Abstract

Block ciphers are among the mostly widely used symmetric-key cryptographic primitives, which are fundamental building blocks in cryptographic/security systems. Most of the public-key primitives are based on hard mathematical problems such as the integer factorization in the RSA algorithm and discrete logarithm problem in the DiffieHellman. Therefore, their security are mathematically proven. In contrast, symmetric-key primitives are usually not
constructed based on well-defined hard mathematical problems. Hence, in order to get some assurance in their claimed security properties, they must be studied against different types of
cryptanalytic techniques. Our research is dedicated to the cryptanalysis of block ciphers. In particular, throughout this thesis, we investigate the security of some block ciphers constructed
with new design strategies. These new strategies include (i) employing simple round function, and modest key schedule, (ii) using another input called tweak rather than the usual two
inputs of the block ciphers, the plaintext and the key, to instantiate different permutations for the same key. This type of block ciphers is called a tweakable block cipher, (iii) employing linear and non-linear components that are energy efficient to provide low energy consumption block ciphers, (iv) employing optimal diffusion linear transformation layer while following the AES-based construction to provide faster diffusion rate, and (v) using rather weak but larger
S-boxes in addition to simple linear transformation layers to provide provable security of ARX-based block ciphers against single characteristic differential and linear cryptanalysis.
The results presented in this thesis can be summarized as follows:
Initially, we analyze the security of two lightweight block ciphers, namely, Khudra and Piccolo against Meet-in-the-Middle (MitM) attack based on the Demirci and Selcuk approach exploiting the simple design of the key schedule and round function.
Next, we investigate the security of two tweakable block ciphers, namely, Kiasu-BC and SKINNY. According to the designers, the best attack on Kiasu-BC covers 7 rounds. However, we exploited the tweak to present 8-round attack using MitM with efficient enumeration cryptanalysis.
Then, we improve the previous results of the impossible differential
cryptanalysis on SKINNY exploiting the tweakey schedule and linear transformation layer.
Afterwards, we study the security of new low energy consumption block cipher, namely, Midori128 where we present the longest impossible differential distinguishers that cover complete 7 rounds. Then, we utilized 4 of these distinguishers to launch key recovery attack against 11 rounds of Midori128 to improve the previous results on this cipher using the impossible
differential cryptanalysis.
Then, using the truncated differential cryptanalysis, we are able to attack 13 rounds of Midori128 utilizing a 10-round differential distinguisher.
We also analyze Kuznyechik, the standard Russian federation block cipher, against MitM with efficient enumeration cryptanalysis where we improve the previous results on Kuznyechik, using MitM attack with efficient enumeration, by presenting 6-round attack.
Unlike the previous attack, our attack exploits the exact values of the coefficients of the MDS transformation that is used in the cipher.
Finally, we present key recovery attacks using the multidimensional zero-correlation cryptanalysis against SPARX-128, which follows the long trail design strategy, to provide provable security of ARX-based block ciphers against single characteristic differential and
linear cryptanalysis.