For data centers, cooling systems consume the most power, and so reducing the energy consumption of cooling systems represents a first crucial step towards a greener data center. While innovation and technology has brought much advancement in the use of natural cooling sources, cooling systems need smart brains that are capable of smart adjustments and on-demand cooling with the constant changes in data center loads and their environments.
For example, China Unicom’s Henan branch (China Unicom Henan) deployed Huawei’s iCooling@AI solution which integrates big data and AI, enabling data centers to learn to save power and automatically optimize their power efficiency, improving data centers’ PUE by 8 to 15 percent.
Located in Henan’s Zhengzhou, the Central Plains Data Center is one of China Unicom’s 12 planned, ultra-large, national data centers. It’s also the only core data center in central China to be built to the T3+/T4 standard. At the start of this strategic partnership project between China Unicom and the Henan provincial government, China Unicom Henan put forward rigid energy consumption requirements for the data center. Its priority was to achieve reliability and a PUE that would be first-class both in China and globally.
In the days before the iCooling@AI solution was adopted for optimizing power efficiency, data center cooling systems were mainly configured manually. This made it difficult to achieve satisfactory results, as the load and environment were constantly changing.
In contrast, iCooling@AI can efficiently and accurately collect all data relating to a data center’s power efficiency. It then uses the deep neural network for modelling and accurately compares the created models with the data center’s operating status, which is optimized every hour. As a data center’s load increases, the cooling mode changes and the AI becomes more capable of learning. This leads to constant improvements in power saving that significantly reduce power consumption and waste.
With the adoption of the iCooling@AI solution, components of a data center cooling system will collaborate intelligently and operate efficiently, optimizing the entire system’s power efficiency.
Source: Huawei
