On April 18, at the "Tencent Big Data Summit Forum", Tencent officially released the self-developed fourth-generation digital-intelligence fusion computing platform "Tencent Big Data-Tiangong" . "Based on the technical concept, under the important premise of ensuring data security, truly realize unmanned "autonomous driving" of trillion-level data analysis, and then promote the integration of big data and artificial intelligence technology, leading the global big data computing to enter the next stage. era.
Mei Hong, Academician of the Chinese Academy of Sciences, Wei Kai, Deputy Director of the Institute of Cloud Computing and Big Data, China Academy of Information and Communications Technology, Yuan Hong, Managing Director of Accenture Strategy and Consulting, Jiang Jie, General Manager of Tencent's Data Platform Department, Cheng Congchao, Chief Engineer of Intel Big Data, Liu Yuhong, Deputy General Manager of Tencent Data Platform Department, Tao Yangyu, Director of AI Platform of Tencent Data Platform Department and other domestic leaders in the field of big data delivered speeches. At the meeting, Liao Peng of China Minsheng Bank, Zhang Yutian of China Electric Power Research Institute, Wen Senmao of China Telecom Cloud Computing Branch, Nie Jing and Chen Peng of Tencent held a roundtable dialogue on the opportunities and challenges of privacy computing.
Jiang Jie said at the meeting: "As an important part of the new infrastructure, the big data industry will usher in a new stage of development. The fourth-generation digital-intelligence fusion computing platform released this time will open up data silos and bridge data in a safe way. Multi-party data, using a unified data processing engine to mine data value more efficiently, drive the entire data processing closed loop in an intelligent way, and lay a solid foundation for the digital and intelligent upgrade of developers, enterprises, and governments."
Secure self-developed privacy computing technology leads the world
To break down data silos and realize data collaboration, data security is the top priority. "Tencent Big Data-Tiangong" adopts self-developed privacy computing technology to provide comprehensive protection for various scenarios from machine learning to big data analysis. The decentralized architecture can avoid the risk of single-point privacy leakage. At present, "Tencent Big Data-Tiangong" can support massive data training with a scale of 100 billion, and its performance parameters are 5 times ahead of the industry. .
With this technology, Tencent's big data related team won the iDash 2020 World Privacy Computing Competition champion, leading the world in security performance. It is understood that Tencent's privacy computing technology with financial-grade security strength has been widely used in many fields such as medical care, financial risk control, and digital government affairs.
Intelligent trillion-level data analysis will realize "autonomous driving"
From offline computing, real-time computing to machine learning, big data intelligence is the only way to plan the fourth-generation digital-intelligence computing platform. Tencent is building a platform brain to promote trillion-level big data analysis to gradually realize "autonomous driving". From quickly discovering big data operation problems, proactively finding problems, and proactively solving problems, the platform brain is expected to increase data center R&D efficiency by 60%, operational efficiency by 50%, and platform service quality by 80%.
Unifying big data and artificial intelligence
The trend of big data intelligence is becoming more and more clear, and the integration of big data and artificial intelligence technology is also accelerating the digital upgrade of various industries. In view of the similar nature of the two technologies, the "Tencent Big Data-Tiangong" platform can better adapt to hardware such as CPU, GPU, NPU, and FPGA by solving the unification of big data and artificial intelligence computing frameworks. Including the construction of big data and AI basic operators, unified metadata for execution optimization, unified batch, stream, and graph computing forms to unify computing engines, and timely compilation, code generation to adapt to heterogeneous hardware.
After ten years of deep cultivation, "Tencent Big Data - Tiangong" leads the global big data wave
Jiang Jie said that Tencent itself has a massive data scale. At present, the daily access message volume of Tencent's big data platform exceeds 55 trillion, the daily real-time computing volume exceeds 65 trillion, the overall computing power of the platform exceeds 5 million cores, and the daily analysis tasks reach 15 million, Tencent Cloud has become the cloud manufacturer with the strongest computing power in China, and it is also the company with the largest daily real-time computing volume.
Since 2009, Tencent has been deeply cultivating the field of massive big data processing. The first generation big data platform relies on the Hadoop ecosystem to build a data processing platform that can stably support computing tasks at the hour/day level around the offline computing model. The platform was launched in 2016 In 2012, it won the world ranking championship; around 2012, Tencent's second-generation big data platform introduced real-time computing and processing frameworks such as Spark and Storm, so that the processing performance of the big data platform reached the millisecond level, and at that time it ranked first in the domestic real-time computing scale. A throne.
In 2015, Tencent's third-generation big data platform developed towards machine learning. Angel, the self-developed machine learning framework, became the first top AI project in China to graduate from the Linux Foundation, and promoted domestic big data processing to officially enter the era of machine learning. The fourth-generation digital-intelligence fusion computing platform "Tencent Big Data-Tiangong", which has been planned since 2020, will undoubtedly continue to lead the global big data computing on a new journey of security, unity and intelligence.
It is worth mentioning that, in order to better explore the balance between development and security in the digital age, Tencent also released the "Tencent Privacy Computing White Paper" at the summit (click the original link at the bottom of the article to get the download method of the white paper). The white paper was co-authored by Tencent Research Institute and other departments to explore privacy computing in an all-round way from the perspectives of privacy computing development background, technical system, key application industries and scenarios, data security compliance, and future development prospects. By building consensus with the industry , and jointly promote the development of privacy computing technology.
At the meeting, Tencent Big Data signed a strategic cooperation agreement with the Cloud Computing and Big Data Research Institute of the China Academy of Information and Communications Technology.
Original link: https://mp.weixin.qq.com/s/vUIgPT7pcYV50bVXk7Tw1A