A Quantitative Goal Setting Method for Software Reliability in Early Stages of Software Development

As software size and complexity have been increased, the importance of software quality, especially software reliability, has been more emphasized than ever. However, it is impossible to guarantee 100% software reliability and it requires a lot of development cost and efforts to develop high reliable software products. Therefore, a software product should be developed with an appropriate software reliability goal that is fit to software development project and the usage of target software. In this paper, we proposed a technique to support a quantitative software reliability goal setting in the early stage of a software development life-cycle. Existing software reliability studies focuses on the qualitative expert's knowledge and experiences in the specific domain. However, this qualitative way makes it difficult to accomplish the software development and reliability analysis because software development resources, costs and the reliability goals are different each other even if software products are developed in the same domain. Our contributions of this paper are followings: (1) we define sets of the cost related parameters and product related parameters to identify the tradeoffs between software development cost and software reliability goal. (2) we resolve the trade-off results to maximize the software reliability using optimal software development cost and resources. Finally, (3) we show that the proposed way is useful through the case study with industry data which comes from our past projects.