Automated SQA Framework with Predictive Machine Learning in Airfield Software

Given the intricate composition and complex nature of modern software systems, it is necessary to ensure sufficient software quality throughout their entire life cycle. This paper highlights the development efforts made toward delivering an automated solution for software quality metric acquisition and the analysis of quality-related data for a real-world airfield operations software. The target software system, at the time of producing this paper, consists of over 110K lines of code, requires over 10K developer minutes to address quality issues, contains over 140 identified bugs, has approximately 50 security hotspots, and includes nearly 3K code smells. Considering the abundance of quality-related items uncovered by the solution being developed, the airfield software was presented as an exemplary case study. This paper introduces a novel dual-framework architecture for software quality assurance that specifically targets the airfield software system in focus. This unique approach combines data logging for metric acquisition and machine learning for predictive analysis. This helps address real-time operations, integration challenges, and security concerns in the target software. This paper highlights the tools and technologies selected, the architecture implementing the frameworks and processes used, and the results of preliminary experiments and analysis activities.