Towards a Failure-Aware SDLC for Internet of Things

Internet of Things systems carry substantial engineering risks including catastrophic physical failures. To aid software engineers in developing reliable IoT systems, we conducted an experiment to evaluate the influence of learning treatments on design decisions. Specifically, we compared the influence of a set of design guidelines (current practice) and failure stories (proposed learning treatment) on developers' design rationale. We conducted an experiment with 21 computer engineering students using a questionnaire. We observed that both treatments helped subjects reason about criticality as a part of their design rationale. However, failure stories had a greater effect at enabling subjects to reason about safety as a part of their design rationale. We share our results illustrating the effects of a failure-aware design process and propose new research directions to enable a Failure-Aware Software Development Life Cycle for IoT development.