Want to learn how an Internet of Things (IoT) technology stack is built?

Internet of Things (IoT) infrastructure consists of hardware and software components (collectively referred to as a stack) that reside both at the edge and in the cloud. When getting started, there is no better place to begin than with the edge hardware.

Edge hardware is responsible for ingesting and processing data, controlling local devices, and communicating with higher-level cloud and data center systems. To understand how edge hardware tackles these tasks, it is first necessary to dive into three key areas:

1. The diversity of endpoint requirements
2. The need for architectural flexibility and adaptability
3. The challenges of scaling IoT designs

This post touches on each topic to get you started on your IoT journey.

Heterogeneous Endpoint Requirements

IoT endpoints span a vast spectrum, from simple sensors to complex industrial robots, each with unique requirements. Computational needs vary widely, ranging from basic data collection to intensive AI inference at the edge.

Connectivity requirements are equally diverse, encompassing low-bandwidth, intermittent communications to high-speed, constant connections. Power and thermal considerations add another layer of complexity, with solutions ranging from battery-powered devices with strict energy constraints to industrial PCs requiring active cooling systems.

This diversity in endpoint requirements necessitates a range of edge computing solutions. To address these varied needs, SECO offers a comprehensive portfolio of products, from simple , high-performance industrial servers. The following table illustrates SECO’s solutions across this spectrum:

Architectural Flexibility and Adaptability

The diverse and evolving nature of IoT deployments demands architectural flexibility in edge hardware. This flexibility hinges on three key elements:

• Modularity, which enables easy upgrades and customization without complete redesigns
• Scalable processing, ideally with the ability to freely select performance levels and processor architectures
• Versatile I/O and connectivity to meet various sensor, actuator, and network requirements

Computer-on-Modules (COMs)

The Computer-on-Module (COM) approach embodies these principles by separating core compute elements (microprocessor, memory, and often common wired and wireless interface circuitry) from application-specific I/O. This separation allows for customization while maintaining a standardized core.

A COM (also often referred to as a System-on-Module, or SOM) mounts to a carrier board that is optimally designed for the application by implementing specialized circuitry and nonstandard peripheral interfaces. This COM carrier assembly must then be incorporated into a packaged device.

Single-Board Computer (SBC)

The single board computer (SBC) is an intermediate solution, which incorporates the core computing elements of a COM with standard interfaces such as USB, Ethernet, and video, on a single circuit board. The SBC is most usable when standard peripherals are sufficient to satisfy application needs. A custom circuit board design is usually not required. The SBC is incorporated into a machine, which requires enclosure design and device-level test.

Fanless Embedded Computer

A fanless embedded computer is a fully packaged device, complete with regulatory and sometime industry-specific certifications. Similar to SBCs, fanless embedded computers often feature optional I/O modules so they can be adapted to various roles. Support for plug-in AI accelerators and other performance-enhancing add-ons further extends this flexibility. The main advantage of a fanless PC is that it is a ready-to-use solution.

This modular strategy enables system integrators and developers to balance customization and cost-effectiveness. It’s particularly valuable in scenarios where requirements vary across deployments or evolve over time, allowing for efficient adaptation without the need for complete system overhauls.

Horizontal & Vertical Platform Scalability

Scalability in IoT platforms encompasses both horizontal and vertical dimensions.

• Horizontal scalability, or scaling out, refers to the ability to economically deploy solutions across many locations. It’s about finding the right balance between hardware optimization and engineering costs. The COM approach, for instance, offers a good balance for medium-volume applications, optimizing hardware without incurring excessive non-recurring engineering costs.
• Vertical scalability, or scaling up, involves enhancing the capabilities of individual nodes over time. This includes both software updates and hardware upgrades. Software scalability requires initial performance headroom, necessitating a careful balance between future-proofing and current costs. Hardware scalability is facilitated through modular designs, allowing for component upgrades (such as swapping in a newer COM) without full device

SECO supports both forms of scalability with modular hardware designs that offer long-term availability. Plus, our Clea platform offers holistic remote device management, including secure over-the-network software updates, enhancing software-based vertical scaling.

Building a Solid IoT Foundation

The diverse and evolving nature of IoT deployments requires edge hardware solutions that can address a wide range of computational, connectivity, and power requirements while offering flexibility and scalability. By understanding these key aspects, businesses can select solutions that not only meet current needs but also adapt to future challenges.

SECO’s portfolio of edge hardware solutions, combined with our Clea platform, offers a comprehensive approach to addressing these complex requirements across various industries and use cases. Our strategy balances performance, adaptability, and security in IoT deployments, providing a solid foundation for both current and future IoT infrastructure.

Interested in developing your next IoT solutions with our edge hardware? Contact SECO today to find the perfect solution for your IoT deployment.