PUE | 電力使用效率

  • 是什麼?
    PUE(Power Usage Effectiveness,電力使用效率)是計算資料中心節能省電的標準,計算的方式是資料中心的「總用電量」與「供應IT設備的電量」之比。PUE值越低,代表機房空調冷卻時所耗的電力就會更少,理想的 PUE 比率為 1.0:100%,意即數據中心所提供的電完全用於電腦運算。然而事實上,機房內卻有大量的電力被用於散熱(機器高速運轉時會產生大量熱能,若沒有相應的冷卻系統,將導致機器因過熱而停擺)、照明和其他耗電的開銷,也因此統機房的 PUE 比率始終大於 1.0。

  • 為何需要? 誰會用到?
    維持大型伺服器每日運作的費用(如電費)相當可觀,長年累月下來,甚至可能比當初建立機房設備、購買硬體的初始成本更大。 因此,企業不斷思考如何降低物理伺服器及散熱設施所需的用電量,而其中高耗電的傳統冷卻系統(如冷氣空調),即是降低PUE值的重點項目。

    降低機房 PUE 值有許多方法,如:捨棄傳統風扇冷卻,改將伺服器本身直接浸泡非導電流體中,此種液態冷卻裝置,能有效地減少(或根本不再需要)額外的空調成本。市面上許多大型企業,也透過直接將數據中心設置於寒冷的氣候中來降低 PUE 值,北歐、俄羅斯或阿拉斯加都擁有良好的環境條件。微軟甚至直接將資料伺服器沉入海底,嘗試更好的冷卻效果。

  • 技嘉的特色
    GIGABYTE 與閉環液體冷卻系統(如 CoolIT 和 Asetek)以及浸入式冷卻系統(如 Allied Control 和 Submer)的供應商建立了緊密的合作關係,以提供各種不同的解決方案,幫助客戶降低 PUE值。 通過與這些公司合作,GIGABYTE 可以提供經過修改,測試和驗證的服務器,以便與這些不同的解決方案配合使用,從而實現快速,輕鬆的部署。

    GIGABYTE Servers Become Part of a European Space Agency Data Center

    As environmental protection is becoming a pressing issue in recent years due to climate change and global warming, data centers – which traditionally consume a huge amount of electricity – are being designed to better meet the goals of energy efficiency and carbon reduction. When a national aerospace center in Europe wanted to expand their data center, with a requirement for servers that could maintain normal operation in an ambient temperature environment of 40°C without the need for air conditioning systems, GIGABYTE’s server team provided a solution with Hyper-Converged Infrastructure (HCI), combining compute, storage and networking into a single system, and equipped with liquid cooling technology to successfully overcome the harsh high temperature environment, enabling the data center to efficiently process an enormous amount of space-related research data within a limited space.

    [Video] GIGABYTE Server Liquid Cooling Solutions

    GIGABYTE can now offer various server platforms with liquid cooling technology, such as direct-to-chip liquid cooling (liquid to liquid or liquid to air), one-phase oil immersion or two-phase liquid immersion technologies. Liquid cooling systems can support a greater density of CPUs and GPUs, enabling better compute performance in a given amount of space, as well as helping customers to reduce their power consumption for cooling infrastructure, to achieve a better data center PUE.