Choosing the Right SBC for Industrial IoT: Key Considerations

For industrial Internet of Things (IIoT), cloud-based data centers are a critical resource that provides scalable, high-capacity storage, processing, analysis, regular security updates, and continuous feature enhancements. However, relying solely on cloud services has its drawbacks. For instance, in mission-critical functions such as process and motion control, latency between the cloud and the edge can be excessive. Moreover, data may be vulnerable to hacking, and massive data transfers could quickly strain communication resources. The solution to these issues is to perform processing and analysis as close to the edge as possible.

In the realm of edge processing for industrial IoT, single-board computer (SBC)-based computing platforms offer designers a growing array of well-supported, cost-effective solutions. To cater to this new market, embedded systems manufacturers have introduced 64-bit processor SBCs equipped with high-speed memory, support for solid-state high-capacity storage, multiple operating systems, high security levels, dozens of GPIO pins, and many other features.

This article explores the demand for SBCs in edge applications and highlights important considerations for selection and integration when implementing edge computing within industrial IoT. Additionally, it illustrates how designers can apply SBCs from Advantech, Digi, UDOO, and VersaLogic, as well as Raspberry Pi, to create computing solutions tailored for industrial IoT edge environments.

Why Use SBCs in Industrial IoT

For designers in industrial IoT, the standout advantage of SBCs is their highly flexible platform configuration, allowing the selection of only the necessary features while also being scalable—unlike some specially designed PCs that require the entire platform to be scrapped. Another advantage is that by designing systems from scratch, designers can accumulate valuable knowledge, which becomes essential when expanding computing resources in the future.

SBCs come in various types, just like their target applications. For example, SBCs commonly used in defense and aerospace systems are available in 3U and 6U sizes, typically based on the Open VPX standard. Host processors often utilize high-end Intel processors and may include Xilinx’s Virtex series FPGAs or GPUs as hardware accelerators, 12-bit and 16-bit ADCs and DACs, and large-capacity DDR4 memory. Additionally, backplanes consist of multiple channels of 4th generation PCIe and switching structures such as RapidIO and PCIe. Of course, these features come at a significant cost.

On the other hand, thanks to the widespread popularity of Raspberry Pi, Arduino, and other platforms, DIY or “maker” SBCs have sold millions worldwide. While they are much cheaper, these SBCs can also connect a range of sensors, perform moderate processing, and send results to field or edge computers, which are based on one or more powerful SBCs. This clearly increases the bill of materials but is reasonable, as it enhances the intelligence of edge sensor device clusters, allowing edge computers to make certain decisions locally.

In between defense/aerospace applications and DIY solutions are SBCs designed for industrial applications. While they may also support Raspberry Pi and Arduino, they must exhibit superior performance and durable environmental characteristics; thus, host processors often utilize Arm® Cortex® series or mid-range Intel Core series. These SBCs have board sizes less than 6 in.2, yet perform comparably to mid-range laptops, equipped with DDR3 or DDR4 memory, and allow designers to choose their own storage options.

Other standard features include support for SPI and SPX, gigabit Ethernet, low-voltage differential signaling (LVDS), and PCIe, various types of security features such as Trusted Platform Module (TPM), audio and video input and output, 8 to 12 USB ports, and support for dual-channel and quad-channel SATA 3.0 storage. Typical accessories include various types of mounting hardware, cooling units, and cables. Many SBCs also support daughter cards to extend communication standards not included on the motherboard, and in some cases, support 4G cellular networks. Additionally, these SBC manufacturers provide a wealth of technical resources, such as development boards and prototyping kits.

A typical example of an Intel-based SBC is Advantech’s AIMB-581WG2-00A1E (Figure 1). This 9.6 in.2 board features Intel Xeon E3-1275 and Core i7-2600 processors, supporting up to 32GB of DDR3 memory. Another example is UDOO’s SC40-2000-0000-C0-V, which is a 4.72 in.2 board based on the AMD quad-core 2GHz Ryzen embedded V1605B CPU, equipped with AMD’s 8 GPU Radeon Vega 8 graphics accelerator, supporting up to 32GB of DDR4-2400 memory and various high-capacity storage options.

Choosing the Right SBC for Industrial IoT: Key Considerations

Figure 1: Advantech’s AIMB-581WG2-00A1E SBC exemplifies how SBCs integrate critical and extended functionality within ultra-small dimensions. (Image source: Advantech)

Unlike many other industrial SBCs, VersaLogic’s Liger VL-EPM-43SCP-08 operates using both Windows and Linux (Figure 2). This board, measuring PC/104-Plus 4.2 x 3.7 in., can increase functionality through stacking boards. Unlike previous PC/104 versions, it supports PCI bus and ISA. The VL-EPM-43SCP-08 is based on a 2.8GHz Intel Core i7-7600U CPU, equipped with 8GB of DDR3 memory (expandable to 16GB) and SATA 3.0 high-capacity storage. Other interfaces include: microSD slot, I2C interface; RS-232, RS-422, and RS-485 (optional); two miniDisplayPorts, one HDMI output, with a display resolution of up to 4096x 2304. Additionally, this board meets MIL-STD-202G for vibration and shock resistance.

