In our paper, we investigate a statistical property of random functions we named drainage. (Definitions for drainage, random function, etc. will be given shortly.) Our motivation for doing so is twofold. First, it generally assumed that a good cryptographic system will exhibit no simple statistical regularity. For example, the function from key to ciphertext when a block cipher is used to encode a fixed plaintext should appear to be completely random. We were therefore interested in studying the behavior of drainage for a random function, and then comparing it to the measured behavior for a real cryptosystem, the DES. Secondly, drainage is closely related to statistical properties which are important to the performance of the generalized cryptanalytic attack proposed by Hellman , discussed below.
KeywordsDirected Graph Random Graph Random Function Large Component Block Cipher
Unable to display preview. Download preview PDF.
- M. E. Hellman, “A cryptanalytic time-memory tradeoff,” IEEE Trans. Inform. Theory, vol. IT-28, pp. 401–408, July 1980.Google Scholar