Design and Implementation of IoT Data Platform Based on Object Model and Rule Engine

With the formulation of the “Made in China 2025” strategy and the urgent need for digital transformation and cost reduction in manufacturing enterprises, the industrial Internet of Things (IoT) has developed rapidly, with increasing industrial applications and increasingly rich usage scenarios. However, many issues have arisen in the development of IoT. The data generated in complex industrial scenarios is massive and heterogeneous. How to parse this data and assign semantic information to it has become a challenge. Additionally, how to utilize the collected data to maximize its value through processing has also become another difficulty. To address these challenges, He Jingjian from the Metallurgical Automation Research and Design Institute has designed and implemented an IoT data platform.

Author He Jingjian designed a practical system with functions such as data collection, data parsing, data storage, data processing, and data sharing with other business systems based on the characteristics and requirements of IoT data. The overall technical architecture consists of five layers: data collection layer, data parsing layer, data storage layer, data processing layer, and data sharing layer. The data collection layer integrates various IoT communication protocols, where the gateway converts these protocols into MQTT protocol and publishes the data to the cloud. The data parsing layer parses the subscribed data according to the configuration information of the object model and ultimately persists the data into the time-series database TDengine. The advantage of using the object model to parse data is that it can first assign semantic information to real-time data, simplifying the difficulty for users to understand the data. Secondly, it can improve the efficiency of data parsing. Finally, using the object model can decouple physical devices from cloud devices, simplifying the development process and enhancing development efficiency. The data processing layer’s functionality is implemented by the rule engine, which realizes the linkage between data processing and device control based on different business requirements.

The rule engine designed by He Jingjian decouples business rules from application code, covering comprehensive data processing logic. It leverages the high concurrency and low latency features of the Go language for performance. Moreover, it uses Docker-compose for container orchestration, significantly enhancing the scalability and availability of the rule engine. The data sharing layer primarily implements external data distribution through RESTful interfaces and publish-subscribe methods.

According to the design plan, the author’s solution has realized an IoT data processing platform. Taking the precision forging machine, a typical device in steel mills, as an example, the complete process from connecting to the cloud, data reporting, data parsing, data storage design, to final data processing has been implemented, demonstrating the overall functionality of the IoT data platform and providing new ideas for the collection and utilization of industrial data.

References:

[1] He Jingjian. Design and Implementation of IoT Data Platform Based on Object Model and Rule Engine [D]. Metallurgical Automation Research and Design Institute, 2022.