Energy Efficiency and Increasing Data Center Temperatures Without Risk
Even when ASHRAE Technical Committee 9.09 released the Thermal Guidelines for Data Processing Environments, a path to improve efficiency was laid out to be followed. And with each update additional energy savings could be found as the group found that IT equipment was becoming much more robust along with other findings. More resources and tools came from that group, The Green Grid, IEEE, and others as each reacted and added their knowledge.
The trend that promises significant energy savings by increasing data center temperatures is certainly not new. Decades ago data centers might have been kept at low temperatures to prevent overheating and ensure optimal performance of IT equipment. However, advancements in technology and better understanding of thermal management have shown that it is possible to safely raise data center temperatures without compromising equipment reliability. Again, even in 2024 is getting more facilities, operations personnel, IT professionals, and more comfortable with shedding their sweaters and jackets before entering data halls.
The Benefits of Higher Data Center Temperatures
1. Energy Savings: Cooling systems are among the largest consumers of energy in a data center. By increasing the ambient data hall temperature, the load on cooling systems, from the air circulated to chilled water or other means of moving the heat out, is reduced, leading to significant energy savings. Under ideal circumstances for a vapor compression coefficient of performance, a maximum savings of over 7% per degree F is theoretically possible. According to the U.S. Department of Energy, for every 1°F increase in the supply air (aka server inlet) temperature, there is a potential 3-5% energy savings in cooling costs. However, given real world applications and measurements, 1-2% is much more likely – which is still a large amount given a data center operates 7/24/365.
2. Environmental Impact: Reducing the energy consumption of cooling systems not only lowers operational costs but also minimizes the carbon footprint of data centers. This aligns with the growing emphasis on sustainability and environmental responsibility in the tech industry.
3. Cost Efficiency: Lower energy consumption translates to immediately reduced operational costs. This can free up resources for investment in other efficiency measures, such as upgrading equipment or adding AI to help with manage the overall efficiency scheme.
Challenges and Considerations
1. Equipment Tolerance: Not all IT equipment is designed to operate at temperatures above 100-120 degrees F. It is crucial to understand the thermal tolerance of the hardware and the other components that may see these higher temperatures. The least robust equipment will often dictate how high the temperature might be raised, along with safety or bypass factors.
2. Thermal Management: Effective thermal management remains critical at higher temperatures. This includes ensuring proper airflow and regularly monitoring temperature and humidity levels with relative granularity to monitor for hot spots or other issues.
3. Risk of Overheating: While modern equipment can tolerate higher temperatures, there is still a risk of overheating if temperatures are not properly controlled. Overheating can lead to equipment failure, data loss, and downtime.
4. User understanding: One of the main approaches to share is that the data center temperatures vary greatly in an optimized data center, from hot to cold aisles. When they do not, it means there are bypass or other losses that are decreasing the overall efficiency and should be examined further.
Air Management Audit
Before adjusting any temperatures, the airflow and overall management approach should be defined, as each data center is unique. This can include underfloor, overhead, rear-door or overhead heat exchangers, in-row cooling units, liquid, combinations of these, and more. Aisle, row, and rack air containment implementation needs to be included in addition to the overall air management scheme.
When optimized, bypass air is kept at a minimum (ideally zero) which allows an increase on the temperature exchange for the volume of air circulated. This also leads to reduced energy used by fans and a better heat exchange rate across the cooling equipment; again, ideally the temperature differences are maximized and equal for the IT equipment and cooling equipment.
An engineering and/or operations team should be able to define whether the air management is acceptable to proceed with adjusting the temperatures. Locations with poor or unknown air management will exacerbate any issues if the temperatures are increased, affecting the performance of the entire data hall and facility. Adjusting and improving the air management is one of the main methods of energy savings as well, so this may be the first step in decreasing overall cooling energy and improve the PUE.
Increasing Data Center Temperatures
Fundamentally, the savings achieved by increasing the temperatures of a data center are derived from reducing the amount of energy needed for mechanical cooling, whether by a chiller or other cooling equipment. The temperature increase is passed along to the heat rejection equipment, which then can reduce the amount of lift (or compressor energy) needed to reject the heat to the atmosphere, thereby saving the facility energy.
1. Understand Equipment Specifications: Before raising temperatures, review the thermal specifications of all IT equipment. Most modern servers and networking gear are designed to operate efficiently at higher temperatures, often well over 90°F or higher. ASHRAE designations (A1, A2, A3, A4 or A5) may also be used, which correlate to specific temperature ranges.
2. Implement Advanced Cooling Solutions: Use advanced cooling technologies such as liquid, direct-to-chip, rear-door, or in-row cooling to manage heat more effectively. These solutions can help maintain optimal localized operating conditions even at higher temperatures.
3. Regular Monitoring and Maintenance: Implement a robust monitoring system to continuously track temperature, humidity, and airflow. Regular maintenance of cooling systems and IT equipment is essential to prevent overheating and ensure consistent performance.
4. Gradual Temperature Increase: Avoid sudden temperature changes. Gradually increase the temperature in small increments while closely monitoring the impact on equipment performance and stability. This approach allows you to identify temperature limits without risking equipment failure.
5. Leverage Free Cooling: Utilize outside air economization or "free cooling" when environmental conditions permit. This technique leverages cooler outside air to reduce the load on traditional cooling systems, further enhancing energy efficiency. Enabling this method may take additional considerations or facility improvements, including filtration to ensure poor outside air quality does not affect the equipment. Once implemented, the air economizer will see greater use throughout a year, yielding immediate savings with every degree of increase that can be implemented.
6. Adopt Industry Standards: Follow industry guidelines such as the ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers) recommendations for data center temperature and humidity ranges. ASHRAE provides the most well-researched guidelines that balance energy efficiency and equipment reliability. Included are guidelines for taking measurements, equipment placement, airflow considerations, and other metrics that can be applied for improvements.
Case Studies and Success Stories
How much energy saved is relative to each data center and the increase of the temperatures. With air economization, the hours of free cooling can be increased significantly. Less compressor use is another outcome of increasing the temperatures, with or without air economizers, as well as less energy for the lift to provide chilled water or air to the data center.
Google and Facebook have reported significant energy savings by increasing their data center operating temperatures, boasting 40% or more savings on their overall energy for some data halls. These companies have invested in advanced cooling technologies and robust thermal management practices to ensure the reliability and efficiency of their operations, along with other measures are aren’t reliant on increased temperature savings alone.
In the quest for greener and more efficient data centers, raising operating temperatures represents a forward-thinking approach that aligns with both economic and environmental goals. With careful planning and execution, this strategy can deliver substantial benefits without compromising the performance and reliability of critical IT infrastructure.
