Catch the replay of the Apsara Conference 2020 at this link!
On June 2020, a news about "autonomous ride-hailing vehicles" quickly made the headlines of major news websites. The news reported that Shanghai had opened the first batch of pilot projects of autonomous ride-hailing vehicles. Citizens can directly hail autonomous vehicles online at fixed stations. This news showed that autonomous vehicles have gradually developed from the early phase of research and development to the actual use.
According to the Research Report on the Development of the Autonomous Driving Industry (2019), the year 2020 will be a crucial year for the autonomous driving industry. In the Roadmap for Energy-saving and New Energy Automobile Technology officially released by the China Society of Automotive Engineers (China-SEA) in 2016, it was clearly pointed out that "by 2030, China would have 38 million fully autonomous vehicles." The roadmap clearly set the development goals for China's autonomous vehicles in three five-year periods. The year 2020 is the starting year, and the goals set for 2020 are the number of automobiles reaches 30 million and partially autonomous vehicles achieve a market share of 50%.
Autonomous driving is the most popular start-up track in 2020. Therefore, Alibaba Cloud officially released the solution that integrates data collection, transmission, storage and computing for the autonomous industry at the Apsara Conference on September 18, 2020. The solution can facilitate the implementation of new technologies and mass production for the autonomous industry, ultimately achieving real autonomous driving.
During the research and development of autonomous driving, the infrastructure is required to quickly and steadily collect and process massive data. In intelligent automobile scenarios such as AI and business of the Internet of Vehicles (IOV), data will be generated in TB level every day. If such huge amounts of data are directly written into a hard disk, the performance cannot be guaranteed, nor can the data be protected. At the same time, it is also very complicated and difficult to transfer massive amounts of data to cloud computing clusters. In the transferring process, the O&M costs are also very high. In addition, the material volume is often as high as 100 TB in daily model training scenarios. If centralized training of such materials is required, the GPU needs to repeatedly and randomly access these materials. Therefore, the file system is required to provide low-latency file access capability to speed up the training process. If this is the case, the traditional offline Network Attached Storage (NAS) suffers from single point performance bottleneck, and the capacity and performance cannot be expanded elastically, which are unable to meet GPU's requirement for low-latency file access.
To meet diverse needs of the autonomous driving industry, Alibaba Cloud tailors a data storage solution that integrates data collection, transmission, storage and computing. The AI-based operating system is the core technology for intelligent automobiles. Massive amounts of data and behavior data collected daily by sensors need to be simulated and deeply learned, which requires higher performance of storage throughput, latency and flexibility.
For data collection and transmission, Alibaba Cloud Lightning Cube can upload up to 100 TB of data to Object Storage Service (OSS) each day, with a maximum data transmission speed of 100 Gbps. In addition, the Lightning Cube applies AES256-based end-to-end encryption and Cyclic Redundancy Check (CRC) consistency to ensure data security and reliability during fast data transmission.
For data storage, Alibaba Cloud OSS provides 99.9999999999% data security and 99.995% Service Level Agreement (SLA) availability to comprehensively protect data. In addition, the file lifecycle management and data hierarchical archiving functions can automatically store data in low access frequency or archive type OSS. This simplifies operations, improves efficiency and greatly lowers data storage cost.
For data computing and analyzing, CPFS of Alibaba Cloud OSS can easily withstand performance stress and its throughput capability can be scaled up to 100 GB/s. As a result, the latency for random access to small files is reduced by eight times. In some training and deep learning scenarios, the speed of random access to small files is increased by three times. This greatly improves the efficiency of file computing and analyzing.
Alibaba Cloud's integrated data solution for autonomous is committed to helping autonomous driving enterprises improve time consumption, costs, security, and computing efficiency in terms of data acquisition, transmission, uploading, and computing.
As a leading intelligent automobile manufacturer in China, Xpeng applies Alibaba Cloud's integrated data solution for autonomous driving to store hundreds of TB of data per day. It also helps the AI system process data quickly and accelerates the training speed of driving skills of automobiles and the speed in complex road conditions The solution helps Xpeng improve the overall cooperative R&D efficiency of the autonomous driving technology by more than 40%.
Alibaba Cloud believes that cloud computing can offer the greatest value of data, and migration to the cloud can be beneficial for the development of all industries. In the future, Alibaba Cloud will continue to provide reliable solutions for all industries to accelerate their development and help enterprises achieve milestones of intelligent development.
Catch the replay of the Apsara Conference 2020 at this link!
Alibaba Clouder - April 17, 2020
Alibaba Clouder - October 15, 2018
Alex - July 9, 2020
Alibaba Clouder - June 11, 2018
Alibaba Clouder - July 16, 2020
amap_tech - August 27, 2020
ET Brain is Alibaba Cloud’s ultra-intelligent AI Platform for solving complex business and social problemsLearn More
This solution provides you with Artificial Intelligence services and allows you to build AI-powered, human-like, conversational, multilingual chatbots over omnichannel to quickly respond to your customers 24/7.Learn More
An end-to-end platform that provides various machine learning algorithms to meet your data mining and analysis requirements.Learn More
This technology can be used to predict the spread of COVID-19 and help decision makers evaluate the impact of various prevention and control measures on the development of the epidemic.Learn More
More Posts by Alibaba Clouder