Informatica

A key exchange (or agreement) protocol is designed to allow two entities establishing a session key to encrypt the communication data over an open network. In 1990, Gunther proposed an identity-based key exchange protocol based on the difficulty of computing a discrete logarithm problem. Afterwards, several improved protocols were proposed to reduce the number of communication steps and the communicational cost required by Gunther's protocol. This paper presents an efficient identity-based key exchange protocol based on the difficulty of computing a discrete logarithm problem. As compared with the previously proposed protocols, it has better performance in terms of the computational cost and the communication steps. The proposed key exchange protocol provides implicit key authentication as well as the desired security attributes of an authenticated key exchange protocol.

Recently, Tseng et al. proposed an improvement on Peyravian and Zunic's protected password transmission scheme and protected changing scheme to remove some security flaws. However, as we will point out in this paper, any adversary can intercept the request for changing the password sent by a legal user and modify it with a wrong password. Furthermore, we shall also propose an improved version of their protected password changing scheme to help it out of the trouble.

Smart card has been adopted to various applications. In 2000, Hwang and Li proposed a remote user authentication scheme, which is also using smart card. Nine months later, Chan and Cheng pointed out that there is a weakness in the remote authentication scheme proposed by Hwang and Li. In this paper, we show that Chan and Cheng's attack does not work well because they did not consider the format of user's identity. In addition, we propose several ways to solve the problem of Chan and Cheng's attack.

In an internet environment, such as UNIX, a remote user has to obtain the access right from a server before doing any job. The procedure of obtaining acess right is called a user authentication protocol. User authentication via user memorable password provides convenience without needing any auxiliary devices, such as smart card. A user authentication protocol via username and password should basically withstand the off‐line password guessing attack, the stolen verifier attack, and the DoS attack. Recently, Peyravian and Zunic proposed one password transmission protocol and one password change protocol. Later, Tseng et al. (2001) pointed out that Peyravian and Zunic's protocols can not withstand the off‐line password guessing attack, and therefore proposed an improved protocol to defeat the attack. Independently, Hwang and Yeh also showed that Peyravian and Zunic's protocols suffer from some secury flaws, and an improved protocol was also presented. In this paper, we show that both Peyravian and Zunic's protocols and Tseng et al.'s improved protocol are insecure against the stolen verifier attack. Moreover, we show that all Peyravian and Zunic's, Tseng et al.'s, and Hwang and Yeh's protocols are insecure against DoS attack.

Yamaguchi, Okayama, and Miyahara proposed a simple but efficient authentication system, SPLICE/AS. In this article, we show that their method is vulnerable to the guessing attack. An attacker can obtain the password, private-key, and public-key of the user. To overcome the vulnerability of SPLICE/AS to the guessing attack, we propose an improvement of their system. In our scheme, we not only prevent the guessing attack to obtain secret messages but also enhance the security of the SPLICE/AS authentication system in WIDE.