Stateless Job Offloading for Mobile Robots in Kubernetes

Edge devices are increasingly requiring more and more intelligence, hence asking for an amount of computing power that that is not always sustainable on the individual device itself. In particular, this applies to IoT devices where battery consumption and low processing power are constraints that limit the amount and complexity of tasks that can be performed on the edge, hence demanding for a tighter edge-to-anything interaction for job offloading. This paper focuses on mobile robots running the ROS operating system, presenting an offloading approach for long-lived stateless services based on a switching approach, an algorithm that optimizes for a faster switching between local and remote execution, minimizing downtime of the service.