Self-equivalence encodings and improvements of white-box implementations

In the white box attack environment, the attacker can not only access the input and output of the cryptographic algorithms, but also obtain the internal details of the algorithms and control the terminal. In this environment, CHOW et al. constructed the look-up tables by using network encodings, embedded the key in the look-up tables, and designed the white-box implementation scheme for the AES algorithm and DES algorithm. The white-box implementation of the cryptographic algorithm based on self-equivalent encodings design is a new implementation method. RANEA et al. designed a white-box implementation scheme for substitution replacement cipher by using the self-equivalent encodings of the S-box. The size of encoding space completely depends on the S-box self-equivalence of the cipher, and the security analysis also shows that the application scope of this scheme is limited. In view of this situation, this paper considers the impact of self-equivalence of the S-box on the security of white-box implementation, and proposes two improved schemes for expanding the encoding space of the white-box implementation scheme by adding self-equivalence encodings to the linear layer or linear encodings to the affine layer. Security analysis shows that the two improved schemes can effectively resist the attacks from RANEAet al, and expand the application scope of the scheme. Finally, based on the above two design schemes, this paper constructs two white-box implementations of the AES algorithm, and compares the security with the white-box AES scheme of RANEA et al. The comparison results show that the two improved schemes can resist protocol attacks based on the centralization problem and asymmetric problem. © 2022, The Editorial Board of Journal of Xidian University. All right reserved.