Shifting Sands of MCUs

The rise of artificial intelligence (AI) is reshaping the landscape for simple microcontrollers (MCUs). With AI-generated code becoming more prevalent and the increasing demand for AI inference, along with the necessity for secure systems in light of upcoming standards like the Cyber Resilience Act (CRA), the role of microcontrollers in system designs is evolving. The shift towards more advanced process technologies at 22nm is enabling the utilization of new memory technologies such as MRAM, leading to enhanced performance at a lower cost. These developments are paving the way for the introduction of new MCU architectures into the market, with many of these innovations on display at the Embedded World exhibition in Nuremberg, Germany.

Memory Safety

SCI Semiconductor is showcasing the first silicon of its ICENI secure 32-bit microcontroller, which is the first commercial device to incorporate the CHERI (Capability Hardware Enhanced RISC Instruction) secure memory architecture. By combining the RISC-V RV32E instruction set with the CHERI hardware architecture, the ICENI device allows legacy code and AI-generated code to be significantly more secure through the utilization of embedded security features in the hardware.

Haydn Povey, CEO of SCI Semiconductor, emphasizes the importance of addressing vulnerabilities in code generated by AI. The integration of the CHERI architecture not only enhances security but also streamlines the development process by enabling code reuse and leveraging tools like CodeForge and AI-generated code. This approach ensures that AI-generated code can run safely on the microcontroller without the need for additional measures like thread isolation in Linux.

The memory safety achieved through the CHERI hardware-enforced capability model, developed in collaboration with Microsoft and the University of Cambridge, replaces traditional pointers with unforgeable, bounded capabilities that specify memory regions and permissions. By restricting software components to access only delegated memory, ICENI enforces robust spatial memory safety, preempting potential exploits before they can manifest.

Povey highlights the ease of implementing memory safety measures with the CHERI architecture, emphasizing that transitioning to a higher resilience layer through compartmentalization incurs minimal code impact. This compartmentalization strategy limits the impact of potential attacks, offering a more efficient and effective approach compared to legacy microcontroller protection models that often come with performance penalties.

• Europe’s drive for digital sovereignty accelerates

The ICENI microcontroller, built on the low-power 22FDX silicon on insulator process at GlobalFoundries in Dresden, ensures European sovereignty in microcontroller production. Povey underscores the significance of this sovereignty, particularly for companies operating in critical infrastructure sectors, such as aerospace, defense, and the digitalization of the electricity grid.

Silicon Labs

Meanwhile, Silicon Labs is gearing up for its Series 3 range of MCUs ahead of its acquisition by Texas Instruments. The Series 3 22nm platform distinguishes itself with enhanced digital content, enabling customer software to run alongside the wireless stack seamlessly. This integration caters to high-value applications that necessitate real-time operating systems and off-chip flash for execute in place (XIP), offering a more scalable software solution.

Daniel Cooley, CTO of Silicon Labs, underscores the significance of AI inference in the current landscape. With a focus on sensor data processing and machine learning, Silicon Labs is incorporating multiple accelerators to meet the evolving demands of the market. Cooley also acknowledges the shift towards licensed accelerators in the microcontroller space, emphasizing the importance of security in light of impending regulations like the CRA.

Both SCI Semiconductor and Silicon Labs are at the forefront of innovation in the MCU market, addressing the growing need for secure, efficient, and high-performance microcontrollers that can cater to a wide range of applications.