The digital transformation agenda is spreading across many industries, in which connected devices and the internet of things (IoT) help to improve processes, operations, and many aspects of businesses.
In the past, a connected device might have been a simple sensor with a communications system integrated to transmit the measured data to a control system. In such simple systems there was very little intellectual property (IP) sitting near the connection so the risk of IP theft was minimal.
However, the need to address issues like latency and security as well as enabling real-time feedback and control has meant a lot more processing and neural network accelerators are being integrated into the edge device or sensor, making them much more sophisticated and also holding a lot more value in the microcontroller or the MCU. In other words, more IP is contained within the edge device and that means the value of the IP at risk is much higher – and the cost of theft of that IP is potentially much more damaging to a business or its brand.
To understand the scale of this challenge, market analyst International Data Corporation (IDC) has said worldwide shipments of edge artificial intelligence (AI) processors will reach 340.1 million units in 2019, and it expects unit shipments of 1.5 billion by 2023. Every key supplier is trying to develop solutions that address the new computing paradigm in the IoT and addressing edge AI. This is beyond the general-purpose processors and graphic processor units (GPUs) that might in the past have enabled high performance compute capabilities within mobile, battery-powered devices.
Many vendors are now bringing AI-optimized processors, both discrete accelerators and host processors with neural network accelerators integrated into the processor. Key application areas for edge AI processors include automotive advanced driver assistant systems (ADAS), gaming systems, smart home, and video surveillance. More broadly, they will be utilized more often in sectors like industrial automation, medical devices, AR/VR devices, and robots and drones.
Hence, we are seeing the value of the IP within small MCUs within edge IoT devices increasing due to the growing implementation of AI at the edge. This is especially the case as each vendor uses AI and highly valuable algorithms to provide their competitive advantage. Even logic from the cloud is moving to the edge.
The effect of this is that the value of the software executed on the MCU is becoming significantly higher. It is therefore vital that this IP is protected.
The software development tool Embedded Trust from IAR Systems/Secure Thingz provides an efficient way to design-in protection of valuable software IP during the development phase. The necessary cryptographic mechanism is provided by the toolchain and requires no cryptographic expertise by the firmware developer – allowing the developer to focus on implementing software and complex AI algorithms.
A cryptographic hybrid scheme in Embedded Trust combines symmetric and asymmetric algorithms in such a way that the integrity and confidentiality protection of the software is guaranteed based on topical crypto technologies. Protection of these keys is anchored in the available hardware security depending on underpinning SoC (system-on-chip) security.