Research On Cube Attack Of Lighetweight Stream Cipher Algorithm

In modern society,more and more devices require the use of cryptographic algorithms to protect their sensitive information,such as smartphones,Io T devices,sensors,and so on.However,unlike high-performance devices such as personal computers or servers,their computing power,storage space,and energy consumption are often very limited.Due to the large computational and storage resources required by traditional symmetric cryptographic algorithms,they cannot be effectively used in these resource constrained environments.In response to this issue,the National Institute of Standards and Technology（NIST）of the United States has launched the Lightweight Cryptography（LWC）.The purpose of this competition is to solicit and evaluate lightweight cryptographic algorithms suitable for resource constrained environments.In addition,in recent years,many lightweight stream cipher algorithms targeting resource constrained environments have also been included in international conferences and journals in the field of cryptography.For the lightweight stream cipher algorithms that have emerged in recent years,the relevant security analysis is often not sufficient.How to provide new security assessments for lightweight stream cipher algorithms is currently a research hotspot.To solve this problem,based on new technologies such as mixed integer linear programming（MILP）and division property,this thesis applied cube attack on the TinyJAMBU cryptographic algorithm that entered the third round of the LWC competition and the Atom cryptographic algorithm that released at the 2021 Fast Software Encryption（FSE）,and proposes a fast method for solving superpoly for stream cipher algorithms,The main research work is as follows:1.A single bit key leakage detection method for TinyJAMBU algorithm is proposed.This method utilizes the MILP model and division property to construct a cube attack model against the TinyJAMBU algorithm,and achieves single bit key leakage of superpoly.For the TinyJAMBU algorithm,the initialization and encryption processes were first modeled using the MILP automated analysis tool.Then,by adding MILP constraints to the keystream of the TinyJAMBU algorithm,a single bit key leakage experiment was achieved,and the time complexity of restoring its superpoly was controlled within an effective time range.Compared with existing cube attack results,under the condition of limited cube set selection,this thesis increases the number of cube attack rounds from 2176+0 in the encryption stage to 2176+345,and the time complexity of restoring its superpoly is only2¹⁰ operations.2.A new cube attack method against Atom stream ciphers is proposed.Based on the MILP model and division properties,combined with algebraic degree evaluation techniques,a cube attack model was constructed for the Atom stream cipher algorithm.For the Atom stream cipher algorithm,the initialization process and keystream generation process of the Atom stream cipher algorithm are first modeled using the MILP automated analysis tool.Then,the cube attack based on division property and algebraic degree evaluation method are combined to achieve the key recovery attack of the Atom cipher algorithm.Compared to previous cube attack results,this thesis increases the number of cube attack rounds from 67 to 73.3.A new method for quickly reconstructing superpoly in cube attack is.The combination of two-subset division property,reconstruction technology based on the property of Boolean function,and three-subset division property greatly improves the efficiency of solving superpoly in cube attack.Firstly,a MILP model was established using two-subset division property,and by combining flag techniques,algebraic degree evaluation,and term enumeration techniques,possible monomials in the subfunctions of the superpoly were screened out;Then,using the properties of Boolean functions,the monomials that may be included in the superpoly are determined according to the monomial of the subfunction,and finally the specific expression of the superpoly is accurately solved by using the three-subset division property.Compared with existing techniques for solving superpoly,this thesis combines the two-subset division property and three-subset division property,providing a new approach to improve the efficiency of solving superpoly.