-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 17/07/12 17:15, Tony Arcieri wrote: > To actually trick a collaborative filtering algorithm that's > looking for self-similarity, a "Sybil" (I'm not even sue if that > word makes sense in this context) would need to participate in the > network as a good-faith peer.
I'm not sure that's true in the general case. For example, if the collaborative filtering algorithm looks for "files rated highly by peers that give similar ratings to mine" then the attacker can create one Sybil for each legitimate peer that echoes the peer's ratings and adds a high rating for virus.exe. The Sybils don't have to do any useful work, just retrieve and publish ratings. Each legitimate peer will find that its perfect "taste buddy" gives a high rating to virus.exe. Does the Cryptosphere prevent that sort of attack? Cheers, Michael -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iQEcBAEBAgAGBQJQBZxzAAoJEBEET9GfxSfMo1kH/1HK52AmPG7EEM094e1jEd2C 6N6ymr2eSv73DLQJ958nIT8Al2qVZbJR/JoNqbb7a7VFxAdqsqa6Wou32wE/Pwg6 3cRJyiiaB3ExBxBVVeeKn+4ayihb6WdDdIQ5kOg/dplanmDvQ4+2WnLjaGJpMzhB o8mTYvHhG8xMMI+yAFvkgniqUwoL+BE+EWUBARWAY/vyHjXOaKqAOny3C6grdMo4 sgiTMfdZn+vV3EUVhMyAwJtmn0G7IFLXg4RRmeUl79L8rIpOZXAuj6DD2p+5J1uk jPIcmkWYoGLW4tj3bNztkzy2rzOu9acUS30FqUGF4g3gDf/ycOs5cGGdk9aWzBM= =ruuB -----END PGP SIGNATURE----- _______________________________________________ p2p-hackers mailing list [email protected] http://lists.zooko.com/mailman/listinfo/p2p-hackers
