Figure 2-3 Industrial equipment developers design for 10+ year support cycles amid rapid hardware innovation

For most industrial and automotive applications, AI integration and computation in embedded systems is trending toward hardware closest to the system sensors, rather than being outsourced to cloud services. This trend, edge computing, has a variety of benefits, including cost savings, reduced latency and enhanced security for the consumer. However, this raises unique challenges for software developers. Unlike data centers, edge computing hardware requires extended software lifecycles. In industrial end equipment, such as programmable logic controllers (PLCs) and robotic arms, developers design their applications with 10 years of support or more in mind, often clashing with rapid hardware innovation. Forced migration is a legitimate concern for embedded software developers, and they often consider their suppliers’ software longevity policy with the same weight as the hardware longevity policy. Also, embedded software engineers must design with appropriate security measures. They need in-depth tools for boot-time security, runtime protection and data-at-rest protection to offer comprehensive coverage. Without a robust development kit, these protections are not feasible to implement. On top of this, engineers must protect all layers of the stack, including hardware, firmware and software. Therefore, engineers need an SDK with industry-standard compliance and hardware-backed security, or risk safety issues in the embedded system.

In the automotive industry, edge computing trends are shifting vehicles toward increasingly complex architectures. These structures are revolutionizing the way vehicles compute AI models for safety systems. Rather than a disjointed approach to computing, vehicles are pushing AI models closer to respective sensors, leaving room for a centralized computer to make informed and unified decisions for the entire vehicle. As a result, multiple processors must communicate, creating the requirement for a scalable SDK. Embedded software engineers must verify that they can reuse drivers and toolchains across the vehicle to keep software development on track.