Enterprise IoT Challenges for PCB Designers in 2020
Enterprise IoT became one of the biggest tech buzzwords in 2019, with many analysts promising improvements in business efficiencies and profits simply by connecting everything in the office. As time goes on, real problems that can be solved by connecting business assets start to become clear, and practical opportunities for innovation and problem solving can be identified. I would argue we’re starting to move away from the 1990s-equivalent age of enterprise IoT and into a decade where useful applications can become widely commercialized.
The ideas for enterprise IoT implementation sometimes ignore the hardware challenges as well as the economic feasibility, promising the world in benefits while ignoring the costs. On the software side, major companies are busy developing the software architecture required to bring these applications to reality, ultimately enabling embedded hardware developers to create innovative new products for an enterprise IoT environment. Let’s take a look at some important opportunities in this area and break down the hardware challenges.
Three Enterprise IoT Design Challenges
The list ideas for enterprise IoT applications is long, but they all present some common challenges to PCB design for enterprise IoT devices. Designers will need to confront challenges in one or more of the following areas:
Multiple Wireless Protocols
Unless you plan to connect all enterprise assets with twisted pair cable, you’ll need to choose a wireless protocol for communication. So which is best for enterprise IoT devices? This is not such a simple question to answer, especially when there are over a dozen possible wireless protocols to use for networking. Everyone is familiar with Bluetooth and WiFi, but there is a whole list of other wireless protocols that vary in range, data rate, frequency, and intended application. For enterprise applications, protocols operating in licensed frequency bands are preferred to prevent interference.
You’ll most likely need multiple wireless protocols to provide a range of communication and configuration options. Short-range, moderate data rate applications like Bluetooth will most likely need to appear alongside longer range, lower data rate applications like LoRaWAN. With telecom companies now getting involved in the enterprise IoT market, cellular IoT protocols (NB-IoT and LTE-M) are a useful option for transmission of small chunks of data over longer range. Take a look at this article for more information on IoT protocols that can be used in an enterprise environment.
PCB designers need to be prepared to accommodate multiple wireless frequencies in their boards alongside digital components to provide data processing. Smartphone makers have solved a lot of layout and space problems by integrating everything into a variety of SoCs/SiPs, but custom board designers may not have this luxury. Pay attention to best practices for isolating RF circuit blocks from digital components, planning return paths, and designing a stackup to accommodate these aspects of mixed-signal design.
Low Power Design for Mobile Devices
Just like most enterprise IoT applications won’t be communicating over copper, they probably won’t be receiving power over copper either. Designers need to minimize power usage in mobile devices by opting for lower power components and eliminating unnecessary functions. To extend lifetimes, designers can consider adding some energy harvesting (thermoelectric, RF, piezoelectric, or other methods) capabilities to a device.
Edge Data Processing vs. Cloud Data Processing
This is a major decision that will determine the level of embedded processing power and memory that must be placed in the device. If your device is simply gathering sensor data and uploading it to a base station, then you won’t need much beyond a microcontroller or small FPGA. If your enterprise device needs to run prediction or classification in the field and in real-time, then you need to pack significant processing power into your embedded device. You’ll also need to scale your memory depending on the amount of data being stored or processed; applications involving high resolution images and video require much more memory than sensor data.
This actually works against the previous point, as these types of ML/AI algorithms consume a ton of power on current hardware. However, IC startups are coming up with new ASICs and SoCs that can be dedicated specifically to low-power AI computational tasks while using lower power. Once these new devices become commercialized, designers will have more tools they can use to provide dedicated AI functions directly on their devices, rather than relying on the cloud for these computationally heavy tasks.
Uniform Data Formats
If you’re working over a standard TCP/UDP protocol, packet formats are already rigidly defined and data formats are built into embedded OS images. Depending on your application, you may need to enforce a novel unified data format. In areas like industrial IoT, the IPC-CFX standards enforce a uniform data format for machine-to-machine communication on the factory floor—such a standard can certainly be applied elsewhere. This particular standard is adaptable in multiple protocols, making it more of a messaging structure rather than some new wireless standard.
Enterprise IoT Opportunities for Embedded Developers
Implementing an enterprise IoT vision is as much about software as it is about hardware. All IoT devices are effectively embedded systems with varying levels of complexity. Any of the application areas presented above will require some level of software engineering, both on and off a new device.
Application areas like AI for customer experiences, asset tracking, and predictive analytics have already seen significant innovation within the open-source software developer community. Custom code and application examples for a variety of applications already abound for desktop, web, and cloud applications. If any of these applications are to move to the edge in an enterprise IoT environment, code for these applications will need to be adapted to run on embedded devices. Supporting services that run on a private network, private enterprise cloud, or public cloud will also need to be tailored to receive and process data from embedded devices.
API developers will play a decisive role in bringing this interface between edge devices, an enterprise network, and external stakeholders to reality. Just like connected manufacturing assets in an industrial IoT environment, running everything through HTTP requests with JSON data provides an excellent framework for quickly connecting customized edge devices with enterprise networks. Every software developer I know has experience working with JSON data structures; this type of standardization will allow everyone from experienced developers to new entrepreneurs to participate in enterprise IoT development.
The schematic design, PCB layout, and design reuse tools in Altium Designer® were created to help you design PCBs for any application, including enterprise IoT devices. These tools sit at the cutting edge of the EDA industry. You’ll be able to design compact boards for powerful IoT devices, manage your design data, and prepare your new enterprise IoT product for full-scale manufacturing.
Now you can download a free trial of Altium Designer and learn more about the industry’s best layout, simulation, and production planning tools. Talk to an Altium expert today to learn more.
