Chill factor: Top liquid cooling considerations for high-density environments

Disclaimer: This article delves into the concept of heat reuse, its applications, and the relevant regulatory and legislative initiatives. It draws on experiences and policies from EMEA countries, including the EU, and is intended as a best practice guide for understanding heat reuse within this region. It's important to note that local and regional policies and guidelines must be considered and adhered to for areas outside the EU. For more information, please download the white paper at the end of the article.

As AI applications drive the growing demand for high performance computing, servers are packing more processing power to handle increasingly complex and data-intensive tasks. This increased power results in higher heat generation, pushing the limits of traditional air cooling, especially in high-density rack environments.

Liquid cooling has emerged as a valuable addition to the data center cooling toolbox. Nearly one in five data centers (17%) already use liquid cooling, and an additional 61% of operations teams are considering its implementation. This trend reflects the industry's confidence in the effectiveness and efficiency of liquid cooling in managing heat from high-density IT racks.

However, integrating liquid cooling into existing systems isn't simple. Since liquid cooling is a relatively new addition to the data center cooling landscape, overcoming the hurdles associated with this transition remains a challenge.

Explore the advantages of liquid cooling, key considerations for successful implementation, and available services to help you navigate the complexities of adding this technology into your data center cooling strategy.

Advantages of liquid cooling

1. Enhanced heat dissipation - Liquid cooling utilizes the higher thermal transfer capabilities of water or other fluids to provide efficient and cost-effective cooling for high density racks and can be up to 3,000 times more effective than using air.

2. Improved IT reliability and performance - Liquid cooling systems allow densely packed systems to operate at maximum voltage and clock frequency without overheating, unlike air cooling, which throttles CPU performance as temperatures rise. Liquid cooling improves energy efficiency by using pumps that consume less power than fans.

3. Energy efficiency - Liquid cooling lowers data center energy usage, enhances power usage effectiveness (PUE), and enables efficient heat repurposing for building heating systems.

4. Reduced footprint - Liquid cooling also maximizes space utilization, by enabling higher density within the same physical footprint. By efficiently managing heat, liquid cooling reduces the need for facility expansions, making it a valuable solution for data centers.

5. Lower TCO - Liquid cooling can lower the total cost of ownership (TCO) through higher density, increased use of free cooling, and improved performance per watt, according to the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE).

Key considerations

Although some new facilities will be specifically designed for AI workloads and liquid cooling, most installations will take place in existing facilities. Given the complex nature of liquid cooling systems, data center teams can benefit from full-service liquid cooling solutions that cover design, installation, and ongoing maintenance.

Learn more:
Liquid Cooling Services

As heat surges in high-performance computing environments, the need for liquid cooling technologies in modern IT thermal management becomes critical.

To successfully implement liquid cooling and maximize its benefits, consider the following key factors:

1. Installation

The incremental addition of liquid cooling, whether in an existing data center or a new facility, often requires significant changes to the infrastructure. In brownfield projects, new piping systems, pumps, and heat exchangers may need to be installed, potentially requiring modifications to server racks to accommodate liquid cooling components. In greenfield projects, the infrastructure can be designed from the ground up.

Liquid cooling systems can optimize your facility's space by enabling denser server configurations within each rack. However, the cooling units themselves may require additional space and might require reconfiguring your facility's layout. For brownfield deployments, careful planning is critical to provide enough room for the new equipment without disrupting the current operations. In greenfield environments, space allocation and layout can be tailored to the cooling system, maximizing efficiency and performance from the outset.

To streamline this process, whether you're retrofitting an existing data center or building a new one, consider partnering with experienced providers specializing in liquid cooling solutions. Vertiv offers comprehensive services that cover every aspect of the installation process for both brownfield and greenfield projects, including program and design, factory witness testing, delivery, quality assurance (QA), and installation assembly.

2. Cleanliness

The performance of liquid cooling systems can be significantly impacted by cleanliness because direct-to-chip cooling relies on clean, uncontaminated fluids to function efficiently. Any dirt, dust, or debris introduced during installation or servicing can contaminate the coolant, impact server operations, reduce thermal performance, and lead to potential system failures.

Data centers house sensitive electronic equipment that can be adversely affected by dust and other contaminants. During the installation or servicing of liquid cooling systems, it is crucial to maintain a clean environment to prevent particles from settling on electronic components, which can cause overheating, short circuits, or other malfunctions.

