For the Industrial Internet of Things (IIoT), the edge of the IoT – where the physical world meets the virtual world of data and the cloud – is perhaps the most important part. It’s here where data is collected and gathered, and also where the action happens. The insights gathered from the cloud are all implemented at the edge.
But what do these edge devices look like, and how do we implement them? While edge devices share some of the capabilities of any other embedded system, they are often space, power and resource constrained. This requires a unique approach that optimizes the use of their limited processing capabilities.
This is reflected both on the software side, where embedded OSs and RTOSs are often deployed, and on the hardware side, where custom SoC designs allow for the best use of limited resources.
Rather than the heavyweight, full-fledged operating systems that were used to in consumer products, like Windows or MacOS, IoT devices use specialized operating systems designed for embedded systems. Depending on the real-time requirements of the device, this can be either a generic embedded system OS such as uClinux or OpenWRT or if real-time capabilities are required, an RTOS such as FreeRTOS or ARM’s Keil RTX.
The choice of OS depends on the resource constraints of the device, as well as its desired functionality. Higher-level IoT devices such as IoT gateways can use more resource-intensive OSs such as embedded Linux distributions. Lower-level edge devices will often use a more lightweight OS, often an RTOS, which is not only lightweight, but provides the real-time determinism which allows for direct control by the device over time-dependent industrial processes, or if extremely regular sensor measurements are required.
Besides the OS itself, edge devices will often also require some kind of middleware to handle networking and security.
Like IoT software, IoT hardware is also often resource constrained. Cost, power budget, and space are key considerations. The core components of an IoT edge platform are the sensing and actuating components as well as the communication interfaces, either wired or wireless, or both.
Wireless communications are the most prevalent way for IoT edge devices to connect to the cloud. These include shorter-range PAN protocols such as Bluetooth or RFID, longer-range WLAN protocols such as WiFi, as well as cellular-style WAN protocols such as NB-IoT or LTE-M. Choosing the right wireless protocol requires understanding the power budget of the application, network topology (star, mesh or cellular) and desired signal range.
Wireless modules exist for all protocols on the market, allowing for easy testing and prototyping. However, for large-scale deployment, an integrated approach with the radio incorporated into the SoC often scales better for the mass market.
Besides the communications component, the other major function of an IIoT node is the sensing and actuation. This interface to the physical world often entails some analog and mixed-signal components including sensor interface and analog front ends (AFEs), actuator interfaces and control circuitry, and sometimes both on the same package.
For the IIoT, industrial Ethernet is the most relevant wired communication protocol. In this case, the device is usually not a strict edge device but more a “fog” device. This component generally is a gateway, with industrial Ethernet which controls other edge devices. Industrial Ethernet protocols such as PROFINET or EtherCAT provide the real-time deterministic communication to enable industrial control and automation. The industrial gateway would control edge devices through its industrial Ethernet interface while relaying useful data to the cloud.
An Integrated Approach
These days off-the-shelf modules exist for wireless and wired communication, as well as mixed-signal capabilities to allow for sensor and actuator interfaces, but such an approach yields a device that’s bulkier, less efficient and less optimized than desired. Especially when analog and mixed-signal components such as wireless radios, sensor calibration or AFEs are involved, an integrated SoC approach can yield much better results.
Modern SoCs with all the components needed for a proper IIoT edge device can be created today much more quickly and cost-effectively than ever before. By using its extensive IP libraries, S3 Semiconductors SmartEdge Platform is able to integrate all of the functions of an IIoT edge device from wireless and wired communications to analog and mixed-signal functionality, all onto a single chip, saving space and energy, and reducing BOM costs.