Analyzing the Behavior of Software Reliability Execution Time Models for Different Agile–Scrum Based Projects

Abstract

Software reliability is an essential component of software development stages. The reliability of a software system plays a vital role in the overall development and success of that software. Reliability is broad and linked to other areas of technologies and approaches. This requires a proactive mechanism, which includes not only technical aspects but also legal and ethical considerations. Maintaining reliability in Agile-based software is an arduous exercise. This happens because in an Agile-based project, frequent changes in requirements are expected during the development cycle. To develop high levels of reliability in an Agile environment, the Quality Assurance (QA) engineer needs to carefully select the appropriate reliability model. In this research paper, we studied the performance of two most popular reliability models namely Basic Execution Time Model and Logarithmic Poisson Execution Model. Since the basic idea behind developing software is quite different in an Agile environment compared to traditional software development processes; a dataset of 30 Agile-based projects has been prepared for the purposes of calculation. These 30 projects in this dataset are divided into three groups as low, medium and high-level projects based on the number of sprints required to complete the work. This paper presents a comparative analysis of these two reliability models on various parameters. As a result, we found that the Logarithmic Poisson Execution Model produces optimal results for most Agile-based projects in all 3 project categories.