Optimal Authentication Systems

Abstract

Abstract. In this paper we define an optimal authentication systems as a system whose minimum probability of deception is k / M , k and M being the number of source states and cryptograms respectively, and satisfies information theoretic bounds on the value of impersonation and substitution games. We will characterize order-1 perfect systems and C-perfect systems and prove their optimdity when E, the number of encoding rules, satisfies certain bounds. We will show that both types of systems, in this case, also have best game theoretic performance. Thia will be used to prove that optimal systems exist only if E 2 M 2 / k 2 and for less value of E probability of deception is always greater than k / M . We w i l l prove that doubly perfect codes are optimal systems with minimum value of E and perfect systems are not optimal. Characterization of doubly perfect systems follows from characterization theorems mentioned earlier. We give constructions for each class.

Fail-Stop Signatures

“…In practice, these parameters correspond to the communication and storage complexity of both key generation and the transmission of signed messages. Similar lower bounds have been investigated for other cryptographic schemes with information-theoretic security requirements, e.g., encryption schemes [42], au-thentication codes (starting with [16]; see, e.g., [40]), and unconditionally secure signature schemes [23]. 3 Section 6.1 sketches the information-theoretic background, section 6.2 introduces the random variables, section 6.3 investigates the length of the secret key, section 6.4 investigates the length of signatures and public keys, and section 6.5 answers the questions posed above, i.e., it compares the lower bounds to the known upper bounds and to ordinary digital signature schemes.…”

Information-theoretic bounds for authentication codes and block designs

“…In Section 2 we prove our main theory (Theorem 1), which was first presented in the rump session at Asiacrypt '91. (The latter approach was used in the original version of this paper, see Appendix 7.3C of [8].) It can also be proved by the method of Sgarro [10].…”

Коды Аутентификации С Секретностью (Обзор)

“…[42]) . Для A-кода равенства в (1) и (2) выполняются в том и только том случае, когда для любых m, n ∈ M, n = m, e∈E(m,n)…”