Shulou(Shulou.com)11/24 Report--

CTOnews.com February 23 news, according to JD.com cloud announced that JD.com cloud fourth-generation self-research server has been launched recently.

Jingdong cloud self-research server adopts innovative system architecture and uses the fourth generation Intel Xeon scalable processor. According to the internal test results of JD.com, the performance can be improved by 1.7 times, and the delivery of the whole cabinet can be supported, which increases the delivery efficiency by 5-10 times. In terms of power supply and heat dissipation, the use of a new server architecture can improve the efficiency of full-link power supply by 3.67%, and the use of EVAC radiator and liquid cooling technology can increase the efficiency of heat dissipation by 50%.

In architecture design, by adopting modular design, flexibly supporting pre-I / O or post-I / O, and supporting tool-free maintenance, the material reuse rate can reach 70%, the development cycle can be shortened by 60%, and the development cost can be reduced by 60%.

Jingdongyun fourth-generation self-research server has designed a multi-power supply scheme, which can improve the full link efficiency by 3.67% under the condition of single cabinet 20kW. This result can save 8500 kilowatt-hours of electricity per year and reduce carbon emissions by 8.7t.

At the heat dissipation level, JD.com Cloud's fourth-generation self-developed server uses a high-performance EVAC radiator, which can save power consumption by 8%, reduce fan speed by 34.3%, and improve noise.

For the heat dissipation scheme of the data center, the liquid cooling module is used. JD.com cloud liquid-cooled server cold board with modular design, can support multi-platform CPU, up to 750W chip power consumption, and can intelligently switch air-cooled and liquid-cooled fan control strategies to ensure energy-saving optimization, and the use of UQD quick-release connector to meet the modular design. The most important thing is that JD.com cloud liquid-cooled server has passed rigorous reliability tests, including high and low temperature, humidity, vibration, impact, ultimate pressure and other reliability tests, which improves the reliability of liquid cooling design.

From the application data, the Jingdongyun data center in Langfang has improved the PUE to less than 1.1 through the large-scale use of liquid cooling, heat recovery and other technologies.

CTOnews.com Note: PUE, short for Power Usage Effectiveness, is an indicator of data center energy efficiency and the ratio of all energy consumed by the data center to the energy consumed by IT loads. PUE = total energy consumption of data center / energy consumption of IT equipment, in which total energy consumption of data center includes energy consumption of IT equipment and energy consumption of refrigeration, power distribution and other systems. A value greater than 1 indicates less energy consumption of non-IT equipment, that is, better energy efficiency.

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.