On Wednesday 20 February 2008 14:17, Michael Rogers wrote: > On Feb 7 2008, Matthew Toseland wrote: > >> Perhaps a node should only attempt to reduce its > >> own stress (in Robert's terms), rather than the product of its stress > >> and the other node's stress? > > > >That's an interesting idea. Wouldn't it result in much slower network > >evolution? Simulate it, if it looks promising we should ask Oskar. > > The results seem pretty similar to the current method, at least in terms of > minimising the distance between neighbours (I haven't tested routing yet).
The Pitch Black paper's attacker source code is available here: http://crisp.cs.du.edu/pitchblack/code/source/freenet/src/freenet/node/LocationManager.java (Hello to any bad guys listening in!) AFAICS the attack is as follows: - Bad nodes conspire to divide up the keyspace amongst them - if there are 3 bad nodes, the locations would be 1/6, 1/2, 5/6 ((i+0.5)/numBad). - Their locations vary very slightly around these target locations. - To do a bad swap: - Pick a random peer. - If its location is very close to 0.25 or 0.75. return. (Huh?) - Send a swap request with the same number of friends as we really have. But set each friend's location to be very close to the random node's location. - The rest of the swap attempt proceeds normally. It doesn't seem an exceptionally effective attack: It's guaranteed to succeed if it reaches the chosen random peer node, but that's not likely, and if that was the idea why use HTL 6? In the paper they suggest to take up the target location m, then try to force a swap, by either sending locations close to the target (impossible because of our commit/reveal protocol, unless you're targeting a specific node), or close to m+0.5 modulo 1.0. As far as I can see your proposal would help against such an attack. > > 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/20080221/793f4551/attachment.pgp>
