Research On Graph Reduction For Call Graph Mining-Based Software-Defect Localization

With the rapid development of information technology in various fields, Internet applications have penetrated into all walks of life. They are changing the way people live and the quality of life. At the same time, they are impelling the progress of social economy. Meanwhile, people began to rely on these application systems a lot, not only dependence increased and relationship became close, but also in many areas the application systems closely related to people’s vital interests, even related to personal safety. Therefore, software security has direct effects on the interests of the whole society. But the reality is not just as one wishes, software defects can be seen everywhere, they were huge obstacles and bad influence for software production and quality, what’s worse, the obstacles and influence are getting worse. The great harm of software defects pulled the people demand for application systems of software defect location. In order to reduce the cost of software testing, to ensure software security, and to locate system defects fast and accurately. It is of great significance to establish a reliable defect location system which gains good efficiency and accuracy.Call graph mining-based software-defect localization technology as an important branch of defect location, which used graph mining technology to mine characteristic information from call graph, adopted formula to analyze and calculate that characteristic information. It aimed at getting the methods ranking which represent the possible location of the defects. The ranking suggests an order for programmers testing which can significantly improve the efficiency and accuracy of defect localization. The technology includes the following five parts: analysis the defect’s nature, graph reduction approach, graph mining technology, the characteristic information analysis and the evaluation of the location result. Nowadays, research on call graph mining-based software-defect localization technology is hot and challengeable. Any innovation and improvement in the process of defect location can greatly reduce the cost of software testing, which is of great significance to ensure the reliability and security of industry application systems.In view of the existing shortcomings of graph reduction approach for call graph mining-based software-defect localization as following: they couldn’t keep as much information as possible of the original graph which leaded to inaccurate locating and the reduction efficiency is very low, this proposed a novel reduction approach for call graph reduction which introduced parallel subtree reduction. The experimental results show that parallel subtree reduction approach not only can achieve reduction of the graphs which have large number of nodes but also can significantly improve the efficiency of defect localization. The major work and innovations in paper are as follows:(1)Proposed parallel subtree reduction for call graph reduction: This approach achieves efficiency by adopting the Fork/Join framework for parallel computing, this algorithm had been always recursive and splitting subtasks until they were small enough to solve. What’s more, this framework is not only simple in design but also obtain good parallel performance. The approach not only keeps the raw call graph information as much as possible but also greatly improves the efficiency of graph reduction, which can significantly improve the efficiency of defect localization.(2)Application Aspect J framework to getting call graph: Aspect J framework is a kind of aspect oriented framework. It is different from the traditional artificial way of instrumentation. Application Aspect J framework to getting call graph which is not only simple, but also especially suit for the practical large projects. Aspect J framework is very flexible and convenient to modify, which has important practical significance to improve the efficiency of graph reduction.