Cortex-M55 based Arm Virtual Hardware is now available in AWS Cloud

The Arm DevSummit 2021 is taking place on October 19-21, and the first announcements from Arm are related to IoT with “Arm Total Solutions for IoT delivering a full-stack solution to significantly accelerate IoT product development and improve product ROI”, “Project Centauri” aiming to achieve for an extensive Arm Cortex-M software ecosystem in the way that Project Cassini does for the Cortex-A ecosystem, starting with support for PSA Certified and Open-CMSIS-CDI cloud-to-device specification, and Arm Virtual Hardware based on Corstone-300 IoT platform with a Cortex-M55 MCU core and an Ethos-U55 microNPU accessible from Amazon Web Services.

The first two are quite abstract right now, and more information may become available in the future, but the Arm Virtual Hardware is available now from AWS as a public beta, with 100 hours of free AWS EC2 CPU credits for the first 1,000 qualified users.

The virtual hardware does not emulate only the microcontroller core and microNPU but also peripherals, virtual I/Os, and debug interfaces. The goal is to help software developers, OEMs and service providers to start development without the need for physical silicon. Arm Virtual Hardware is said to “brings modern agile software development methodologies like continuous integration / continuous deployment (CI/CD), DevOps and MLOps to IoT and embedded platforms” without having to setup your own hardware farms.

Documentation can be found on Github.io, and a blog post on the Arm Community further explains the rationale behind the initiative with the embedded software development flow having become more complex with machine learning and firmware updates of fleets of IoT devices.

So what that means is that in the past you could develop the software, create a binary image, and flash it to a single device on your desk, but now IoT solutions must be able to update firmware from the cloud to multiple targets, plus CI and QA requirements may also mandate those. Reproducing this setup in the office would require multiple hardware targets which are usually expensive (e.g. FPGA based designs) and in short supply during development, and since developers already probably need the cloud for training AI data, Arm Virtual Hardware makes perfect sense allowing you to save time and development costs.

There will be a free workshop leveraging Arm Virtual Hardware entitled “IoT DevOps made Simple and Scalable in the Cloud” that will take place on October, 21st at 10am PT. Further details may also be found on the product page.

Jean-Luc Aufranc (CNXSoft)

Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.