On Thursday 12 June 2008 23:45, Michael Rogers wrote: > Matthew Toseland wrote: > > The information is available now from swapping, it's a question of making it > > more reliable, up to date and accurate. > > Right, I'm just pointing out that attackers like reliable, up to date, > accurate information too. :-) > > > Long-term intersection attack: Correlate the time at which inserts, requests > > etc which can be tied to a specific identity happen with the time at which a > > specific node is known to be online or offline. Generally this is much less > > of a threat to Freenet than to a real-time mixnet. > > Depends on the application - it could be used to identify Frost/FMS > posters, for example.
Right. A general solution is to introduce arbitrary delays at each hop during the encrypted tunnel; this is what mixmaster remailers do. Of course if you want your post to arrive quickly this doesn't work. And we'd have to figure out a way to build tunnels anyway! > > > The mechanism we are > > describing would make it easier but only when attacking your peers' peers. > > My peers' peers are about 400 nodes (more if I raise my connection limit). Significantly less than that surely; small world has significant overlap between my peers and my peer's peers for any given peer, that's a fairly basic property without which routing won't work well. Small world networks have *both* fast mixing and clustering... > > Cheers, > Michael -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080612/18b429d2/attachment.pgp>
