Abstract
Recently S. Papadimitriou et al. have proposed a new probabilistic encryption scheme based on chaotic systems. In this letter, we point out some problems with Papadimitriou et al.'s chaotic cryptosystem: (1) the size of the ciphertext and the plaintext cannot simultaneously ensure practical implementation and high security; (2) the estimated number of all possible virtual states is wrong; (3) the practical security to exhaustive attack is overestimated; (4) the fast encryption speed is dependent on the first defect; (5) problems about the dynamical degradation of digital chaotic systems; (6) no explicit indications are given to explain how to construct the virtual state space with the 2^d virtual attractors, the 2^e virtual states and the permutation matrix P. The detailed analyses and discussions on the above problems show that the proposed chaotic cipher is insecure and unpractical. Also, we give our suggestions on the design of general digital chaotic ciphers, and give some open topics in this area.