Automotive transformation to software-defined vehicles: Red Hat point of view and synergies with state-of-the-art IT

In this post:

• Get an overview of what software-defined means, particularly when it comes to a vehicle. 

• See how open source and Linux can be used to deliver vehicle onboard functionality. 

• Get a glimpse of Red Hat’s vision for the software-defined vehicle and how open source contributors can get involved. 

Building a dream car doesn’t just revolve around factors like horsepower and interior finishes anymore. The automotive industry is going through some big changes as the electric vehicle market revs up and edge computing extends new capabilities for vehicle connectivity. 

At Red Hat, we are working to enable a number of edge use cases—from telecommunications use cases with core network infrastructure for radio access networks to distributed computing for retail environments and manufacturing environments.

During the “Automotive transformation to software-defined vehicles” Ask the Expert session at Red Hat Summit this year, we discussed the car as the ultimate edge server and the role of Red Hat and other open source contributors in this fast-evolving industry. This post will cover some of the highlights from the session. 

First things first: What is a software-defined vehicle?

Making something software-defined means taking its existing hardware platform and moving capabilities that might have been hard-coded in, say, firmware or ROM and bringing them into a software layer that runs on standardized hardware. The software layer also adds the ability to bring in new software-defined functionality.

The simplest versions that you already have today may be an app store or the ability to connect to your mobile phone and have a different audio stream for your infotainment system. 

That’s software-defined, but that's not nearly the level of software-defined we're talking about. A software-defined vehicle encompasses the core aspects of the vehicle that make it possible to drive: the instrument cluster or advanced driver assistance system (ADAS). Having those all be on a software platform allows for the original equipment manufacturer (OEM) or their Tier 1 integrator and supply chain to easily select and change the software components that make up that solution even if the hardware remains the same.

In the past the vehicle was defined by hardware with software that was put on top of it—now it is shifting to the other side. Companies like Tesla are showing us that they build the software and then the car around it. We’re seeing the growing importance of vehicle operations centers and of vehicle software platforms in general.

In the car architectures on the road today, there can be 150 or more small computers — electronic control units (ECUs) — working side by side, each developed to meet a specific use case (like controlling your motorized seat position or the ABS brake function) and using bespoke design. 

The industry is being shaped by a trend to consolidate those to a few and powerful compute units that will serve a multitude of functions while still meeting all the requirements of functional safety (i.e., changing your seat position shouldn't affect the ability to brake or your instrument cluster).

Different divisions within the car manufacturers (OEMs) can no longer just ask their traditional Tier 1 suppliers to design a motorized seat system that brings its own ECU but will need to coordinate the software development to target a common and standardized approach for in-vehicle platforms.

Datacenter on wheels

A datacenter convergence has been happening in the vehicle as more functionality is delivered—like infotainment platforms that may control multiple screens, multi-zone audio, advanced streaming, advanced mapping, heads-up display, etc. 

For example, advanced driver assistance systems that provide automation to the driving experience and zonal control and telematics systems are separate compute systems that are delivered across dozens or more ECUs in a vehicle. Now, these are converging onto a few but more powerful systems (ECU consolidation) that look very much like servers.

As you get more power in a vehicle, it’ll also face challenges that servers and datacenters face: constant security updates and the customers’ expectations for ongoing functional updates.

The future of software-defined vehicles is part of the edge computing future. Just as organizations move toward an open hybrid cloud strategy to take advantage of edge computing capabilities, cars—these datacenters on wheels—are poised to be a prime example of edge computing. 

Software-defined vehicles are a fundamental component of the edge computing strategy, and the strategy doesn’t just include components in the vehicle, but also the network as well as the cloud that supports the operations happening in the car. 

We see the near future of the automotive edge comprising a modern Linux-based operating system and a modern ecosystem that consists of open source and proprietary components that work together to deliver choice and agility in hardware and software components. State-of-the-art software patterns for the distribution and the update of software from the backend (vehicle onboard) into the car will also be in demand.

