Industrial IoT Architecture Design and Implementation

Preface

Although the concept of the Internet of Things (IoT) appeared in 1999, it was not until IBM introduced the "Smart Earth" idea in 2009 that many countries elevated IoT research to a strategic level. In China, development still relies heavily on government projects, making the future outlook uncertain.

The government, prompted by internet companies, proposed the "Internet+" concept. While both focus on networking, "Internet+" emphasizes the internet itself with peripheral smart modules, whereas IoT centers on sensor data collection, device control, and remote monitoring.

Many internet companies market their work as "IoT" while actually delivering "Internet+" solutions, causing confusion about the true positioning of IoT.

From a technical standpoint, IoT can be seen as an extension of traditional industrial control networks, which are typically isolated for safety and costly to maintain. With the maturity of internet and mobile networks, remote monitoring needs have emerged in agriculture, forestry, aquaculture, and other fields.

"Internet+" projects prioritize user count and data traffic, aligning with cloud‑platform business models, whereas small‑scale IoT projects focus on system stability, reliability, ease of development, and long‑term maintenance.

Understanding these differences, this article introduces the architecture and implementation of industrial‑grade IoT projects.

Concept and Features of Industrial‑Grade IoT

Industrial‑grade IoT refers to projects with industrial‑level characteristics (e.g., agriculture, forestry, livestock, fisheries) that require stable, reliable devices, easy adjustment of control strategies, upgradeability, extensibility, and maintainability.

Traditional industrial control projects are large, costly, and time‑consuming, while IoT leverages existing internet infrastructure and cloud services to deliver solutions more quickly and cheaply.

The reduced development cost expands the applicable domains, creating a virtuous cycle of increased reliability and lower price.

Industrial‑Grade IoT Architecture Design Philosophy

A typical IoT system consists of three parts: device side, cloud side (public cloud), and monitoring side.

1. Device‑Side Architecture

The device side handles data acquisition, process execution, and control. Devices act as IoT gateways that communicate with the cloud, optionally processing raw data and executing business logic. Communication may use HTTP (GET/PUT) for low‑volume data or TCP/UDP sockets for high‑volume, low‑latency streams.

Functionally, the device side can be divided into three layers: data acquisition & control output, process execution, and data upload & command‑receiving communication.

2. Cloud‑Side Architecture

The cloud side typically includes a web front‑end, web back‑end, and middleware.

The front‑end displays process diagrams, data reports, curves, logs, and diagnostic information. The back‑end handles HTTP requests, real‑time data via WebSocket, and establishes relationships between device data and reports. Middleware provides long‑running services for device communication, complex business logic, data conversion, and storage.

3. Monitoring‑Side Architecture

The monitoring side may be a PC, mobile phone, or tablet. Mobile apps enable remote monitoring in the era of mobile internet.

Typically, the monitoring architecture consists of a UI layer and a data‑communication layer that exchanges information with the server.

4. Summary

Implementing the three major components can be done with any platform or language, but a good architecture must also consider reliability, scalability, and maintainability.

Industrial‑grade IoT projects span agriculture, livestock, and fisheries, requiring customized code and extensions. Engineers with limited expertise must be able to develop robust code quickly.

These projects involve embedded device development, web front‑ and back‑end development, service programming, and mobile app development, each with its own technology stack (e.g., Windows CE/.NET Micro Framework, Linux, FreeRTOS, iOS, Android, UWP, various web frameworks).

Choosing different technology stacks for each component often necessitates separate development teams and integration effort, as well as ongoing maintenance for both developers and field operators.

General IoT Middleware Platform Architecture Design

The platform is a generic middleware for multiple industries, focusing on reusable technical components and a common runtime.

4.1 YFIOs Embedded Data‑Configuration Architecture

YFIOs adopts a three‑layer design: driver layer, strategy layer, and core layer.

The driver layer supports most physical communication interfaces to acquire sensor data and send control commands.

The strategy layer can load system policies and user‑defined policies, enabling on‑site process control without server interaction.

The core layer creates two in‑memory databases: IODB for point data (e.g., temperature) and IOBC for block data (e.g., images). Drivers and strategies interact through these databases, achieving loose coupling.

Key advantages over traditional configuration software include .NET‑based driver and strategy development with garbage collection, C#/VB.NET SDK for driver development, hot‑swappable drivers, and remote upgrade/debug capabilities at runtime, configuration, and driver/strategy levels.

2. YFCloud Cloud‑Side Middleware Architecture

YFCloud extends YFIOs to a networked version, removing the IOBC database and simplifying the driver layer to a TCP/IP socket interface. Each remote device maintains a socket connection, sending heartbeats if idle.

YFCloud integrates a WebSocket server for real‑time web communication and manages projects via templates and a platform‑level API for creation, editing, and lifecycle control.

3. General IoT Platform Architecture

YFIOs runs on the IoT gateway and communicates with YFCloud. The web front‑end provides process visualization, reports, curves, logs, parameter configuration, and camera monitoring. The web back‑end handles user/role management and project/template administration.

4. Summary

The platform’s main advantage is that it is entirely built on the .NET stack, reducing the need for diverse technical expertise and simplifying secondary development and maintenance.

IoT Project Case Summaries

1. Home Remote Health Monitoring System

Integrates blood‑glucose meters, blood‑pressure monitors, cameras, temperature/humidity modules, RFID, and 3G. Doctors can view data via a web portal and send messages to the device. The configuration‑based architecture allows driver adaptation for various sensor models.

2. Agricultural Greenhouse Monitoring System

An IoT gateway connects cameras and environmental sensors, transmitting data over Ethernet, Wi‑Fi, or 3G. Users monitor crop growth via PC, tablet, or mobile.

3. Offshore Fishery Monitoring System

Collects water quality via Modbus RTU, images via camera, GPS coordinates, and sends data through GPRS to a remote server.

4. Village‑Level Wastewater Treatment Monitoring System

Gateways link RS485/CAN devices to collect data and control equipment, communicating with the server via Wi‑Fi or GPRS. The web interface displays process diagrams, reports, and parameter settings.

Future Development Directions for IoT Projects

While consumer‑oriented IoT (smart homes, vehicle‑IoT) focuses on high volume and uniformity, industrial IoT faces diverse processes, harsh environments, 24/7 reliability requirements, and the need for extensibility and maintainability.

Future efforts will prioritize reliability, extract common components to minimize per‑project modifications, and, as cloud computing matures and standards solidify, enable different vendors to specialize in hardware or software parts under coordinated customer oversight.