Queueing-Theory-Based Models for Software Reliability Analysis and Management

Software reliability is one of the most important internal attributes of software systems. Over the past three decades, many software reliability growth models have been proposed and discussed. Some research has also shown that the fault detection and removal processes of software can be described and modeled using an infinite-server queueing system. But, there is practically no company that can afford unlimited resources to test and correct detected faults in the real world. Consequently, the number of debuggers should be limited, not infinite. In this paper, we propose an extended finite-server-queueing (EFSQ) model to analyze the fault removal process of the software system. Numerical examples based on real project data are illustrated. Evaluation results show that our proposed EFSQ model has a fairly accurate prediction capability of software reliability and also depict the real-life situation of software development activities more faithfully. Finally, the applications of our proposed model to project management are also presented. Our proposed model can provide a theoretically effective technique for managing the necessary activities of testing and debugging in software project management.