Smart cars are setting off a new wave, cars from the mechanical device of transportation to the software-defined intelligent mobile space, the change of attributes has brought about changes in the core of competition, from product performance, functional competition, to product life cycle scenario, personalized experience and service competition, intelligent cars into more IT technology and Internet concepts.
At this moment, auto people with mechanical backgrounds will inevitably be confused about the future of the car, and IT personnel who have participated in the transformation of the PC Internet and mobile Internet are full of familiarity with the wave of intelligent car changes. From the perspective of IT, we combine the evolution of smart phones to judge the development trend of smart cars.
Part I: Analogy with smartphones, smart cars tend to converge, but the process is confusing
The development of smartphones:
Step1: Interactive transformative keyboard + small screen multi-touch + high-definition large screen, open a new interactive mode
Step2: Architecture upgrade For hardware function development, software and hardware separation software definition are laid to lay the foundation for ecological applications
Step3: The evolution of ecology is based on new interaction methods + OS+ architecture, and ecological applications have begun to explode, generating multiple tens of billions and hundreds of billions of applications
For smart cars, a similar change is taking place:
Step1: Interactive transformation instrument + button voice + large-screen multimodal interaction
Step2: Upgrade of architecture Distributed E/E architecture Domain Centralized EEA +SOA+SOC Software Definition + Data Driven
Step3: The evolution of ecology is based on spatial interaction technology and spatial OS, and the automotive application ecology has begun to explode
Compared with smart phones, smart cars not only have more complex software, hardware, and power systems, but also have more spatial attributes, and there will be ecological prosperity before the development of space interaction technology and space operating systems of smart cars. Tesla is developing an app store for smart cars, which may promote the rapid development of the smart car application ecosystem.
We can make a clearer judgment on the general trend of smart cars, but based on the current situation, it is full of "uncertainty", what is the final form of spatial interaction technology and spatial OS? Who will become IOS and Android in the era of smart cars? When will the application ecosystem break out? Now we don't know.
Part II: The "determined" trend in the evolution of smart cars
The evolution of smart cars is full of confusion, but the pursuit of higher efficiency and lower costs is an inevitable trend. Today, when smart cars are rapidly evolving, there are several major challenges:
Challenge 1: After a large amount of data is stored in the data center, java, Spark, Python and other languages need to be used in the cloud to achieve modeling and analysis, which is a high threshold for automotive engineers;
Challenge 2: It is difficult and inefficient to deploy and iterate functions in the embedded environment at the vehicle end;
Challenge 3: Cloud (distributed) and vehicle-side (embedded) heterogeneity, as well as heterogeneity between different models, lead to fragmentation of the Internet of Vehicles, making OEM and Tier1 business development and product iteration costly and inefficient.
Faced with such challenges, the IT field has redefined efficiency and cost through Codeless, Serverless, and Middleware, accelerating the change of times, and the same applies to today's automotive industry.
Codeless: Democratization of technology
In 1980, IBM's Rapid Application Development Tool (RAD) was given a new name: Low Code, thus making the concept of low code first available to the general public. Its full definition is "a technology and tool that leverages little or no need to write code to quickly develop applications and can be quickly configured and deployed."
Today, in the computer industry, low code has formed a complete ecosystem with its characteristics of "empowering business personnel to develop independently".
In the automotive industry, the earliest engineers in the desktop level system processing data, with Datalogger record data, copied to the computer, through MATLAB plus VECTOR CANoe to analyze data, when the scale of data increases, can only be stored in the data center, business personnel often need to rely on IT staff through Spark, Java, Scala, Python and other processing data, resulting in high labor and communication costs, and extremely low efficiency.
vStudio is a graphical, low-code modeling and analysis tool developed for the automotive industry, covering almost all the operators used in automobile development, and has proprietary modules such as signal filtering, window functions, and machine learning.
Using The AdvancedData vehicle-to-transport solution, vStudio's algorithms can be deployed to the cloud and on the vehicle side with one click.
2. Virtualization & Microservices: Serverless
In 2012, the word Serverless first appeared, after Microsoft, Google, IBM have launched their own Serverless products, in 2019, Serverless was called by Gartner the most potential cloud computing technology development direction, and was identified as an inevitable development trend.
