Objective of authentication is to detect attempted deceptions in a communications channel. Traditionally this has been restricted to providing the authorized receiver with a capability of detecting unauthentic messages. The known codes have all left open the possibility for either the transmitter to disavow a message that he actually sent to the receiver, i.e., an authentic message, or else for the receiver to falsely attribute a message of his own devising to the transmitter. Of course the party being deceived would know that he was the victim of a deception by the other, but would be unable to ''prove'' this to a third party. Ideally, authentication should provide a means to detect attempted deceptions by insiders (the transmitter or receiver) as well as outsiders (the opponent). It has been an open question of whether it was possible to devise authentication codes that would permit a third party, an arbiter, to decide (in probability) whether the transmitter or the receiver was cheating in the event of a dispute. We answer this question in that both permits the receiver to detect outsider deceptions, as well affirmative by first constructing an example of an authentication code as permitting a designated arbiter to detect insider deceptions and then by generalizing this construction to an infinite class of such codes. (ERA citation 12:038398)
展开▼
