Improved Meet-in-the-Middle Attacks on AES

We improve the existing distinguishers of AES. Our work is mainly built upon the works by Gilbert& Miner [17] and Demirci & Selcuk [14]. We find out that some part of the inner encryption function of AES can be expressed with relatively few constants under certain conditions. These new distinguishers are exploited to develop a meet-in-the-middle attack on 7 rounds of AES-128 and AES-192, and on 8 rounds of AES-256. The proposed attack is faster than the existing attacks [15,17] for key size of 128 at the expense of an increase in the complexities of memory and precomputation.