3 Steps to Transition to Cloud-First Vehicles
As the automotive industry transitions to software-define vehicles, the cloud will become a key enabler. But there are three steps that need to be taken first.
The automotive sector stands at the dawn of a new, connected revolution. Software-defined vehicles (SDVs) will completely redraw the automotive landscape, reimagining the vehicle as we know it.
These sophisticated vehicles will be able to deliver digital experiences comparable to those expected on other consumer devices. Vehicles can even be enhanced remotely via over-the-air (OTA) updates, opening new revenue streams for original equipment manufacturers (OEMs).
In the race to develop SDVs, however, we’re seeing newer entrants like Tesla, Rivian, BYD, Lucid, and Nio stealing a march on longer-established automotive manufacturers. To keep up, traditional OEMs will need to adopt a cloud-first approach to accelerate their transition to SDVs.
The Business Case for Cloud-First SDVs
There is no doubt that this migration to cloud-first SDVs will cause a major upheaval to processes, but the business case is strong.
First, the cloud will enable automotive development to happen more quickly and in a highly scalable way. Tasks can be automated, and the testing of new concepts can be accelerated by using simulators and cloud-native solutions, independent of hardware. This will reduce both cost and time to market.
Second, once SDVs appear in the market, OEMs can start providing custom vehicle features to their customers, uncovering new upsell opportunities. OEMs can also administer updates and fix bugs remotely, reducing the need for costly vehicle recalls.
SDVs will enable vehicles to become an extension of the driver’s domestic digital ecosystem with the ability to sync with other internet-of-things (IoT) products, such as digital voice assistants, becoming a part of a digital continuum for their users.
So, how do we get there?
The Roadmap to Cloud-First SDV Development
Moving to a cloud-first development environment will likely be a case of taking short-term pain for long-term gain. This will involve the industry taking three crucial steps:
1. Committing to a centralized architecture. Currently, vehicle computing is too fragmented. Modern vehicles often have over 100 separate electronic control units (ECUs), managing everything from windscreen wipers to heads-up displays. This is extremely inefficient and creates too much complexity when making OTA updates or deploying systems using a containerized approach.
The industry needs a centralized architecture that will provide a standardized common application platform for vehicles. This will enable greater collaboration between OEMs, software vendors, semiconductor designers and cloud service providers, and it will enable new digital automotive experiences.
2. Investment committed to shortening the architectural transition. The transition to cloud-first SDV development will take several years. Manufacturers cannot simply tear up their existing products and processes overnight, while jumping to a centralized architecture in one step will also be too risky.
In the short term, OEMs will likely need to move to zonal architectures first and, at some point, embrace a centralized architecture alongside this shift. This parallel approach will be more expensive, but it will reduce disruption and costs in the long term.
3. New automotive hardware with built-in redundancy and regulatory collaboration. Vehicle safety must be core to any centralized architecture. This likely will require OEMs to introduce some initial redundancy, and not just on the level of hypervisors that can isolate critical and non-critical systems. The entire centralized computing unit must be made redundant, to avoid a catastrophic breakdown in case of a hardware problem (single point of failure). This approach has been applied widely in many other functional safety-relevant industries, like aerospace or nuclear power for many decades.
Close collaboration with regulators will be key so we can create appropriate standards with adequate goals, requirements and procedures for establishing the safety of SDVs.
What Is the Industry Doing to Create a Unified Approach?
A driving force behind the industry achieving a consensus on standards is the SDV Alliance. It was created in March 2023 by four consortia, AUTOSAR (AUTomotive Open System ARchitecture), COVESA (The Connected Vehicle Systems Alliance), Eclipse SDV, and the SOAFEE (Scalable Open Architecture for Embedded Edge).
The organization’s formation demonstrates an industry-wide commitment to achieving a common approach. It is seeking to agree what defines an SDV and explore the technologies, methodologies and standards required to make development a success.
The longer the industry debates the best route to SDV development, the more likely it is that other innovators will fill the void. If this happens, traditional OEMs may struggle to close the gap on faster-moving competitors.
Agreement of an initial fundamental centralized architecture framework will be crucial if all industry players -- OEMs, Tier 1 suppliers, semiconductor designers and cloud providers -- are going to work together on projects.
OEMs must be prepared to over-invest, mindful that in the long term, this injection will lead to a smoother transition, and will ultimately prove cost-effective. OEMs will also need to accept more hardware redundancy while the industry agrees on the path way ahead.
There is evidence that the industry is ready for this fundamental change, but it will require collaboration across the sector to make this a reality. So, ask yourself, are you ready to work with others to enable this change?
About the Author
You May Also Like