The digital lifecycle revolution in a software-defined world
Dec 6, 2022 by Marek Jersak
We’re moving toward a software-defined, data-driven automotive world. Vehicle capabilities are increasing quickly in electrification, connectivity, autonomy and shared mobility. With opportunities come challenges for OEMs and Tier 1s to create offerings and business models around software and data. Add to that the new technology- and software-players who are already doing things differently on the market, and it’s obvious that traditional players need to transform to remain successful.
This article is the second in our series on software-defined vehicles (SDVs).
The challenge of transforming the automotive industry is multi-dimensional. Let’s start with customer expectations — everyone is used to digital experiences, but the question is, how can we project those into the vehicle? The digital experience of today is very personalized especially in mobile devices, internet services and smart homes; everyone wants to be seen as an individual and that’s the experience users expect in cars, too. There’s a big opportunity to reimagine the automotive experience around electrification, connectivity, autonomy, and shared mobility. Software and data are key enablers here, as functions and services transcend into industries around automotive (energy ecosystems or increasingly smart infrastructure in cities for example), giving automotive players a lot of opportunities to connect the vehicle to X devices. The most critical aspect for success, however, will be transforming business models toward recurring revenue streams built on top of software and services. To be clear, these have to be innovative, as clients are not ready to pay for something that they feel is already an integral part of the vehicle they purchased.
Opportunities clearly abound in the increasingly digital automotive world, but there are a lot of challenges too. With so many different dimensions in motion at the same time, businesses must ask themselves a series of questions about:
Industry leaders are increasingly focusing on software and data.
The goal is to develop and deploy software and the associated services largely independently of vehicle product lifecycles, keeping vehicles and services ‘fresh’. Monitoring them in the field and understanding how they’re performing is key to learning how to monetize them, independent of selling the vehicle itself. Software transcends the vehicle into the back-end and connected mobile devices — it's a big ecosystem. A software deployment factory is needed that operates highly effectively, taking updated software from the development team and driving that into vehicles over the air (OTA) at proper cadences for safety updates, regular feature updates, or novel software- and data-based services. The industry refers to the software-defined vehicle (SDV) in this context.
But how do you build an SDV? A system-thinking approach is key, breaking down the challenge into a manageable, evolutionary set of sequential and parallel steps. In our experience, software and data require an ecosystem of partners and suppliers, global standards, and suitable in-vehicle, cloud backend, and development platforms, plus seamless processes for software deployment and monitoring. It’s about looking holistically at the technologies that are evolving, capabilities being developed, the cultural/organizational change, the spirit to innovate, and the rigor to automate and ensure quality, performance and safety.
Starting with electric/electronic (E/E) architectures, the trend is a push away from domain architectures, toward zonal architectures, and then probably toward a brain-spine network. Evolving the E/E architecture allows OEMs to create layers upon which they can flexibly deploy software largely independent of the underlying hardware. Such hardware-software separation creates a very flexible and reusable system where features and functions evolve at different speeds. Achieving this requires an organizational transformation because the automotive industry is traditionally structured by function from business case to requirements, to hardware-software architectures embedded in dedicated ECUs. This prevents the level of flexibility required for SDVs.
One layer above the E/E architecture, the in-vehicle software platform is also transforming. Here we see a lot of OEM.OS projects in the industry, but there is clearly a need for a more standardized approach. Firstly, because such an approach takes a lot of effort and creates a lot of redundancy. More importantly, the most efficient, consistent, and widely adopted platform is also the most attractive for companies who are developing functions and apps — thereby enriching the ecosystem and reinforcing the pull to said platform. Therefore, OEMs and suppliers ideally need as few standardized software platforms as possible on which a large ecosystem of software companies can deploy apps and services in a very effective way. These apps are no longer confined to the vehicle themselves, but they transcend into the cloud and mobile devices. This triggers the need for a service-oriented architecture where apps and services can migrate or partially run in one place and partly another.
On top of the traditional engineering disciplines that mechanical and electrical engineers are so good at in the automotive industry, it’s vital to then add capabilities around software and data in a way so they’re run in a completely different manner: It’s really a tightly integrated and highly automated process which is walking us through different fidelities of data and phases of software development. In addition, automation needs to work on multiple layers — from digital-twin simulations to fully integrated production vehicles. These combined capabilities enable the automation of software development, and data-driven development and are at the core of the digital transformation of the automotive industry.
OEMs and their supply chain need to move away from traditional hierarchical organizations toward agile streams that work on different features and functions, combining the right people from different teams and restructuring them for the time being for the task at hand. This can be one of the most challenging parts of the transformation: Building interdisciplinary and cross-company teams, capable of working together effectively.
In the agile way of working, one of the things that happens is that different pieces of the traditional decomposition tree have different stages of maturity — you can already start to implement certain pieces of your software while other pieces haven’t even been fully specified yet. So, the whole organization is flipped onto its side — instead of a tree, it becomes agile streams that flow into sprints that the whole agile community knows how to follow. It’s very important to control and understand all the dependencies you have (both visible and invisible) to build connections here diligently and meticulously.
Luxoft, together with our parent company DXC, is a software and data company at its very core and culture. Global collaboration and flexible teams are the norm. We’re organized in a multi-dimensional matrix and work in guilds and chapters to build and scale our expertise. We work with hundreds of clients in automotive and neighboring industries, bringing the experience that the automotive industry now needs. We consistently attract the best talent and provide a fresh influx of ideas and experiences to clients. If you’d like to discuss your transformation, or if you have any other questions, please reach out to us at AutomotiveAdvisory@dxc.com.
Written with contributions from Suat Kusefoglu, Advisory Director Architectures and Ecosystems, Manish Singh Dhek, Senior Advisor Architectures and Ecosystems, Matthias Bauhammer, DXC Director Data and Analytics