The essence of Serverless is not that there is no need for a server, but that the server is fully hosted to the platform, the user focuses on business and development, and the entire compilation, build, packaging, and deployment can be automatically completed by the Serverless platform. Serverless can also configure computing resources on demand, which not only ensures the normal operation of the business, but also ensures high cost performance, and fully empowers data-driven services.
In the automotive industry, the cloud is a high-dimensional world, using Java, Python, SPARK and other high-level languages and tools, its development of the model to deploy to the car-side embedded environment, in the face of heterogeneous environment of the algorithm migration, dependent library migration and implementation, communication matrix adjustment and other difficulties, so that the completion of an algorithm model of the car-side successful deployment, the cycle is often in months or even years of computing, the efficiency is extremely low.
For intelligent cars, the closed-loop capability of vehicle-cloud data is the core, so the core challenge of Serverless is the integrated computing architecture of vehicle-cloud, and the algorithm model developed by the cloud can be deployed in seconds at the end of the car. After years of research and development, Zhixi Huitong has overcome more than 100 technical difficulties in the whole chain of Cheyun, the first in the industry to realize the real sense of the car-cloud integrated computing architecture, business personnel can focus on business logic, without caring about the underlying implementation process, through the vStudio developed algorithm model in seconds can be issued in the car side, more efficient real-time computing, in a low-cost way to achieve data seconds to the desktop, compared to traditional methods, comprehensive data cost savings of more than 85%, efficiency increased by a hundred times.
3. Complexity packaging: Middleware
We introduced Codeless and Serverless earlier, and the final implementation vector is middleware software across the edge and the cloud. Middleware is a large category of basic software, in the middle of the operating system and applications, the role is to shield the complexity of the underlying operating system, to achieve compatibility and openness to various hardware platforms, operating systems, network database products and Client terminals, greatly improving development efficiency.
In the automotive industry, one of the goals of SOA and AUTOSAR is to decouple hardware and software, atomize the underlying functions for the invocation of upper-level services, and some OEMs and head suppliers even try to rewrite AUTOSAR AP completely, which also shows from another perspective that the automotive industry currently lacks effective middleware to shield the complexity and heterogeneity of systems and platforms.
The automotive middleware of The Auto-Edge includes the edge computing engine vCompute and the timing signal database vData, which are developed specifically for automotive and can:
Shielding vehicle-cloud heterogeneity (cloud-based distribution, vehicle-side embedded)
Masking of vehicle-side complexity (communication matrix, C++ programming, dependency libraries)
Shielding vehicle heterogeneity (hardware, software, communications, etc.)
ExD vehicle-cloud integrated computing architecture supports the cloud algorithm model to be sent to the car side in seconds, enabling car companies to efficiently realize function iteration, performance iteration and data acquisition rule configuration, realize efficient and flexible data closed-loop capabilities, and meet the evolutionary needs of intelligent vehicles. Since EXD shields model heterogeneity, the algorithm model developed by the supplier can be efficiently deployed across models and OEMs, eliminating the need to repeat the underlying development and adaptation work for different models.
Codeless, Serverless, and Middleware are successful experiences in the development of the IT industry, redefining efficiency and cost, and promoting the rapid development of the Internet era and the mobile Internet era, especially the rapid subversion of smart phones to feature phones.
Nowadays, entering the era of a new generation of Internet of Things with smart cars as the carrier, OEMs have begun to dominate more software development, hundreds of millions of lines of new code have appeared in cars, the complexity of smart car software has risen sharply, the vehicle research and development cycle has been greatly shortened, and smart cars need to achieve rapid iteration based on data after they are listed, etc. In the face of these challenges, the automotive industry is in urgent need of new technologies and development methods, codeless, Serverless, Middleware will burst into new energy.
Relying on its core underlying capabilities and first-mover advantages, the EXD vehicle-cloud integrated computing solution developed based on the concepts of Codeless, Serverless and Middleware has produced in-depth cooperation with a number of leading car companies that develop SOA models, including Chinese Express Gaohe (HiphiX and HiphiZ), SAIC Passenger Cars (R and Roewe), SAIC Zero Bundle (Zhiji), FAW Group (Hongqi), Dongfeng Lantu, etc. The ExceedData vehicle-cloud integrated computing solution has become the standard for SOA models, laying the foundation for the era of software-defined smart cars!