The Google Cloud Next ‘18 conference in San Francisco is less than a week away, and with the release of Google Cloud IoT Core managed service earlier this year, there is likely to be a strong emphasis on securely connecting IoT devices at potentially massive scale into the Google Cloud IoT Platform. Arm and its partners will be at the show demonstrating the latest secure client platforms, along with low-power machine learning (ML) at the edge, connecting to Google Cloud IoT Core.
Here is an overview of what you can see on Arm’s booth (S1716) and how you can get involved, either in person or remotely!
Arm’s booth at Google Next ‘18
Machine Learning on Low-Power Microcontrollers
ML at the edge is a major driver for Industry 4.0 and smart cities. Being able to deliver low-power ML solutions that can perform rudimentary image recognition, anomaly detection and more, will drive the future innovations in these markets, with billions of end-node devices collecting and transmitting data to the cloud.
At Google Cloud Next ‘18, Arm will be demonstrating its Project Trillium heterogenous platform with ML image recognition on the high-performance STM32F7 device from STMicroelectronics. Specifically, the STM32F7 processor will classify live images from a camera attached to the board and send this data to an application on the Google Cloud IoT Platform via the Google Cloud IoT Core Service.
Our demo is based on three key open-source components to help engineers use compute resource efficiently:
- CMSIS-NN: a collection of efficient neural network kernels developed to maximize performance and minimize memory footprint of neural networks on Arm Cortex-M
- Google Cloud IoT Device Connector: a middleware library for connecting microcontroller IoT applications to Google Cloud IoT Core
The demo is built with Keil MDK, the ST STM32F7 Discovery Board and the STM32F4DIS-CAM Camera. You can learn more about the MDK Google IoT connector on the Arm Keil website. The neural network code used in the CIFAR10 image classification is on GitHub.
IoT Security Baked into Arm-Based Microcontrollers at the Edge
The Arm Platform Security Architecture (PSA) is the framework for consistently securing IoT, helping to make security a reality across the full IoT ecosystem. PSA sets out three key stages for security design and includes a holistic set of deliverables for each stage (including threat modeling documentation, hardware and firmware architecture specifications, and an open source firmware reference implementation). It’s the perfect starting point for anyone designing an IoT device!
At Google Next ’18, Trustonic will demonstrate Kinibi-M, which is an implementation of Trusted Firmware, aiming to adhere to PSA principles. The implementation is currently running on Arm Musca-A Test Chip Board and Microchip Armv8-M microcontroller boards. This demo will show how a board using an Arm Cortex-M23 processor can securely and automatically enrol in Google Cloud IoT Core. If you’d like to learn more about how Arm and Trustonic have been working together on PSA, you can check out our recent webinar.
Zephyr is a small, scalable, open-source RTOS designed for IoT embedded devices. Linaro (a leading Arm open-source company) has worked with the Zephyr community to deliver connectivity to Google Cloud IoT Core in Zephyr along with support for the Arm PSA, making Zephyr an important platform for future IoT. On the Arm booth, Linaro will demonstrate secure connections to Google Cloud IoT Core and the use of Trusted Firmware-M (TF-M).
Controlling IoT at the Edge with Google Assistant via the MATRIX Creator from MATRIX Labs
MATRIX Creator, developed by the Arm Innovator, Rodolfo Saccoman is a powerful development/prototyping IoT daughter board for the Raspberry Pi family, which allows developers to drive IoT applications from the Google Assistant via it’s onboard eight microphone array. The MATRIX Creator includes the MATRIX Open System which uses Google Cloud to provide capabilities such as a real time dashboard, remote deployment, and more! The MATRIX Creator also has a broad array of sensors: temperature, pressure, UV, motion and orientation, as well as support for wireless communication standards such as ZigBee, Z-Wave, and NFC. Hardware prototyping is also supported via its Spartan-6 FPGA and Atmel Arm Cortex-M3 microcontroller.
And there’s more..
In addition to the partners showcased on the Arm booth at Google Next ’18, there are many others listed on the Google IoT Partners Page as well as some great kits on the IoT Developer Prototyping Kits page.
How to get involved
If you’re attending Google Next ’18, we’d love to meet you there. However, even if you can’t make it, there are still lots of ways to get involved.
Why not learn more about the Arm Platform Security Architecture and Project Trillium, or the Arm partners who are exhibiting at the show?