Open source software reliability model with nonlinear fault detection and fault introduction

Abstract

AbstractIn recent years, open source software (OSS) has been widely used and developed. The reliability of OSS has also become a hot research topic. Due to the complexity, dynamics, and uncertainty of the development and testing process of OSS, it is also a difficult task to accurately evaluate the reliability of OSS. In this paper, we consider the nonlinear changes of fault detection and fault introduction in the development and testing of OSS and propose a reliability model of OSS based on the nonlinear changes of fault detection and fault introduction. Experimental results show that the proposed model is more in line with the actual changes of fault detection and fault introduction in the processes of OSS development and testing. Compared with other models, the proposed model has better fitting and predictive performance. We also give a simple OSS optimal release method considering the testing‐effort and reliability changes in the processes of OSS development and testing. The proposed model and optimized release method can effectively assist OSS developers to evaluate software reliability and determine the optimal release time.