This forum is about enticing the ROS community to take edge-native principles more seriously as there is good reason to believe they offer significant benefits for robotic systems.
There will be more to come, please feel free to share any questions or anything in general.
As edge computing brings processing closer to the data source,
providing reduced latency,
efficient bandwidth management,
increased privacy for sensitive data,
and ensuring robots can continue to perform critical tasks even in the event of network disruptions or communication failures.
While cloud-native technologies empower organizations to build and run scalable applications in modern environments.
Edge-native applications leverage cloud-native principles while addressing the unique characteristics of the edge, such as resource constraints, security, latency, and autonomy.
Embracing edge-native principles enables ROS developers to build applications that are portable, observable, manageable, and agnostic to language and framework, thus making it easier to scale ROS deployments across diverse edge environments.
So, by designing ROS nodes to scale across multiple edge devices or robots, developers can create distributed systems capable of handling a wide range of tasks in parallel, such as simultaneous localization and mapping (SLAM), navigation, and perception.
This scalability is essential as ROS-powered robots become increasingly ubiquitous in various industries and applications.
Please have no hesitation to reach out to me, eager to hear your thoughts and problems.
Best regards,
Fitz
Here’s two white papers about Edge-Native published by the Cloud Native Computing Foundation (CNCF).
Here’s a great repo created by tomoyafujita:
fujitatomoya/ros_k8s: Kuberenetes / ROS&ROS2 Cluster Samples (github.com)
A tool to consider and one that will be more talked about:
KubeEdge (https://kubeedge.io/)