A Data Flow Programming Framework for 6G-Enabled Internet of Things Applications

The adoption of 6G networks and related technologies by the Internet of Things (IoT) are expected to unlock novel advancements in several vertical industries, like Automotive and Robotics. One of the key enablers consists on expanding IoT applications across the entire Cloud-to- Things continuum. Thus, components can be (re)deployed anywhere and anytime according to the requirements at each point in time. As a consequence, application developers face new challenges to transparently interconnect these components, leading to patchwork designs that increase complexity, time to market, and development cost. This paper introduces a novel data flow programming framework, named Zenoh-Flow, that addresses the aforementioned challenges by providing unified abstractions, automated allocation, and deployment across the Cloud-to-Thing continuum. Zenoh-Flow additionally provides key functionalities for Automotive and Robotics, e.g., deadlines, time-stamping and progress tracking. Preliminary results show that our proposed solution outperforms existing solutions, like ERDOS and ROS/ROS2. In particular, Zenoh-Flow achieves lower latency and higher throughput in synthetic benchmarks as well as a reduced computing and networking overhead in a real-world robotic scenario.