-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Ian Clarke wrote: > Well, B will be less likely to forward inserts to C because B will have > incurred a debt with C, and nodes should try to reduce their debt - or > at least, spread it evenly among their neighbors.
I'm not sure this would work - wouldn't it result in fewer inserts being forwarded to the most reliable neighbours and more inserts being forwarded to the most unreliable neighbours? In the other reciprocation mechanisms I've come across, you reward a good neighbour by giving it a larger share of your resources, not by reducing your demands on it. Going back to the example, C could obtain a larger share of B's bandwidth by pretending to be highly reliable. > But this wouldn't be a much easier attack than simply setting up a node > that actually had lots of bandwidth? Why not both? However much bandwidth you have, the attack can multiply it. > Remember also that there is a trust relationship between B and C, so > yes, C's neighbors will be hurt by trusting someone that they shouldn't > have trusted, but that is true in-general with Freenet 0.7, if you > trust people improperly, you suffer. Fair point - I'll be happy if we can confine the effects of flooding and black hole attacks to the vicinity of the attacker. Cheers, Michael -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.3 (GNU/Linux) iD8DBQFEdkb9yua14OQlJ3sRAmnxAKDdqdzbNuC4u2F9+Hh+km2ZikZdhgCff9UA L6U0cYpXaQwMn0cMOqrWIpo= =k7MS -----END PGP SIGNATURE-----
