Architecture for the Automotive Cloud

What is the automotive cloud?

The automotive industry is going through a major transformation, being on a path to recovery from the slowdown in the pandemic years and with the advancements in the connected technology space. The automobile industry has been guided by the acronym CASE (connected, autonomous, shared, electric) since early 2018. There is a recognised need for an automotive cloud that will help reduce the costs of massive compute and data storage requirements, and will bring in new ways of working, tools, and technologies that are enabled by cloud. The demands have led to major shift in customer preference and expectations from OEMs.

Three key imperatives that support CASE and will define the industry future are: 

• Development of sustainable products and driving new revenues through a push towards electric vehicles. According to the International Energy Agency (IEA), the global EV stock will reach almost 70 million vehicles in 2025 and 230 million vehicles in 2030 (excluding two/three-wheelers). EV stock share in 2030 will reach 12%.

• Enhanced car sales through connected infotainment providing commerce and personalisation. In larger vehicles and fleet industries, there is a higher focus on driver safety and evaluation of the driver’s physical condition which needs camera data as well as vehicle sensor data.

• Thirdly, in order to maintain sustainable operating supply chains there needs to be intelligent manufacturing and connected supply chains running to optimum efficiency with tracking via RFID and IoT devices. In Aug 2021,[2] Gartner estimated that the enterprise and automotive IoT platform market will represent a USD 11.3 billion opportunity in 2025, representing a 33% CAGR from 2020. 

All of these imperatives require dealing with large amounts of data that needs to be processed quickly at the source – thus leading to edge and fog computing solutions – in addition to transporting this data for processing back to the cloud. Thus a lot of OEMs are looking to optimise their cloud strategy and make it the foundation of their R&D, manufacturing, sales, and servicing models.

Recent research shows half of under-35s aim to use public transport less often and take their cars more often in the future, and 44% will make less use of ride-hailing due to health and safety concerns. This is a shift from their earlier stance which was not focused on buying automobiles. A majority of this group have never owned a car and hence the ease of the buying process through digital channels, the ability to evaluate the various choices, as well as the car features are going to be key factors. The auto industry will need to extend the buying and after care process to enable dealer discounts, ancillary OEMs, and tie-ups with telecom, media, retail, and value-added service providers in order to attract their target customers and grab the market share.

In 2020 and 2021, many dealerships started selling their vehicles directly on their websites/mobile apps by allowing buyers to complete their entire sales transaction online. Companies like Vroom and Carvana were some of the first to start focusing on online selling of cars and allowing shoppers to choose their car, apply for the financing process, and have their car delivered to their front door.

How will the automotive cloud solve these problems?

With the advancements in virtual reality (VR) and augmented reality (AR), there is a play for Automotive OEMs to build experiences in the metaverse. It’s only a matter of time before dealers jump on board and develop a metaverse dealership in which users can view inventory, check vehicles, and complete the sales transaction all while wearing a VR headset in the comfort of their office or home. Gartner expects dedicated AR clouds to be formed for each sector to help synergise the content with the underlying cloud infrastructure and remove the requirements of local hardware.

There are many use cases for edge computing within the vehicle. The capability to take data from the nearly hundreds of sensors, compute and take real-time decisions from the data available onboard instantaneously, especially around auto braking, and proximity detection to other vehicles and pedestrians have already been embedded within automobiles. The adoption is already increasing and has potential for many future use cases focused not only on the driver but also the passengers, for example, real time personalised offers when the car is moving past one’s favorite store or restaurant. Connectivity to media streaming apps will allow personalisation of the offer by combining the data available within the vehicle OEM as well as the streaming app tie-up. Adding to this the map data from and monitoring for traffic congestion will allow the autonomous vehicle to reroute its path. Edge will also allow better security monitoring and infrastructure to reduce theft and allow for better insurance policy determination.

What are the major manufacturers doing?

VW has invested in making the largest automobile cloud, and other manufacturers are following suit.

Toyota has come together with leading IT hardware and software providers to form the AECC (Automotive Edge Computing Consortium) to develop use cases and reference architectures for connected automobiles. As per their forecast, there will be 100 million connected vehicles by 2025, with a data volume of 100 petabytes per month at a very conservative estimate of 1 GB per month per car.

Hence there is a need for improved cloud and network infrastructure to support the automotive cloud. In the current V2X (vehicle to everything) platforms, the volume of data that is transferred to a central cloud is too high, and new communications systems as well as distributed cloud resources are required to architect the automotive cloud. The cloud will also be a repository for software updates that are being delivered OTA (over-the-air), especially for electric vehicles. The core concepts of cloud will remain in the architecture of the automotive cloud – security, along with providing infrastructure for compute, storage, and networking. Some of the processing will not be feasible to be transferred over the network to the cloud, hence this is an important edge computing use case.

