Getting Started with Cloud-Native Automotive Software Development

This workshop will introduce a novel automotive-native software development infrastructure able to execute the same containerized workload - with environmental parity - in all the targeted compute elements: an AWS EC2 Graviton2 instance, a Raspberry Pi, and an AVA Developer Platform.

As defined by Kevin Hoffman in his book:

The purpose of […] environment parity is to give your team and your entire organization the confidence that the application will work everywhere.

This is a major element in achieving the objective of automotive cloud-native software-defined vehicles.

You will learn how to utilize AWS to create a CI/CD pipeline that builds, containerizes, evaluates, and enables deployment - at scale in the cloud and on embedded devices - of a perception network, YOLO. This network is used as a stand-in for any automotive application workload to demonstrate the design paradigm. The specific version of YOLO used in this workshop is the one implemented in the Autoware stack (YOLOv2-Tiny), running on Ubuntu Linux 20.04.

The full System Under Test (SUT) stack will include the Operating System (a Yocto-Linux distribution) and Arm’s Edge Workload Abstraction and Orchestration Layer (EWAOL).

Continue to the Workshop