Boosting IoT rollout with virtual hardware

Boosting IoT rollout with virtual hardware

Interviews |
By Nick Flaherty

Last week saw a key launch by ARM to provide IoT software development for massive teams in the cloud. The ‘IoT Total Solution’ is based around fast virtual models running in the cloud and ARM’s Project Centauri.

“From our analysis we need to make it simple for people to test their software as early as possible but also quickly with multiple iterations and in the cloud and Project Centauri is one element. At ARM we are in a unique position to do this to address the fundamental problem of scaling up software development and it is for this reason that we launched Arm Total Solutions for IoT,” said Tran Nguyen, senior director of design services at ARM and a board member at Accellera.

“There is a lot of fragmentation and duplication in the IoT software ecosystem with multiple hardware abstraction layers (HAL) and non-differentiating middleware components, causing inefficiencies, cost and inhibiting scale. Project Centauri aims to deliver a set of device and platform standards, as well as reference implementations for device boot, security and cloud integration. So the idea of Project Centauri is to push for standardisation of a layer of software that allows a wide range of microcontrollers to connect to the cloud.”

The virtual models are of the core processor sub-system called Corstone, currently an M55 microcontroller core and U55 AI accelerator core, although there are plans to use an A53 core in future versions. The key is that the pre-verified model runs in the cloud (on Amazon Web Services at the moment) on a virtual machine for teams of developers to use with their existing software development tools.

“We built the standard hardware platform, Corstone, pre-verified and all ready to go. ARM Virtual Hardware enables software development on a virtual representation of the hardware Corstone platform. Rather than waiting for silicon, or an FPGA image of an RTL design, software developers can use an instantiation of ARM Virtual Hardware running in the cloud. For large teams, developers can each instantiate their own model,” he said.

However the different IoT chips have a wide range of peripherals.

“You take the ARM Virtual Hardware and start to run it as an Amazon Machine Image (AMI) in AWS. As part of the virtual target there are virtual interfaces that help you inject the I/O data that you want,” he said. “We are just starting on the journey and ARM will be providing targets.

“We are talking about accelerating the software development, not replacing the hardware. You still have access to the SDK FPGA emulation and prototyping if you need cycle accurate information,” he said.

“The Total Solutions for IoT enable modern cloud-based development methodologies such as CD/CI integrated with a verified environment from ARM, which means you don’t have to worry about the basics of system verification so you can start to develop and focus on the differentiation immediately.”

He point to the various tracks in DVcon Europe this week looking at the anatomy of the verification flow, how to prototype and verify complex SoC, as well as using machine learning for verification.

“This is very powerful, when you have this in the cloud, you have software binaries pushed to the device, this is monitored and captured and fed back to the training network, and this is the starting of the journey,” he said.


Related articles

Other articles on eeNews Europe  






If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles