Research On Techniques For Software Dependability Growth Test Case Generation

As software complexity and malicious attacks continue increasing, dependability has now become one of the most serious challenges, and how to generate the test cases for software dependability growth system has become a hot topic. Therefore, based on recent studies, relative researches were developed in this paper to tackle the problem in this field. Our main contributions are summarized as follows:Firstly, a software dependability growth system is designed, this system takes components, connectors, interfaces, configuration information as basic elements, and it takes Self-regulation, self-configuration as the growth way. By the time, the system is an open structure and mission-critical system, and it supports top-down and multi-level design for the dependability.Secondly, a method towards architecture to generate test cases is designed. In the paper, a set of test criteria according to software architecture Configuration information were presented and their corresponding computational rules are defined. After structuring the Configuration Behavior Graph(CBG) and intermediate configuration information, based on anglicizing the basic dependencies among the components, with the test criteria, a serious of test cases are generated.Thirdly, basing on learning the properties of the dependability, an evaluation metrics model is constructed. The paper describes the evaluation model formally, and quantifies the contents of the basic indicators. Then, for increasing the application of the model, a self-adapted rule of software dependability growth system is designed.At last, an experiment is conducted by using the method. In addition, an autonomic dependability evaluation tool, CGT-Tool, is developed to meet the requirements of dynamic and large-scale evaluation, and the attribute of the experiment system is evaluated.