Modeling Fault Detection Phenomenon in Multiple Sprints for Agile Software Environment

Abstract

The information technology industry has gone through a revolutionary change in the last two decades. In today’s fast changing business environment, the IT organizations have to be agile and responsive to cater to the needs of the customers. The objective is not just to deliver quickly but also to embrace the change without having any adverse impact on the project. The requirements of the end customers are fast changing and get evolved over a period of time as these are directly aligned with the market needs. This has led the organizations to adopt “Agile” approach based on “lean” principles over the conventional software development life cycle (SDLC) approach. In “Agile” framework, the customer works in collaboration with the project team in prioritizing the requirements. The implementation is done through “Scrum” methodology, having multiple “sprints,” and each sprint has a “working software” as a deliverable. This approach has substantially reduced the “time to market” as the customer can decide which features of the software they would like to be delivered on a priority basis. The release of sprints is similar to multi-releases of a software in which software is tested rigorously to detect the underlying faults at the end of each sprint and remaining number of fault of each sprint is taken forward for the next sprint. Hence to model the fault detection phenomenon and their trend in each sprint, software reliability growth modeling has been used. In the current work, we are using software reliability growth models (SRGMs) to find out the trend over the sprints that could ultimately define the overall quality of the software. Numerical illustration is mentioned at the end of the paper for model validation.