Following strict cleanliness standards and protocols during the installation and servicing of liquid cooling systems helps meet local regulatory requirements and maintain certifications

3. Commissioning

Commissioning involves ensuring that the liquid cooling system operates as intended and meets the specified performance criteria. This process includes testing and validating system components, fluid flow control, cooling efficiency, and overall functionality under real operating conditions.

It also helps identify and address potential issues or defects in the system before it becomes fully operational. This includes detecting leaks, verifying pressure levels, checking for proper fluid distribution, and ensuring all components function correctly.

Through commissioning, adjustments can be made to optimize the system's performance, such as fine-tuning flow rates, adjusting temperature settings, and calibrating sensors for efficient cooling.

4. Fluid management

Proper fluid management involves maintaining the correct fluid levels, flow rates, and temperatures. Any deviations can impact the system's ability to effectively transfer heat away from data center equipment, potentially leading to overheating and performance issues.

Effective fluid management helps prevent leaks, which can cause significant damage to equipment and infrastructure. It also minimizes the risk of contamination, as impurities or foreign particles in the coolant can lead to blockages, corrosion, and reduced thermal performance. Regular monitoring and maintenance are essential to keep the fluid clean and free of contaminants.

5. Recurring services

Recurring services help identify and address potential issues before they lead to system failures. It also helps extend the lifespan of the liquid cooling system and its components.

Additionally, regular servicing helps optimize system efficiency by cleaning components, adjusting settings, and making necessary upgrades, which can also contribute to energy savings. Compliance with industry standards and safety regulations is maintained through periodic checks, and thorough documentation provides valuable insights into the system's condition over time.

By keeping the system in good condition and addressing wear and tear promptly, recurring services can reduce the need for costly repairs or replacements.

6. Spare parts

Having spare parts readily available ensures that any issues or failures can be addressed quickly, minimizing downtime and maintaining continuous operations of your facility. It also reduces lead times and prevents operational disruptions, allowing for immediate action when issues arise and avoiding prolonged periods of reduced cooling capacity.

Establishing a spare parts inventory also eliminates the need for emergency repairs or expedited shipping.

Exploring liquid cooling options

There are several liquid-cooling technologies available to meet the cooling needs of high-density racks. However, installing liquid cooling can be complex. Data center teams will want to work with a partner to address key issues, such as plumbing requirements, cooling distribution, capacity balancing, risk mitigation strategies, and heat rejection systems.

Here are some liquid cooling options to consider:

Rear-door heat exchangers

Rear-door heat exchangers utilize the high thermal transfer properties of liquid without bringing it directly to the server. Passive or active heat exchangers replace the rear door of the IT equipment rack with a liquid heat exchanger. In passive designs, server fans push heated air through a liquid-filled coil in place of the rear door, absorbing the heat before it enters the data center. Active heat exchangers use fans to draw air through the coils and remove heat from even higher density racks. These systems can be used alongside air-cooling systems.

Direct-to-chip liquid cooling

Direct-to-chip cold plates are used to cool heat-generating components like central processing units (CPUs) and graphics processing units (GPUs). They can use single-phase cold plates or two-phase evaporation units to remove heat. It can remove about 70-75% of the heat generated by the equipment in the rack, leaving 25-30% that needs to be managed by air-cooling systems.

Immersion cooling

Immersion cooling submerges servers in a conductive liquid, eliminating the need for air cooling and fans. There are single-phase and two-phase systems. In the single-phase system, servers are submerged in a conductive fluid, with heat removed by a Coolant Distribution Unit (CDU).

Learn more:
Liquid cooling options for data centers

Data center operators are evaluating liquid cooling technologies to increase energy efficiency as processing-intensive computing applications grow.

Deploying liquid cooling

In summary, liquid cooling offers an efficient solution for managing heat in high-density computing environments. Its ability to maintain optimal temperatures ensures reliable performance, making it a compelling choice for modern data centers. However, successful implementation demands more than technology-it requires a comprehensive approach.

Vertiv's Liquid Cooling Services address this challenge by offering design, installation, and ongoing maintenance for operational efficiency and system availability, ensuring compliance with global efficiency standards for thriving data centers.

If you have further questions or need personalized advice, contact us about Liquid Cooling Services.

About the authors