What are the use cases?

There are multiple opportunities for different stakeholders within the industry:

1. OEMs are producing vehicles with high-quality sensors such as LiDAR, radar, cameras, and ultrasonic, which can generate richer telematics data. Automotive semiconductors are now seeing a strong rebound, and IDC expects that Auto ICs will be the fastest-growing industry segment over the next five years, with a CAGR of 10.6%. By 2025, automotive semiconductors will reach USD 65.4 billion in revenue, accounting for 10% of the total semiconductor industry. They are also focusing on supply chain for ancillaries. For example, the Capgemini Intelligent Assistant for Automotive app (CIA4AUTO), built on top of SAP S/4HANA, connects SAP data such as transactions, postings, events, or lack of events, to detect anomalies and indicate potential issues across the supply chain. Variances in transactions trigger an alert to the appropriate employees for corrective action.
2. Dealer focus – The dealers need data to propose personalised offers and insurance based on user interests and their purchase history. Also for marketing and sales, improved customer engagement, driver and passenger focus – infotainment, driver comfort, ecommerce – taking care of individual needs. OEMs can put their data to better use and provide services such as remote troubleshooting.
3. Planner and administration focus – There is a need for city planners to monitor vehicle traffic 24/7, identifying and alerting travelers of issues more efficiently. Also, as electric vehicles rise, the need for city planners to place electric chargers at the optimum points as well as plan dedicated lanes for electric cars and buses.
4. Services and experiences – OEMs are focusing on the development of solutions leveraging IoT, AI, data, software-defined vehicles, among other technologies to enhance services and experiences for their customers. For instance, in February 2022, Jaguar Land Rover partnered with NVIDIA to jointly develop and deliver automated driving systems and AI-enabled services and experiences for its customers. Starting in 2025, all new Jaguar and Land Rover vehicles will be built on the NVIDIA DRIVE software-defined platform – delivering a wide spectrum of active safety, automated driving and parking systems, as well as driver assistance systems. Inside the vehicle, the system will deliver AI features, including driver and occupant monitoring as well as advanced visualisation of the vehicle’s environment. This full-stack solution is based on NVIDIA DRIVE Hyperion, which features DRIVE Orin centralised AV computers; DRIVE AV and DRIVE IX software; safety, security, and networking systems; as well as surround sensors. DRIVE Orin is the AI brain of the car and runs the Jaguar Land Rover Operating System, while DRIVE Hyperion is the central nervous system.

It makes sense for the designers of automotive cloud to build capabilities for these different user groups.

The typical building blocks of an automobile cloud platform include:

• Secure communication for vehicle-to-cloud connectivity
• Data ingestion for a variety of structured and unstructured data sources – this will be process as well as transactional data
• Data migration and cleansing through cloud-native techniques which can scale as per the volume
• Fast and scalable analytics with dashboard visualisations to detect anomalies, and threat modelling including cabin simulations of various scenarios
• Data models and AI to trigger workflows and notifications such as equipment maintenance, driver alerts, and route notifications and provide location-based services
• Secure data storage with encrypted data at rest as well as in transit
• Will need to have a zero trust architecture because of multiple devices and entry points that assumes every device and connection can be breached and hence needs verification and validation
• Portability between on-premises and hyperscalers – hence suited for a hybrid cloud architecture

In closing

In summary, it is not surprising the leading hyperscalers are already engaged with multiple customers building out automotive cloud components. AWS has partnered with Toyota to build a mobility services platform. Microsoft has launched their own Connected Vehicle Platform, which in addition to the common use cases has focused on regulatory compliance and ethics and is attached to the Open Mobility Foundation. Google has developed Android for Auto in order to allow driver-centric apps along with their Android Infotainment platform, which is provided as a standard feature by various OEMs.

The challenge is always that of the right professionals to design and run these clouds. As a global leader in technology, Capgemini is proud to provide thought leadership that helps our clients in the automobile industry to be at the forefront of the trends that will drive the industry for the next 15-20 years.

Capgemini recently launched TechnoVision for Automotive 2022, the industry playbook. It covers a multitude of disruptions that are shaking the automotive industry. Customer expectations are changing. Products are evolving. Ecosystems are growing. Technology continues to redefine drivers’ relationships to their cars. And cultural disruptions reflect the need for manufacturing companies to embrace the very different world of software development. TechnoVision for Automotive 2022 dives into each of these areas with up-to-date details on a host of developments in the automotive world.

Capgemini helps automobile industry players get into the new-age mobility market faster through our specialised offerings for connected vehicles as part of our “Digital Cloud Platform”.

Please watch this video to learn more: