Abstract
Distinguishing analysis is an important part of cryptanalysis. It is an important content of discriminating analysis that how to identify ciphertext is encrypted by which cryptosystems when it knows only ciphertext. In this paper, Fisher’s discriminant analysis (FDA), which is based on statistical method and machine learning, is used to identify 4 stream ciphers and 7 block ciphers one to one by extracting 9 different features. The results show that the accuracy rate of the FDA can reach 80% when identifying files that are encrypted by the stream cipher and the block cipher in ECB mode respectively, and files encrypted by the block cipher in ECB mode and CBC mode respectively. The average one to one identification accuracy rates of stream ciphers RC4, Grain, Sosemanuk are more than 55%. The maximum accuracy rate can reach 60% when identifying SMS4 from block ciphers in CBC mode one to one. The identification accuracy rate of entropy-based features is apparently higher than the probability-based features.
Highlights
The main purpose of cryptanalysis is to study the deciphering of encrypted messages or the forgery of messages[1]
The principle of the cryptosystem identification scheme based on statistical methods is to design the identification index first, and calculate the index value based on the extracted features, and identification result is determined by the size of the index value[5]
In order to make up for this deficiency, this paper aims at the improvement of the Fisher’s discriminant analysis (FDA)-based classification technique proposed in[13], and selects practical 18 algorithms which are 4 kinds of stream ciphers and 7 kinds of block ciphers in ECB mode and CBC mode
Summary
The main purpose of cryptanalysis is to study the deciphering of encrypted messages or the forgery of messages[1]. It means using various methods to try to get all or part information of plaintext by ciphertext, under the condition of not knowing or not fully knowing the details of the decryption key and the cryptosystem adopted by the communicator[2]. Cryptosystem identification is an important part of cryptanalysis[3]. The statistical method and the machine learning have been used to identify cryptosystem from existing ciphertext[4]. The principle of the cryptosystem identification scheme based on statistical methods is to design the identification index first, and calculate the index value based on the extracted features, and identification result is determined by the size of the index value[5]. The identification scheme of cryptosystem is regarded the identification task of the cryptosystem as the pattern recognition task3467. Because of its simple design and stable results, it has attracted the attention of many researchers
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal of Networked and Distributed Computing
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