Choosing the Right SBC for Industrial IoT: Key Considerations

Figure 2: VersaLogic’s VL-EPM-43SCP-08 SBC runs Windows and Linux, using PC/104-Plus dimensions. (Image source: VersaLogic)

Digi takes a different approach with its ConnectCore 6 system-on-module. This module is based on NXP Semiconductors’ i.MX6UL-2 processor series and integrates the application processor and Arm Cortex-A7 core into a single device (Figure 3).

Choosing the Right SBC for Industrial IoT: Key Considerations

Figure 3: The ConnectCore 6 System-on-Module (SoM) is based on NXP’s iMX6UltraLite application processor and integrates nearly all SBC functions into a single device. (Image source: Digi)

The CC-SB-WMX-J97C version of the ConnectCore 6 SoM measures 4.7 in.2 and offers Bluetooth 4 and Wi-Fi, the company’s Digi XBee radio (based on the IEEE 802.15.4 standard), cellular connectivity (optional), and gigabit Ethernet, supporting multiple displays, equipped with camera and expansion connectors (Figure 4).

Choosing the Right SBC for Industrial IoT: Key ConsiderationsFigure 4: The CC-SB-WMX-J97C SoM supports multiple wireless standards and the company’s XBee radio, measuring 4.7 in.2. (Image source: Digi)

Considerations for SBC Selection

For existing industrial IoT systems, the first step in the design process is to assess the current edge application needs of the company and how much they may grow in the future. The latter is more of an assumption than a statement of fact, as we cannot accurately predict when these resources will be needed. Experience shows that companies that implement industrial IoT often underestimate their needs initially, so the best approach is to assume that demand will grow over time.

The next step is to identify the basic resources required, including wired and wireless connections, support for high-capacity storage, and the inputs and outputs needed to drive displays, audio and video, panel lighting, and speakers. These are generally not difficult, as SBCs with the necessary performance for industrial IoT typically include all these features.

Another factor to consider is whether additional functionality can be added to the SBC through expansion boards. For example, while most SBCs integrate Wi-Fi and Bluetooth transceivers, many industrial IoT systems employ Zigbee and other possible short-range wireless standards, as well as low-power wide-area network (LPWAN) technologies such as LoRaWAN, Sigfox, or narrowband IoT (NB-IoT) provided by wireless carriers.

In terms of software, multiple operating systems are available, most of which are based on the official Raspberry Pi Raspbian or various versions of Linux. The Arduino Integrated Development Environment (IDE) supports Windows, macOS, and Linux. Windows 10 is often excluded mainly because it is incompatible with Raspberry Pi, and until recently, industrial IoT applications have only begun to show interest in this operating system.

Finally, designers must consider the environmental conditions of the system installation site, which may require reinforced enclosures or protection against water, dirt, vibration, and shock.

Turning to SBC Clusters

While these SBCs are very useful, the capabilities achievable are limited if designers can only use a single circuit board. However, as application scales grow, circuit boards can also be expanded. To create micro-supercomputers, Los Alamos National Laboratory and NASA, among others, have established SBC clusters. However, SBC clusters are also within the capabilities of industrial IoT designers, as exemplified by a 40-node Raspberry Pi 3 Model B cluster (Figure 5). This 40-node cluster is based on 40 Raspberry Pi 3 Model B boards, equipped with 20GB of memory and supporting up to 12TB of high-capacity storage, all within dimensions of just 9.9 x 15.5 x 21.8 in.

Choosing the Right SBC for Industrial IoT: Key Considerations

Figure 5: This 40-node cluster, based on 40 Raspberry Pi 3 Model B boards, has 20GB of memory and supports up to 12TB of high-capacity storage, all within dimensions of just 9.9 x 15.5 x 21.8 in. (Image source: LikeMagicAppears!)

This type of system demonstrates how to build powerful, scalable SBC clusters for industrial IoT edge computing based on Raspberry Pi and other architectures, which embedded systems developers should take note of. For such applications, the Raspberry Pi Model 3B+ can provide an excellent starting point. Compared to traditional clusters, SBC clusters are smaller, cheaper, and consume less power, making them ideal for the limited space of edge applications.

It has been proven that there are multiple ways to achieve high performance within constrained spaces. For instance, the Pi Stack technology brings DC power into the cluster from a single point and distributes it throughout the cluster (Figure 6). This reduces wiring, allowing more Raspberry Pi boards to be installed within a given space. Communication between nodes can be achieved through the Ethernet interfaces provided by the SBC.

Choosing the Right SBC for Industrial IoT: Key Considerations

Figure 6: Building SBC clusters using the Pi Stack construction method proposed by Philip Basford et al. (Image source: Future Generation Computer Systems)

One of the most striking aspects of SBC clusters is their ability to provide extremely high performance at low cost using off-the-shelf SBCs, power supplies, and various peripherals. This concept has only recently been introduced for industrial IoT edge computing, but it deserves careful consideration.

Conclusion

SBCs designed for industrial IoT applications are gradually increasing, providing designers with many attractive solutions for building edge computing platforms. With just the addition of power supplies, enclosures, and some peripherals, scalable, cost-effective solutions can be customized to meet the demands of various operating environments within ultra-small dimensions.

Choosing the Right SBC for Industrial IoT: Key Considerations

Leave a Comment

Your email address will not be published. Required fields are marked *