多层可信软件代码中程序标注准确性验证仿真

摘要

对多层可信软件代码中程序标注的准确性的验证,能够有效保证多层可信软件的正常运行.对程序标注准确性进行验证,需要建立软件不同程序行为切分的超平面,区分不同的标注错误类型,完成可信软件代码程序标注准确性的验证.传统方法结合机器学习算法,提取可以反映程序标注准确性行为的特征,但忽略了对标注错误类型的区分,导致验证精度偏低.提出基于支持向量机的多层可信软件代码中程序标注准确性验证方法.通过构建不准确标注的行为特征空间,建立不同程序行为切分的超平面,从而对不同的标注错误类型进行区分,对多层可信软件代码程序标注中所有权标注、效果标注、唯一性标注、恒定和只读标注、保护锁标注的内容准确性进行验证,从而实现程序标注准确性验证.实验结果表明,所提方法能够准确对多层可信软件代码中程序标注进行检测,保证了验证的准确性.%The verification of the accuracy of the program annotation in the multi-layer trusted software code can effectively ensure the normal operation of the multi-layer trusted software.To verify the accuracy of program annotation,we need to establish hyperplanes for segmentation of different program behaviors,distinguish different types of annotation errors,and complete the verification of the accuracy of trusted software code program.Traditional methods combine machine learning algorithm to extract features that reflect the accuracy of program annotation,but neglect the distinction between annotation error types,resulting in low accuracy of verification.A method for verifying the accuracy of the program annotation in the multi-layer trusted software code based on support vector machine is proposed.By constructing the behavior feature space with inaccurate annotation,different program behavior segmentation hyperplane is established to distinguish different types of errors of annotation,multi-code annotation in trusted software ownership tagging,tagging,tagging,only the effect of constant and read-only annotation,content protection lock annotation to verify the accuracy,so as to realize the program accuracy verification.The simulation results show that the proposed method can accurately detect the program annotation in the multi-layer trusted software code,and ensure the accuracy of the verification.