A Conceptual Model of Agile Software Development in a Safety-Critical Context: A Systematic Literature Review

Context: Safety-critical software systems are increasingly being used in new application areas, such as personal medical devices, traffic control, and detection of pathogens. A current research debate is regarding whether safety-critical systems are better developed with traditional waterfall processes or agile processes that are purportedly faster and promise to lead to better products. Objective: To identify the issues and disputes in agile development of safety-critical software and the key qualities as found in the extant research literature. Method: We conducted a systematic literature review as an interpretive study following a research design to search, assess, extract, group, and understand the results of the found studies. Results: There are key issues and propositions that we elicit from the literature and combine into a conceptual model for understanding the foundational challenges of agile software development of safety-critical systems. The conceptual model consists of four problematic practice areas and five relationships, which we find to be even more important than the problematic areas. From this review, we suggest that there are important research gaps that need to be investigated. Conclusions: We suggest that future research should have a primary focus on the relationships in the resulting conceptual model and specifically on the dynamics of the field as a whole, on incremental versus iterative development, and on how to create value with minimal but sufficient effort.