website, it'll be revamped later > this month though, I hope. > > > I am sure we could all inform ourselves better about many things > > given sufficient time. > > True indeed, but there aren't that many things going on in the > anonymity field - perhaps a dozen public projects past the "vague > idea" stage? You've also shown that you don't follow the DHT > field either, as we discussed some pretty substantial theoretical > research last time, none of which you were familiar with. What > efforts in the anonymity field do you follow? > > > You are right, we didn't invent I2P here, because I2P doesn't meet > > our requirements, nor does anything that relies on a DHT for its > > operation. > > I thought relying upon a DHT was something redundant, in your eyes, > as "Freenet is like a DHT", not something that failed to meet your > requirements. > > Or are your requirements novelty? > > > >I don't care about technology, I care about results. What will > > >help real live people. > > > > Just not if they live in China, Saudi Arabia, Iran, or maybe even > > the UK several years from now... > > Of course I care about them, but I'm only one person. I've written > 95% of the I2P SDK and router, and probably a third of the LOC in > the client apps, and I'm overwhelmed just doing what I do. I can't > do everything for everybody all at once. > > If Freenet were to run on top of I2P, the reuse made possible would > free up substantial resources (aka hours) across both teams, > allowing further improvement of the capacity to help those in > hostile regimes. As I can't do it myself, I don't promote it > for such a use. > > > you have provided no evidence that your approach is better suited > > to meeting the requirements we share. > > What better evidence can I offer beyond working code? > > You haven't even bothered to see what you're talking about. This > is why I dismiss any notions that you've "seriously considered" > I2P, as you've never actually looked into it. Have you actually > run any of the I2P releases put out in the last 2 years? > > This, in turn, suggests to me that neither have you "seriously > considered" running over Tor. Last I heard, the complaints being > stated about Tor were that "its centralized", without any regard > to the work that the Tor folks are making on decentralization. > > "Serious consideration" means doing some work. We've discussed in > this thread how I2P can offer the same level of operation that > Freenet/dark can, unless you're either disconnected from the > Internet or any notion of a free press is gone (as without the > ability to communicate anonymously, there is no free press). > > The Freenet/0.7 algorithm is neat, but it also isn't too different > from the old CPA (only now there's a theoretical basis for it, > instead of a purely heuristic basis). Not to diminish Oskar's and > your work on that research, from a practical point of view, it > doesn't change Freenet/light too much (and Freenet/dark is just > a different way of distributing references). > > =jr > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.1 (GNU/Linux) > > iD8DBQFDT9lMWYfZ3rPnHH0RAgaaAJ9Ux0IpPf6xFEuShQGDzOBsVv6X2QCfcC22 > 8JnB2JCCvHb3lmPD4D/64Ro= > =Sor5 > -----END PGP SIGNATURE----- > _______________________________________________ > Tech mailing list > Tech at freenetproject.org > http://emu.freenetproject.org/cgi-bin/mailman/listinfo/tech -- Matthew J Toseland - toad at amphibian.dyndns.org Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/tech/attachments/20051014/fb028858/attachment.pgp>
