On Thursday 11 Aug 2011 06:43:16 Volodya wrote: > You have serious misconseptions about the way freenet operates, i believe. > > 1. You do not connect to the nodes that have the stuff you need, you build a > chain of the nodes (so to speak), and therefore you do not know exactly who > has > that block of furry porn you are looking for, only that one of your friends > (or > stranger opennet peers) has a connection which has a connection... to the > stuff. > Think of the social networks, it's a good model. > > 2. If you will have separate stores the way you describe, then you would have > to > do several things things: 1. You would have to declare the type of content you > are distributing. 2. Know that your node is propagating a specific type of > content. 3. When requesting content you would have to declare that you want > it. > This leads to many issues, for example, if furry porn is illegal in Vatican > city > for everybody but muslims and mormons, and i run a node there and i'm not a > muslim nor a mormon, then when the pope inspects my computer and sees > 'cache-furryporn.bin' i will not have plausible deniability, because i could > have been deleting that file or could have altered the code of freenet to > disallow requests to be passed which are for furryporn audience. Also when one > of my friends gets compromised by l33t cardinal's h at x0r skills then his > node > will record the fact that i have been making a lot of requests for furry porn > lately, and that is just not very nice. > > The point of freenet is (and must be) that nobody, not even you, knows what > information you are passing to your peers. This way if somebody asks you "What > is in your datastore?" you laugh at them because that is something you do not > know. > > - Volodya
Of course, it is possible to collect keys (e.g. by spidering the freesites and watching the forums) and therefore be able to identify that this or that block is part of this or that file. But in general, you can't immediately decrypt or identify the content in your datastore (because you don't have the decryption keys unless you have the freenet URIs e.g. CHK at ...) - at least not without a lot of work. The objective is "plausible deniability": You don't know what is in your datastore, or what you are forwarding. Also, the fact that blocks of data are essentially indistinguishable probably makes some attacks harder. Although in some cases it makes them easier - dividing a big file into many small chunks means there might be more samples for an attacker in some cases. Fixed sizes are more for simplicity... Have a look at our website and/or wiki for more details. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part. URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20110818/422998af/attachment.pgp>
