Key Exchange Using 'Keyless Cryptography'

In a classical cryptosystem, as opposed to a public-key cryptosystem, communication between users is based on a secret key known to them, but no others. A key-distribution protocol is employed to distribute copies of this secret key to the users. One simple way to perform key-distribution makes use of out-of-band communication a diplomatic courier, for example. More desirable for a computer network, however, is a key-distribution protocol that uses the network itself for all communication. Protocols that do this are described in [2,1] For users A and B to agree on a secret key k,,, each first chooses a secret bit string. Let k,[l]k,[2] . . . k,[2n] be k,, the bit string chosen by A, and k,[l]k,[2] . . k,[2n] be k,, the bit string chosen by B. Assume that user names are totally ordered so that each can independently compute the same name k,, for the key on which they will agree. A then constructs 2n messages, one for each bit in k,, where each message has the form: