2023，算力霸屏，数据中心出现往往是作为关联词，这一情况下，“算力中心”的概念被更多的提及。
一、核心作用变化，数据中心演变向算力中心
当前，我国数字化、智能化进程不断加快，算力需求不断提升。今年，伴随生成式人工智能的火热，算力的需求量更是快速增长。
数据显示，今年我国生成式人工智能市场规模有望突破10万亿元，生成式人工智能正在加速渗透制造业、零售业、电信行业和医疗健康等四大行业。这带来了海量的算力需求。据数据，2023年我国人工智能算力市场规模将达到664亿元，同比增长82.5%。
数据中心由此开始新的演进，第十八届中国IDC产业年度大典（IDCC2023）上，中国信通院云计算与大数据研究所所长何宝宏表示，应用现代化的概念会随着算力的发展出现在前端，早期是机房，后来叫存储数据中心，近年来，数据中心的作用正在从以存储为中心转向以计算为中心，提供算力资源。因此，行业以后应该用“算力中心”这个词取代“数据中心”，以更好的强调核心使命是算力的供给而不是存储能力的供给。
同时，这种演进才刚刚开始。何宝宏指出，算力中心最初的核心目标是如何生产出更多算力，生产是第一要务，生产到一定阶段，摩尔定律减速后只是生产还不够；如何更好组织好算力、管理好算力、调度好算力成为重点，需要更好的进行算力调度；接下来要研究如何更好地让算力发挥作用，释放算力的价值。
二、专用技术发展，数据中心走向原生时代
数据中心的发展，从核心作用、概念角度讲，其向“算力中心”演化，这一过程中，数据中心技术、产品等也有着新的发展变化。
何宝宏表示，技术融入产业约分为三个阶段，第一阶段是简单使用，将技术作为工具来应用；第二阶段是转型，让技术释放出更好的生产力；第三阶段是原生，专用于数据中心的技术发展与成熟，并扩张到更多的领域中。
数据中心同样如此，何宝宏指出：当前数据中心仍然是传统技术与创新技术混合的存在，随着技术的发展可以判断再过大概七年或十年，数据中心将进入原生时代，这些技术是土生土长的，专用于数据中心这个场景，下一步还会扩到场景之外。
何宝宏强调，当前比较有代表性的技术是近两年兴起的液冷技术，其主要是用于数据中心，市场体量正快速增长。液冷是由数据中心的原生场景持续催熟一个产业，随着数据中心的广泛使用，必然从数据中心外溢出去，慢慢走向原生。
据《中国液冷数据中心市场深度研究报告》，中国液冷数据中心市场快速发展。过去几年保持40%的复合增长率。2022年，液冷应用爆发，市场规模达到100.5亿元。预计在人工智能等新技术普及应用带动下，至2027年，中国液冷数据中心市场规模将突破千亿大关。
伴随数字化、智能化需求的提升，数据中心在今天也来到了新的发展阶段，其计算能力成为中心，开始向算力中心演变，其技术能力逐步发展，逐步来到原生时代。这是数据中心演进的方向，也是数据中心实现突破的契机，在这一情况下，数据中心将更好的满足算力需求，实现市场与产业的发展，最终为数字化提供更好的支撑。
翻译：
Is “data center” really going to be replaced by “computing center”?
In 2023, the data center is often used as a related word, in this case, the concept of “computing power center” is mentioned more.
First, the core role changes, and the data center evolves into the computing power center
At present, the process of digitalization and intelligence in China is accelerating, and the demand for computing power is constantly improving. This year, with the hot generation of artificial intelligence, the demand for computing power is growing rapidly.
Data show that the size of China’s generative artificial intelligence market is expected to exceed 10 trillion yuan this year, and generative artificial intelligence is accelerating its penetration into four major industries such as manufacturing, retail, telecommunications and medical health. This creates a huge demand for computing power. According to data, in 2023, China’s artificial intelligence computing power market will reach 66.4 billion yuan, an increase of 82.5%.
Data center thus began a new evolution, the 18th China IDC Industry annual ceremony (IDCC2023), China ICT Institute of cloud computing and Big data Institute director He Baohong said that the concept of application modernization will appear in the front end with the development of computing power, early is the computer room, later called storage data center, in recent years, The role of data centers is shifting from storage to computing to provide computing resources. Therefore, the industry should replace the term “data center” with the term “computing power center” in the future to better emphasize that the core mission is the supply of computing power rather than the supply of storage capacity.
And this evolution has only just begun. He Baohong pointed out that the initial core goal of the computing center is how to produce more computing power, production is the first priority, production to a certain stage, Moore’s law deceleration after production is not enough; How to better organize, manage and schedule the computing power has become the focus, which requires better computing power scheduling; Next, we should study how to better let the computing power play a role and release the value of the computing power.
Second, the development of dedicated technology, data centers to the original era
The development of data centers, from the core role, concept point of view, its evolution to the “computing power center”, in this process, data center technology, products, etc., also have new development changes.
He Baohong said that technology integration into the industry is about three stages, the first stage is simple use, the technology as a tool to apply; The second stage is transformation, allowing technology to unlock better productivity. The third stage is native, dedicated to the development and maturity of data center technology and expansion into more areas.
The same is true of the data center, He Baohong pointed out: the current data center is still a mixture of traditional technology and innovative technology, with the development of technology can be judged in about seven years or ten years, the data center will enter the native era, these technologies are homegrown, dedicated to the data center scene, the next step will expand to outside the scene.
He Baohong stressed that the current more representative technology is the liquid cooling technology that has risen in the past two years, which is mainly used in data centers, and the market volume is growing rapidly. Liquid cooling is an industry that continues to be accelerated by the native scene of the data center, and with the widespread use of data centers, it is bound to spill out from the data center and slowly move toward the native.
According to the “China liquid cooling data center market depth Research Report”, China’s liquid cooling data center market is developing rapidly. It has maintained a compound growth rate of 40% in the past few years. In 2022, liquid cooling applications broke out, and the market size reached 10.05 billion yuan. It is expected that driven by the popularization and application of new technologies such as artificial intelligence, by 2027, the scale of China’s liquid cooled data center market will exceed 100 billion mark.
With the improvement of digital and intelligent needs, data centers have also come to a new stage of development today, and their computing power has become the center, and they have begun to evolve into the computing power center, and their technical capabilities have gradually developed and gradually come to the native era. This is the direction of the evolution of the data center, and it is also an opportunity for the data center to achieve a breakthrough, in this case, the data center will better meet the needs of computing power, achieve the development of the market and industry, and ultimately provide better support for digitalization.
💬 评论 (0)