Theoretical Analysis of AES Encryption Optimization Based on Quaternions and Rotation Matrices

As a commonly used symmetric encryption method, AES (Advanced Encryption Standard) still has some steps that cannot fully ensure the effectiveness of encryption, and Symmetric Key Encryption has the risk of key leakage in common. This paper mainly researches an optimization model of AES encryption method, which is based on the original AES encryption method. This model adds another password in the SubByte and MixColumns step, in order to ensure the effectiveness of encryption. The optimization model is achieved with programs on python. Experimental data shows that the encryption process is too regular due to the methodologies of the model itself: the location of those bytes after mapping can be separated into some sets, and the transforms of this model cant exchange the mapping of those bytes from different sets. After analyzing and optimizing the model, another improved model is worked out, which can simply confuse the bytes with their locations. The experiments show that the next model has a better encrypting effect than the first one.