Research On Homology Determination Technology Of Multi-form Software Code

In recent years,network security issues using software code as the carrier have frequently occurred.The determination of software code origin has become particularly important as a common key technology to solve security issues such as network attack traceability.Accurately locating the source of software code has become the focus of attention of all parties.As the current software code has multiple morphological characteristics such as multiple operating platforms,various deformation methods,and complex forms,it is extremely difficult to determine the same origin.How to effectively discover the same source of multi-form software codes is of vital importance to enhancing the protection and deterrence of China's cyberspace security.To determine the homology of multi-form software code.Firstly,it is necessary to clarify the concept and boundary of homology of software code,and realize its unified qualitative and quantitative representation.Secondly,it is necessary to pay attention to the homology feature representation of software code and its processing efficiency,and to study new methods to solve the problem of large space-time cost.Thirdly,it is necessary to consider the measures to deal with the loss of homology features of software code and explore new ideas to solve its characterization.Finally,it is need to design a new algorithm to solve the problem of the lack of adaptability of the homologous identification method to adapt to the characteristics of new forms of software that are easily tampered.This dissertation focuses on the issue of multi-form software code homology determination,deeply analyzes the technical development trends and main challenges in related fields,and focuses on the homology determination method of multi-form software codes,and builds a prototype system based on this.The main contributions of this dissertation are as follows:Aiming at the problem of low time and space efficiency of large-scale software code homology determination,a rapid determination method of software code homology based on the Motif structure is proposed.Through the effective characterization of large-scale software semantic structure graphs,the rapid determination of large-scale software code homology is realized.In this method,a feature representation method based on motif structure of function call graph and an algorithm for extracting motif structure are proposed,and the characterization of software code function call graph by motif structure is realized,and the similarity calculation method of motif structure and its frequency distribution is designed,and the rapid identification of software code homology is realized.Experimental results show that the proposed method has obvious advantages in accuracy,robustness and time-consuming,making it applicable to the rapid determination of large-scale software code homology.Aiming at the problem that the deformation of the software code causes the homology feature to be difficult to represent,a binary code homology determination method based on mixed semantics is proposed.By mining the potential homology features of the multi-form binary code,the accurate determination of the multi-form binary code homology is realized.Methods a hybrid feature representation method of text semantics and structural semantics was proposed,and the hybrid semantic representation of text embedding and graph embedding was realized.The hybrid semantic model was embedded into the Siamese neural network framework,and the binary code homology decision model was constructed.Experimental results show that the proposed method can effectively improve the effect of binary code homology identification,and the algorithm has good stability in a variety of test conditions,which can support the accurate identification of binary code homology in a variety of forms.Aiming at the problem of insufficient adaptability of the new form of software code homology determination method,a multi-dimensional mobile application homology determination method is proposed.Through the representation and calculation of the characteristics of the mobile application sub-regions,the accurate identification of the homology mobile application is realized.Methods the multi-dimensional feature representation and targeted processing algorithm of mobile application as a whole,resources and code were proposed,and the homology determination of different sides was realized.The combination and optimization strategy of a variety of similarity algorithms was designed,and the accurate homology determination of mobile application was realized.Experimental results show that this method has significant advantages in accuracy and speed of homology determination,and performs well in real environment,and does not rely on supervised data,which can be used as an supplement to the security strategy of mobile application market.