Driving safely toward a Linux base platform

Red Hat is already well-versed with the necessary technologies for the automotive edge, like Linux, containers, middleware integration, and DevOps architectures. So how can we apply our experience to help build up an enterprise open source layer to run the vehicle from onboard to offboard?

Car manufacturers will need to focus on both differentiating functionalities and integration. Red Hat can help OEMs think about how to best fit those technologies to meet customer expectations for autonomous, electrified and connected vehicles.

Software-defined vehicles stand to prompt demand for reusable and rapidly integrated software components and standardization. There is also a big need to bring state-of-the-art models like virtualization, containers, Linux and open source in general—which all come from the offboard areas—into the vehicle. 

Our goal is to build the basis for vehicle platforms with a Linux-based operating system where the applications and services are running on top of it, and are able to do so safely. Earlier this year, we announced our intent to deliver an in-vehicle Linux platform in collaboration with our functional safety certification partner exida. 

We realize that use cases in a vehicle are demanding. They require certifications such as ISO 26262, an international standard governing the functional safety of electronic systems on-road vehicles. Our work with exida is to help prepare us for certification. 

With Linux as the foundation, we envision working with other vendors in the automotive industry to create blueprints and open components. We want to collaborate with integration partners with application providers in the vehicle itself and also work together with car manufacturers. An open ecosystem approach can help drive the industry forward.

Standardization can open the doors for open source 

The industry is opening up in the space of commodity functionality and base functionality. There’s an opportunity to change the vehicle onboard and offboard environment from vertically integrated to continuous open source innovation. One advantage that Red Hat brings to the industry is our familiarity with the open source community, which can support standardization to increase flexibility.

Open source is fundamentally about collaboration. By collaborating on the base components, the underlying technologies, we’re able to build truly differentiated experiences much more rapidly than software companies creating their own siloed stacks. 

Existing organizations such as Genivi and Automotive Grade Linux (AGL) build either standards or software that provide automotive middleware. But there’s already standardization on abstractions for things like body control and this is just going to continue to grow up to some level that defines the basis of a differentiating experience for the automotive manufacturers.

In the end we expect to see standardization and the ability to use off the shelf hardware components. We’ll probably see more flexibility in bringing in new software functionalities that won’t require a new vehicle to get new capabilities. 

That means there are going to be a lot of ways to get involved as an open source contributor. 

How open source contributors can rev up

As we make progress in reaching a product offering, it'll become much clearer what components are included. Infotainment platforms technologies like multi-zone audio, multi-zone cameras and additional hardware support for ARM or RISC-V platforms are all probably going to end up in the vehicle. 

If you are interested in improving the Linux desktop experience, almost all of those capabilities are going to carry over to the car from an infotainment perspective. Supporting accelerators for ADAS and, basically, anything that brings more platform enablement for Linux is likely to find a path to the vehicle in one way or another. 

We’ll have a number of engineering openings for our automotive activities as we build out our internal team, so please keep checking back on the Red Hat jobs page for those opportunities.

Mapping out the future of automotive

Digital transformation initiatives are well underway for corporate IT and factories, where we’ve helped customers shift to hybrid or multi-cloud scenarios. In automotive, digital transformation will take place in the product, the vehicle itself. 

We’re working with the automotive industry to support its shift from hardware to software-defined technologies. We’re bringing datacenter knowledge to a new use case, but the core of Linux is still there. 

We’ve covered some state-of-the-art capabilities of software-defined vehicles, but are you going to have to be a Linux system administrator to be able to drive a car that is powered by a Linux-based central computer? Of course not. Like Android phone users today, most people probably won't even realize that they have a fully functional Linux system under the hood.

Our 5G vRAN technical wins at the edge of the telco network and our industrial use cases will provide points of view that will help deliver the software-defined vehicle to a new generation of auto-buyers who are looking for connectivity, personalization, and security.

We covered a lot in this session, from the basics of software-defined vehicles to the future of vehicle updates, compatibility with existing systems, and how software security will be handled with this emerging technology. You can learn more by viewing the Ask the Experts session recording on-demand through May 2022. You can also check out our announcement for more information on our work with the trusted functional safety certification body exida.