Block ciphers identification scheme based on randomness test

Abstract

In real scenarios, cryptanalysts can only obtain ciphertext of unknown cryptographic algorithms to conduct ciphertext only attacks. Cryptosystem identification is a prerequisite for further cryptanalysis. This paper proposes a new method of division for ciphertext sequence based on the four widely used cryptographic algorithms DES, 3DES, AES, and Blowfish, which have high security. We redesigned the ciphertext features using five detection methods in the NIST randomness test standard. In the case of using two working modes (ECB and CBC) of block ciphers, the proposed method in this paper obtains lower feature dimensionality and higher final recognition accuracy compared to some previous research works. This study has certain reference significance for the fast and accurate identification of large volume of ciphertext.