In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-03-26 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > IT Information >
Share
Shulou(Shulou.com)11/24 Report--
The 2023 Open data Center Conference, hosted by the Open data Center Committee (ODCC), was held in Beijing from September 13 to 14. Zhang Binghua, director of the ODCC technical expert group and CTO of Qinhuai data Group, attended the main forum of the conference and delivered a keynote speech entitled "embracing the AIGC era and Innovation leading the Future", and discussed in depth the challenges and opportunities that the generative artificial intelligence (AIGC) era brings to the data center industry, as well as the five capabilities that the data center infrastructure must possess. And share cutting-edge insights on industrial change and full-stack solutions to industry challenges.
The surging AIGC and rising energy efficiency requirements are the two core variables that affect the future development trend of the data center. Zhang Binghua said that with the exponential growth of mainstream CPU, GPU power consumption and server node power consumption, there is an urgent need for data centers to carry out technological innovation and model change, with the five capabilities of large-scale, high computing power, high flexibility, high energy efficiency and intelligence, in order to meet the challenges of computing power supply and energy consumption brought about by artificial intelligence models. The five abilities are as follows:
Large scale: a single AZ (available partition) needs to be able to provide 1 million and 50 million calories of GPU large-scale computing and storage resources to meet the data and computing needs of artificial intelligence applications. In the future, the IT capacity of a single data center will exceed that of 30MW.
High computing power: cabinet power density up to 21kW or higher, with high-performance computing and data processing capabilities to support complex artificial intelligence algorithm and model training
High flexibility: distributed, fully prefabricated modular design, air-cooled, liquid-cooled and other compatible architectures, support multiple power density cabinet scenarios, and flexibly adjust computing and storage resources to meet the fluctuating needs of AI applications in different time periods.
High energy efficiency: minimalist power supply, extreme cooling, making full use of natural resources, adopting integrated data center full-stack solution, reducing data center energy consumption and improving energy efficiency by optimizing design and operation strategy
Intelligence: with the help of AI and machine learning technology, realize the automatic management and optimization of data center, realize intelligent monitoring, intelligent prediction, intelligent heat dissipation, intelligent fault location, and improve operation efficiency.
Zhang Binghua, CTO of Qinhuai data Group
Qinhuai data foresee the development trend of computing power, and successfully create a "rock" module data center architecture through continuous technology research and development and innovation. The architecture consists of six systems and 24 product modules, aiming at advanced technology, flexibility, product prefabrication and on-demand construction, covering full-cycle business scenarios from building models to power supply and cooling. Based on this framework, Qinhuai data recently released a full-stack solution for AIGC next-generation data centers, including "Xuantie" intelligent power module 3.0, "Xuanbing" magnetic levitation phase change cooling system, large-scale fully prefabricated data center technology, and Kunpeng IDC operation platform, providing a new technical framework and evolution route for data centers in the big model era.
In addition, at the scene of the conference, the white paper "Standardization and Technical Optimization of Cold Plate liquid cooling Project" and "White Paper on the Development of Source Network, load and Storage Integration under the background of New Power system" compiled by Qinhuai data was officially released. Among them, the white paper "Standardization and Technical Optimization of Cold Plate liquid cooling Engineering" introduces in detail the technical characteristics and standardization development direction of cold plate liquid cooling, and puts forward some suggestions on the establishment of standards for cold plate components, quick joints, cooling capacity distribution unit CDU and other products, which plays a positive role in promoting the full landing of cold plate liquid cooling technology in the industry.
Wang Shun, Director of Product Development and Planning of Qinhuai data Group
Wang Shun, head of the project and director of product development and planning of Qinhuai data Group, said when interpreting the white paper that cold plate liquid cooling was mainly customized in the past, which was not conducive to the large-scale promotion of products. It is hoped that through the technical popularization of this white paper, we can promote the standardization of cold plate liquid cooling products, improve the compatibility of cold plate liquid cooling, accelerate the promotion and application of cold plate liquid cooling system, reduce product cost and promote large-scale application. The white paper also introduces the latest development direction of liquid contact materials, and explains in detail the most advanced cold plate liquid cooling technologies, such as negative pressure liquid cooling, phase change liquid cooling and gas-liquid mixing technology.
Qinhuai data innovation and practice are closely combined, and the standard architecture system of cold plate liquid cooling has been established to realize the large-scale deployment of high-density liquid-cooled liquid cooling of 20~30kW and above. In the actual operation, the annual PUE of the data center using this technology is as low as 1.15 and Wue is as low as 0.2, achieving a win-win situation of energy saving and water saving.
In the face of the challenges of the AIGC era, data centers must carry out in-depth technological change and model innovation. Zhang Binghua said that Qinhuai data will actively embrace this change and use full-stack solutions to help customers build more efficient, greener and more competitive data centers.
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
Views: 0
*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.
Continue with the installation of the previous hadoop.First, install zookooper1. Decompress zookoope
"Every 5-10 years, there's a rare product, a really special, very unusual product that's the most un
© 2024 shulou.com SLNews company. All rights reserved.