Byzantine fault tolerance is a methodology for reasoning about and implementing PeerToPeer reliability and security.  The Byzantine model assumes that a certain subset (in the traditional model, a fraction of one-third) of peers are arbitrarily malicious -- they will forge data, return false results, and so in the worst possible fashion in order to destroy the reliability of the otherwise reliable disjoint subset (two-thirds in the traditional model) of peers.  Named after fictional treacherous "Byzantine generals", who use false signal flags to dupe their fellow generals, a more apropos name might be "RIAA fault tolerance."  :-)  Traditional Byzantine fault tolerance can be improved using probabilistic security, especially cryptographic protocols -- see http://szabo.best.vwh.net/distributed.html.

----
CategorySecurity CategoryDistributed