Amazon Web Services accelerates deployment in the automotive industry by enabling car-cloud integration

In the car-making boom of "new forces" and cross-border enterprises, the new four modernizations have become the absolute core. As the products gain ground, the names of "star" car companies and even "star" suppliers have gradually become familiar to the public. Of course, this rapid and fierce industrial transformation does not only rely on car companies and hardware suppliers. Many cross-industry and cross-field forces have also participated in the redefinition of the concept of "car" and played a vital role. Amazon Web Services is one of them.

On July 14, Amazon Web Services announced that it will fully empower the digital transformation of the automotive industry around the three major digital scenarios of autonomous driving, vehicle networking, and software-defined cars. Through global infrastructure, integrated cloud-edge services, dedicated industry solutions and a broad partner network, it will form a professional service team for the automotive industry to enable automakers to transform to software-defined cars.

Gu Fan, General Manager of Strategic Business Development Department of Amazon Web Services Greater China, believes that the original business model of automakers focused on selling cars, and everyone is a consumer. When we have a desire to buy a car, the automakers will try every means to attract customers who want to buy a car in the next 3 to 4 months, and finally complete the value exchange between the automakers and customers through the act of buying a car. In the future, many automakers are exploring a direction, that is, to extend the point of value exchange between automakers and customers from the moment of selling a car to the entire life cycle of the user's use of the car, rather than just at the moment of selling the car.

Whether using autonomous driving, Internet of Vehicles or software-defined cars, automakers hope to interact more directly with customers and obtain more data, which will guide you on how to provide services that customers will really pay for during the vehicle life cycle. To achieve this service capability, automakers may need to consider the following three aspects in their technical route:

The first is to shift to a new centralized automotive electronic and electrical architecture, truly allowing the vehicle-side computing power to reach a very high level.

Second, we need to have a software-defined car platform based on a new electronic and electrical architecture, and do a lot of software development work across the car and the cloud, because we know that in the future only such service-oriented software will differentiate the value of car manufacturers.

Third, make better use of data. Whether it is the Internet of Vehicles or autonomous driving, they are essentially data-driven software development processes.

According to a McKinsey report, by 2040, the market for autonomous vehicles and related mobility will reach $2 trillion, which is a very large market. At the same time, the application scope of the autonomous driving industry is also expanding from passenger cars to trunk logistics and even last-mile delivery, which are all opportunities. The biggest challenge is that the maturity of autonomous driving that the industry initially promised is still a long way from the reality we see today.

As we all know, autonomous driving requires powerful computing power, machine learning and reasoning capabilities, and relies on massive amounts of data for algorithm training and simulation verification. Cloud computing's massive computing and storage resources can achieve ultra-large-scale data processing, model training and simulation clusters, thereby accelerating the implementation of autonomous driving products. Therefore, in the fierce competition in the autonomous driving track, the technology and capabilities of cloud service providers in these areas have become the key to helping car companies win.

Data runs through every scenario of autonomous driving research and development. With the autonomous driving data lake as the center, Amazon Web Services helps automakers build an end-to-end autonomous driving data closed loop. With the help of Amazon Simple Storage Service (Amazon  S3 , cloud object storage service) with nearly unlimited scalability to build an autonomous driving data lake, data collection, data management and analysis, data annotation, model and algorithm development, simulation verification, map development, DevOps and MLOps can be realized more easily by automakers to develop, test and apply the entire process of autonomous driving.

Amazon Web Services' machine learning service Amazon SageMaker and Amazon Elastic Compute Cloud (Amazon EC2) elastic computing service have a leading edge in the extremely difficult and challenging data labeling, model and algorithm development, and simulation verification stages of autonomous driving technology. In the data labeling stage, Amazon SageMaker Ground Truth can easily complete the automated labeling of various vehicle, scene, and user data, and create high-quality data sets that meet the requirements for training machine learning models. In the simulation verification stage, especially for ultra-large-scale simulations, Amazon EC2's Spot instances of elastic computing services can provide low-cost computing power at the level of millions of vCPUs, saving up to 90% of costs and accelerating the development and implementation of autonomous driving technology.

At the same time, in the process of building Internet of Vehicles services, as the business expands, car companies need to deploy globally in a unified manner, achieve high availability and secure connections, and meet the security and compliance requirements of global operations; in order to adapt to the elastic needs of Internet of Vehicles business and reduce the difficulty of operation and maintenance, car companies need to use modern services such as global unified architecture, microservices and serverless computing to build a flexible and agile Internet of Vehicles architecture. In addition, car companies also need to select global Internet of Vehicles partners, connect richer service ecosystems and content to the Internet of Vehicles, build a comprehensive service system, and provide value-added services based on massive data to enhance customer experience.

This is exactly where Amazon Web Services' advantage lies: the infrastructure of 84 availability zones in 26 geographic regions around the world allows automakers to easily build Internet of Vehicles platforms in various geographic regions, bringing higher stability and lower latency. The security and compliance of Internet of Vehicles data has always been a top priority. Amazon Web Services supports 98 security standards and compliance certifications, almost meeting the security and compliance needs of customers in all regions of the world. Amazon Web Services' Amazon Lambda serverless computing and microservices help automakers build a modern, serverless, elastic and agile Internet of Vehicles architecture, which not only meets the elastic needs of Internet of Vehicles business, but also helps them save costs and reduce the difficulty of operation and maintenance.

Based on the Internet of Vehicles data lake, car companies can also use data analysis and machine learning technologies to deeply explore the value of data and provide more and more data value-added services. For example, by using Amazon SageMaker, a machine learning service of Amazon Web Services, combined with charging and battery management data, it is possible to predict battery failures in electric vehicles and improve user satisfaction. Amazon Web Services' global partner network can provide customers with various Internet of Vehicles overall solutions and platform services such as the Internet of Things, data analysis, interconnected mobility, edge computing, charging management and services, and map navigation. In addition, car companies can rely on Amazon's global business system to integrate Amazon's intelligent personal assistant Amazon Alexa to provide users with in-car voice interaction and easily obtain rich Amazon content, including Amazon Instant Video, Amazon Audible, Amazon Music, Amazon FireTV, etc., to meet customers' personalized needs.

Amazon Web Services also has relevant arrangements in enabling car companies to realize software-defined cars: First, it participates in the scalable open architecture SOAFEE initiated by ARM, strengthens the parity of the car-cloud environment between the ARM architecture processor Amazon Graviton and processor manufacturers on the vehicle side and the edge of the Internet of Vehicles, and unifies the vehicle-side software development and cloud software development; second, it cooperates with BlackBerry to build BlackBerry IVY to help it build an intelligent car software development and data platform, read vehicle sensor data in a consistent and secure way, and ensure that data can be processed and insights can be obtained both inside the vehicle and in the cloud; third, it cooperates with Continental to help Continental develop a full-stack software-defined car development platform Continental Automotive Edge (CAEdge), which provides a high-performance, highly automated development environment to support all processes of autonomous driving system development; fourth, it realizes an end-to-end data closed loop integrating cloud, edge and end, and provides a full range of cloud services; fifth, it builds Amazon IoT specifically for the automotive industry FleetWise service collects and converts vehicle data in real time, economically and efficiently, and transmits it to the cloud. It uses data analysis and machine learning services for data mining, and improves the quality, safety and autonomous driving capabilities of vehicles based on data insights. Sixth, it will form an Amazon Cloud Technology professional service team to provide software-defined vehicle solutions, build a multi-tenant, highly automated, vehicle-cloud integrated, data-driven, secure and compliant software collaborative development and testing environment, and help automotive companies quickly build a software-defined vehicle development environment and capabilities.