摘要
为研究面向交互行为的软件行为可信性分析和动态度量问题,从软件与外界数据交换时所产生的信息流动过程中,确定并提取依赖于输入数据的语句和变量的集合,作为软件交互行为的预期模型.在软件实际运行过程中,通过动态获取软件的实际行为,与分析软件源代码得到的软件预期行为作比较,形成可信性的动态度量结果.采用程序切片技术,将软件交互过程及交互中被赋值变量及后续赋值行为作为切片分析的结果.在此基础上,继续从源代码的抽象语法树和编译后的可执行程序中提取赋值行为本身和被赋值的变量的详细信息,然后将这些信息与切片分析结果相结合,形成以交互行为为关注点的软件行为描述模型.对模型进行了测试,结果表明:该模型能识别出软件运行中的非预期行为.
To research the dynamic trusted measurement issue for interaction-oriented behavior of a software system, the expected behavior description model was obtained through determining and extracting the statements and the variables, which were referred and relied by the input data from outside. During the software executing, the measurement was done by dynamically acquiring the actual behavior of a software system and then the system compared it with the behavior obtained from the source code. By adopting the program slice technology, the interaction behavior and the successive behavior about assigning variables during the interaction were determined. Then, the detailed information and assigning behavior were extracted from the abstract syntax tree of the source code and the compiled executing program. Result shows that this approach can identify the non-expected behavior during the software executing.
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2013年第5期713-721,共9页
Journal of Beijing University of Technology
基金
国家"973"计划资助项目(2007CB311106)
北京市自然科学基金资助项目(4122007)
关键词
可信软件
交互
动态度量
trusted software interaction dynamic measurement