Software Trustworthiness Measurement Technology Research Based On Attribute Evaluation

With the wide application of computer technology in the fields of politics, economy and science & technology, software technology has been applied to all aspects of people's life.Software becomes larger and more complex, and the quality of software is also faced with greater challenge. Software trustworthiness measurement technology, which is used to evaluate software trustworthiness level, is also becoming more and more important. However,the technology of software trustworthiness measurement is not mature, which is showing that software trustworthiness has no unified definition, software trustworthiness attribute has no unified standard and there is not mature software trustworthiness measurement model,existing problems of software trustworthiness measurement methods also exist theory difficult to practice and measurement result is not accurate.To solve the questions above, the paper presents a software trustworthiness measurement method based on Bayesian network, the method establishes Bayesian network based on the software architecture, divides software from the structure, measures software modules in the Bayesian network, and finally obtains parameter distribution relevant to the trustworthiness indexes of the measure modules. Then compute whole trustworthiness attribute based on the modules' trustworthiness indexes and Bayesian network. The paper's main work is as follows:Firstly, the paper analyzed the relationship among the trustworthiness attributes, which lays the foundation for synthesizing and measuring software trustworthiness attributes.Secondly,in order to ensure the operational feasibility of the trustworthiness measurement, this paper proposes a method based on the software architecture to construct the Bayesian network, and studies software modular expression, including the analysis of relationship of the inheritance and calling among the categories and modules of the software,and establishing the Bayesian network according to dependencies among them, exchanging software measurement from system level into module level and reducing the difficulty of the measurement.Again, measuring software components based on the established Bayesian network. The paper first verifies the software function points for getting the function attributes and verifying the integration level of the software. In terms of reliability, measured the indexes of the successful paths and the successful execution tasks of the software. To quickly obtain accurate measurement results, used a priori probability method when measuring indexes; In aspect of software security,in order to cover significant but low frequency impact,a acceleration metric method based on the importance sampling principle used to reach an overall measurement results; The maintainability attributes are evaluated and explored in the Bayesian networks maintenance evaluation.Finally, the vehicle multi cooperation path planning simulation software measurement experiments was carried out,using Bayesian network to measure software by dividing it can reduce the difficulty of the measurement process. Measure pertinence part of software, can reduce the number of measurement cases, So the method is feasibility and efficiency in measuring software trustworthiness.