An Experimental Implementation of an Edge-based AI Engine with Edge-Cloud Coordination

IoT (Internet of Things) has been increasing its coverage of AI (Artificial Intelligence). AI engines are promising solutions for real-time processing and constrained network bandwidth. Intelligent device control consists of an important application of IoT. Device control requires low costs and low latency. A hybrid approach of edge-cloud coordination can exploit the advantages of both of edge and cloud computing. The author presents an edge-cloud coordination architecture for flexible AIoT (a combination of AI and IoT). In order to examine usability and flexibility of AIoT solutions for the proposed architecture, the author develops a testing environment for device-edge-cloud coordination environment. The author describes the challenges of AIoT deployment. From the lessons learned, the author presents a software-perspective of Edge-AI computing in a scope of edge-cloud coordination.