Thermal monitoring and control is the method to solve the increase of heat generation in data center

The maximum power capacity of your data center must cover the energy consumption and thermal requirements of your server and IT equipment. As data center power consumption has increased from an average of 500 watts per square foot to today's average of 1500 watts per square foot, balancing these two "balanced" demands has become more difficult in recent years! The thermal effect of higher-performance density (HPD) hardware often results in greater data center heat generation.

One way to address the increase in heat production in the data center is to achieve a more efficient cooling facility, using the following emerging best practices for thermal monitoring and control.

Establish a real-time thermal data center map

Enable real-time http://www.aliyun.com/zixun/aggregation/18442.html "> temperature sensors on each server platform for real-time data center thermal map construction." Real-time monitoring of power and thermal events on separate servers and racks (except for some rooms) can enable you to proactively detect failure situations and then take action on specific situations, whether it is cold, excessive cooling, hot spots, or computer room air conditioning (CRAC) failure.

These hot maps can be used to create thermal profiles, records, and report hot trends as well as long-term planning.

The hot spot can be further determined based on the CRAC supply temperature monitoring and control of the real-time data center environment. The condition of ultra cold and excessive cooling is often due to lack of information about the actual ambient temperature of the data center rack and the room level.

In addition, real-time thermal profiles report hot trends to show that increased operating temperatures may lead to a significant reduction in the cost of cooling energy. In several data center pilot projects around the world, the Intel Data Center Manager (DCM) Solutions team has witnessed an accurate reading based on increased energy efficiency and temperature: the net savings of approximately 50000 USD/year for each upgrade of the operating temperature of the datacenter.

Actual data and theoretical data

Today, many available power and thermal data are estimated or based on the manufacturer's rated power, rather than actual consumption. These data can be as high as 40% of actual consumption deviations.

The cooling efficiency of the real-time rack and room-level thermal mapping identification allows you to activate the appropriate cooling action as soon as you need. Sun Microsystems reported that the data center could save 4% of its energy costs at a temperature setting with the help of data Center manager for each increase in operating temperature. It is also reported that cooling can account for the total amount of energy used in the data center 40–50%. Increasing efficiency in this area will have a significant impact on the overall operating costs of the data centre.

Case study

Microsoft hopes to find out how much the cost can be saved by increasing the temperature of the data center's cooling setting. The company conducted a slightly higher temperature impact test at its Silicon Valley data center. "We raise the floor temperature by Shito, saving energy costs by 250,000 dollars a year," he said. "Dondenning, key facility manager at Leetechnologies, who works with Microsoft on the project, said.

When CIOs and their infrastructure teams bring more high-performance-density hardware to the data center, requiring more servers and greater thermal cooling efficiencies, be sure to ensure that these best practices are part of your thermal control plan.