On Thu, Oct 13, 2005 at 12:51:41PM -0400, jrandom at i2p.net wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > > Ok, perhaps I'm confused. Do you disagree with the later part - that > > > Freenet/dark will no longer be dark if it grows? Or do you disagree with > > > the conclusion - that not providing the functionality for its users as > > > it grows means it "doesn't scale"? > > > > It may have to provide slightly different functionality for larger > > numbers of users in hostile environments, if and as they become > > increasingly hostile. > > This doesn't make sense - its like saying Tor can provide high latency > comm if it turns into mixminon. Yes, if you build a different system > with different characteristics 2 years down the road, that different > system will be different. The same folks behind Tor are behind > mixminion (at least, to some degree), but that doesn't make them the > same project.
Same routing. Many of the same functions. We can still have requests, inserts, and streams. I don't see the problem. > > > I believe that it will offer means to hide its users. Initially by > > being unharvestable, and then by simple steganography over > > the internet (which can probably be detected by local traffic flow > > analysis), and then by dropping real time delivery and using safer > > steganography (which can probably be detected by semiglobal traffic flow > > analysis), and finally by using non-internet transports. It may well be > > possible for the state to defeat all of these mechanisms, but it will > > cost them significant resources and time, and the system will be highly > > useful in the meantime. And many semi-oppressive states (I would put > > China in this category) won't bother. > > What part of the above couldn't apply to I2P? All of it. I2P is harvestable. Any bored technician can block it. > > > They've only just got around to blocking freenet's session bytes this year! > > Yet they've already started blocking skype. (matter of economics) I didn't know that, interesting... > > > > You've already agreed that both I2P and Freenet/dark offer essentially > > > the same functionality in hostile regimes, including resistance to > > > harvesting [1]. The only difference is that you believe "Freenet/dark > > > will scale better". > > > > > > [1] http://dev.i2p.net/pipermail/i2p/2005-October/000975.html > > > > No, I haven't. Freenet/dark will allow a large darknet inside china with > > a relatively number of external links. I2P/RR will allow a small darknet > > attached to each link. > > Connolly said: > "Thus, I2P 2.0 and Freenet 0.7 both offer the same restricted routing > possibilities to Chinese dissenters." > > To which you replied: > "Essentially yes." Read the "but". It's an important but. > > In the context of that email, there is no reason to constrain the 'B' > subset to any particular size, to limit the 'A' subset to communicating > with just one of those peers, or to limit the 'B' subset's interactions > amongst each other. How exactly is I2P going to route within a large restricted-routes network? My understanding is that it's simply a system of proxies with largely fixed routes. If they're not fixed, they'll be random, and harvestable. If they are fixed, they won't provide much useful internal functionality. Right? > > > The former is preferable for various reasons, the obvious one being > > efficiency. > > Ok, so the essence of the difference is that you believe Freenet will > be more efficient than I2P? That Freenet's routing will have less > hops than I2P's O(1)? That Freenet will require less bandwidth > overhead? You really believe that? As far as I can see your restricted routes system will be a matter of putting as many users behind one proxy as possible. Which in practice will be very few. And there is *no internal routing*. You have on many occasions explained that restricted routes is simply a kludge to get around firewalls etc and it does not route, because I2P's routing does not work unless there is global reachability. So in restricted routing, a node can be contacted through a tunnel from a node on the "outside". Right? Maybe I am completely wrong about I2P's proposed restricted routes mechanism, but everything I have heard says there is no real routing going on, and that it is entirely parasitic. > > > Additionally, freenet/dark can function in the absence of _ANY_ external > > links. > > This is true, without using peers outside the adversary's region of > influence, they'd have to have their own internal 'C' peers, which would > likely be shut down. Though at the scale such a network would run at, > the anonet thing would probably work fine. The what? > > > Freenet/dark is a stego *network*. It just doesn't have any real stego > > *transports* yet. > > If a pony sprouted wings, you'd have a flying pony, too. > > There is no evidence that stego transports will ever happen, or at least, > happen anytime soon. There is neither any practice nor theory suggesting > how they /could/ operate, which leads me to leave it to others researching > in that field, rather than hoping someone will figure it out. I do look > forward to hearing progress on such an effort, as it'd be really quite > kickass to have a workable stego transport. But i'm not holding my breath. It would be totally useless without some sort of darknet routing. Stego transports of the first type are possible today. You simply wrap whatever it is you are sending in SSH, SSL, VOIP, Quake, or whatever else. Stego transports of the second type are also possible. You can for example use irregular VOIP connections. Either you piggyback on real VOIP connections at a lower quality level, or you have it randomly connect in a plausible pattern. Since most people only talk to their friends most of the time, any detection that did not do more-than-local traffic flow analysis would produce many false positives. And you haven't explained why freenet over sneakernet + wifi wouldn't be useful, either. But what I really want to know is how I2P is going to do what Freenet 0.7/Dark will do - you seem to be saying above that it will provide a scalable darknet. Really? How? > > =jr -- 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/20051013/8be815cf/attachment.pgp>
