An Automated Edge Computing Approach for IoT Device Registration and Application Deployment

As the Internet of Things (IoT) evolves rapidly across various industries, the number of IoT protocols and applications is growing with vast number of heterogeneous components and entities. In setups with thousands of IoT devices, manual deployment of applications and device registration become impractical due to their time-consuming and costly nature, as well as the requirement for background knowledge of IoT devices and protocols. Furthermore, IoT devices often have resource constraints that prevent them from running complex software. Therefore, there is a significant need to enhance and optimize edge computing systems for IoT, making them suitable and dynamic for automated IoT device registration and heterogeneous application deployment. In this article, we present an edge-based framework designed to facilitate the automated registration of diverse wireless IoT devices and the deployment of IoT applications. To validate our approach, we use a smart irrigation system enhanced with a containerized machine learning model as a proof of concept. Our evaluation of the implemented prototype demonstrates that our system is scalable and feasible.