Applying Agile to Large-Scale, Safety-Critical Systems: A Comparative Analysis of Agile and Waterfall Approaches in Satellite Development

Abstract

Large-scale, safety-critical cyber-physical (LS/SC/CP) systems, such as satellites, face significant challenges in balancing the need for safety, regulatory compliance, and documentation with the demand for faster development cycles. This study examines the impact of applying Agile methodologies to the LS/SC/CP system by modeling the development of a fictional mid-size Low Earth Orbit (LEO) satellite using Innoslate. We created two development models: one following NASA’s traditional Waterfall process from Phase A to Phase D, and another using an Agile approach with incremental Minimum Viable Products (MVPs). The models were compared regarding schedule and cost, revealing that the Agile approach delivered the satellite two times faster with reduced costs. However, applying Agile to safety-critical systems introduced several challenges, including regulatory compliance, safety assurance, integration complexity, bi-directional traceability, documentation requirements, and cultural barriers. We applied specific adaptations to the Agile model to address these challenges, including automated compliance checks, integrated hazard analysis, added traceability mechanisms, and streamlined documentation practices. Our findings suggest that these adaptations significantly mitigate the risks associated with Agile adoption in LS/SC/CPS. The study concludes that a tailored Agile approach—augmented with industry-specific adaptations—can improve development speed and flexibility while maintaining compliance, safety, and quality standards, thus providing a viable alternative to traditional Waterfall processes for future satellite development projects.