When design firms, semiconductor companies, and industry experts talk about IoT products, they always focus on the PCB and the firmware. One aspect of product development that is discussed less often is the link between the end product and the cloud, as well as the off-board computing architecture needed to support these products. IoT products use their own computing resources to capture and process data, but not all IoT products use a standalone application. These products require off-board computing resources to perform their core functions and provide services to end users.

Today, computing resources are being moved away from large, centralized data centers and being distributed around the network edge. IoT and edge computing products work together to provide additional computing resources needed for IoT-driven services. Edge computing resources sit between the cloud and the field, and they enable multiple applications that are unsuitable for cloud deployments. Many of these IoT and edge computing applications are focused on smart infrastructure, where computing power and network access must be as close as possible to the end user.

IoT and Edge Computing Infrastructure

Edge computing products play an important role in the modern IoT infrastructure, particularly in smart cities where many IoT products are deployed at high density. The end IoT products serve as nodes that act as the interface of the network, where data is captured and processed locally as part of service delivery to users. Meanwhile, edge servers provide higher-level compute and storage for network nodes. When needed, such as when the ecosystem must interface with a web application or much higher compute is required, the network can connect back to the cloud via the internet.

This infrastructure aids several important tasks in advanced applications like robotics, sensor fusion, environmental monitoring, AI/ML, and connected automobiles. Edge computing systems can be highly generalized with high compute density, which allows multiple IoT nodes to offload their computational tasks onto an edge server or a proprietary edge data center. This decentralized, more advanced infrastructure will be the new cornerstone of smart cities, where localized computing supports a huge number of IoT products.

The design challenge in these products exists at the hardware level, particularly when end products need to be custom designed from existing modules. In addition, IoT products require an application that involves development on multiple levels.

IoT Product Design in Smart Cities

All IoT products will have some common characteristics that determine how the connect to a larger network, provide services to users, and interface with the outside world. Some of the basic characteristics of IoT products include:

• A host controller that enables data collection and processing from external interfaces and devices
• Analog interfaces for signal capture and conversion to digital data, particularly when sensors are necessary
• Network connectivity method, which could be wired, wireless, or both
• Connections to external devices or modules that expand a product's feature set and capabilities

These products are only successfully designed when the PCB is properly laid out and follows best DFM practices. Designers often take a modular approach with IoT products, where hardware modules are grouped together into a single electronic assembly. However, with newer IoT products having more aggressive form factor and higher feature density, a modular approach may not be the best choice. Companies that want to build an IoT ecosystem need an engineering and manufacturing partner that can coordinate all parts of the design and production process.

Edge Computing and Application Development

Systems providing high compute at the edge have similar characteristics as in-the-field IoT products within smart city deployments. The difference is one of scale; edge computing products must interface with multiple devices over the required network protocol, functioning as a miniature data center. In smart cities, these products need to be rugged enough to provide services with high uptime, while having enough compute for data processing and storage from a potentially huge number of network nodes.

Edge applications also need to be developed and optimized around the embedded hardware in these systems to ensure the appropriate level of service can be provided to end devices. Finally, a development team needs to integrate the embedded application deployed in the field with the edge and cloud applications that provide core services. Development teams from each discipline must work together to ensure a developed product will be functional at all levels.

To overcome these design and development challenges at the edge and in the cloud, edge computing systems can also take a modular approach in terms of their hardware and software. Edge computing systems that interface with the cloud can leverage existing hardware modules as part of their design, while commercial or proprietary cloud services provide the final link between an IoT ecosystem and a web interface. Experienced developers should know how to mix and match these hardware and software options, depending on the desired user experience and services required in the product.

Producing Your IoT Ecosystem for Smart Cities

Each of the above aspects of IoT development and hardware design should occur in parallel, where hardware designers and application developers work together to create new products that are manufacturable at scale. EMS companies that provide value-added engineering services for IoT products can help implement the development process and produce new products at scale. Bringing both aspects of design and development for IoT and edge computing systems into one firm lets application developers take an integrated approach, leading to higher product quality.

An experienced EMS provider that brings extensive design, engineering, and process automation experience can help you build IoT and edge computing systems to deliver services in a smart city environment. PCI has more than 30 years of EMS experience focused on consumer goods, industrial, automotive, and medical devices, as well as in Printed Circuit Board Assembly (PCBA) and box builds. Our Lean Six Sigma manufacturing expertise enables us to customize our manufacturing line to meet our partners' requirements.

If needed, we provide our customers the flexibility needed to quickly scale production as needs arise. We provide our partners with high-quality products at lower manufacturing costs thanks to our shorter change-over time and leaner material control. Contact PCI today to learn more about our capabilities.