Written by Eamonn Ryan
Rack densities may be the hottest thing in a data centre. Literally. The Pan-African Data Centre Exhibition & Conference boasted a high proportion of HVAC&R companies. It also hosted a panel discussion on the subject of Key Considerations in Future Cooling Applications for Data Centres. The latter was moderated by Jonker Bester, director: Spoormaker & Partners, with the following industry expert panellists:
• Vincent Laidet, managing director: Systemair
• David Bourke, director of sales: data centre technologies EMEA, Munters
• Matteo Faccia, product management director and tech dep operation coordinator, HiRef
The following is a summary of their discussion. Efficient cooling is a critical aspect of data centre operations, directly impacting equipment performance, energy consumption, and overall reliability. The panellists shared insights on cooling strategies, common mistakes, and best practices to achieve optimal airflow management. By examining design considerations, temperature control, and infrastructure maintenance, data centre operators can enhance cooling efficiency and reduce energy costs.
Understanding airflow management: To ensure effective cooling, it is essential to focus on airflow management within data centres. The panel emphasises the importance of correct positioning and sizing of equipment to maintain optimal conditions. Simply increasing airflow speed may not yield the desired results. Instead, the goal should be to provide the right amount of airflow consistently, optimising energy usage and equipment performance. The discussion highlights the need to align with basic principles and avoid common mistakes that hinder proper airflow management.
Avoiding common mistakes: The panel points out two common mistakes frequently encountered in data centres:
• The first involves placing server cabinets too close together, causing short cycling and inadequate airflow circulation. This error limits cooling efficiency and may lead to equipment performance issues.
• The second mistake pertains to incorrect installation of the cooling unit. Sometimes, the cooling unit is mounted upside down or obstructed by design elements, impeding its functionality, and forcing the system to work harder to achieve desired results. Such errors can be avoided with attention to detail and adherence to best practices.
Temperature control and humidity considerations: Proper temperature and humidity control play a crucial role in data centre cooling. The panel emphasises that achieving the right temperature range is essential for optimal performance and energy efficiency. Even minor deviations in temperature can lead to higher energy consumption and equipment strain. Maintaining a balanced airflow that efficiently removes heat from the server cabinets and controls humidity levels ensures a conducive environment for equipment operation. The panel stresses the importance of maintaining the right balance in airflow and temperature for maximum efficiency.
Best practices and infrastructure maintenance: Implementing best practices is vital to achieving efficient cooling in data centres. Properly sizing the data centre and adopting conservative expenditure graphs help align infrastructure with operational needs, reducing unnecessary energy consumption. Regular maintenance and infrastructure audits allow operators to identify potential issues and implement corrective measures promptly. The panel emphasises the significance of maintaining a pre-emptive approach to infrastructure maintenance to ensure optimum cooling performance.
Choosing the right cooling technology: The panel emphasises the importance of choosing the appropriate cooling technology for each specific case. They acknowledge that there is no onesize-fits-all solution and stress the need to analyse requirements and goals before making a decision. Factors such as system efficiency, cost, and ease of implementation are taken into account when considering options like water cooling or air cooling. While water cooling may offer greater cooling capacity, it also requires a larger volume of cooler air. On the other hand, air cooling is more cost-effective but may result in quicker temperature increases within the data centre. Ultimately, the choice depends on the specific needs and circumstances of the data centre.
Maintenance and efficiency: The significant role of regular maintenance in ensuring optimal cooling system performance is emphasised. Proper maintenance not only contributes to the longevity of the system but also helps maintain efficiency. Neglecting maintenance can result in decreased efficiency, increased energy consumption, and potential system failures. Regular inspections, cleaning, and filter replacements are essential to avoid additional strain on the system and unnecessary energy wastage. The discussion underscores the need for competent individuals to carry out maintenance tasks to maximise efficiency and minimise operational issues.
Configuring cooling units: The panel addresses the importance of appropriately sizing and configuring cooling units within the data centre. Factors such as the position of the cooling unit, duct sizing and configuration, and external static pressure influence the energy consumption and efficiency of the system. Placing the cooling unit closer to the centre reduces energy consumption, while the sizing and configuration of ducts affect the external static pressure required for optimal performance. Regular maintenance of filters and ensuring proper airflow also play a role in maintaining efficiency and minimising energy waste.
Sensitivity ratio and energy optimisation: The significance of the sensitivity ratio in energy optimisation efforts was highlighted. Aiming to achieve a sensitivity ratio close to 1 helps reduce wasted electricity and improve overall energy efficiency. This involves maintaining the right temperature and airflow conditions within the data centre. By considering factors like row design and temperature control, data centre operators can minimise energy wastage and achieve optimal cooling efficiency.
Understanding the factors that influence cooling solutions, such as system configuration, maintenance, and sensitivity ratio, allows data centre operators to make informed decisions and optimise energy consumption. By prioritising these considerations and adopting best practices, data centre operators can enhance cooling efficiency, improve equipment reliability, and reduce overall energy costs.
Local impact and modularity: The discussion begins by highlighting the importance of considering local factors when designing data centres. Factors such as load shedding, power usage, and temperature variations across different markets and regions need to be taken into account. The panel emphasises the need for data centre designs to be adaptable and modular, enabling them to be deployed worldwide. By adopting modular solutions and considering commonality within each market, data centre operators can achieve consistency and efficiency across various locations. The panel discusses the benefits of modular solutions, which allow for flexible expansion and efficient use of resources. Modular data centres can be easily integrated into existing infrastructure, supporting scalability and adaptability to changing demands.
Efficiency in system design and management: The panel emphasises the need for efficient system design that goes beyond considering worst-case scenarios. Instead, the focus is on developing systems that can adapt and perform optimally during regular operations. The discussion explores the concept of managing systems efficiently by switching on and off cooling elements based on real-time requirements. The use of variable speed compressors and fans is highlighted as a way to enhance cooling efficiency, especially for large machine units. By optimising cooling distribution and considering customer needs, data centre operators can achieve greater flexibility and energy efficiency.
Addressing rattling issues and scalability: Another crucial aspect discussed is the increasing challenges posed by rattling in data centre environments. The panel acknowledges the rise in rattling incidents and emphasises the need to design data centres that can handle higher heat loads per rack, ranging from six to eight kilowatts and even up to 20 to 24 kilowatts in certain instances. Data centre operators need to devise consistent solutions that can effectively handle higher heat loads while considering future scalability. Exploring alternative cooling methods, such as immersion cooling, is also mentioned as a consideration for future infrastructure planning.
Future-proofing infrastructure: The panel emphasises the importance of building data centre infrastructure with an eye toward the future. While air cooling remains prevalent, the discussion suggests the potential for alternative cooling solutions, such as immersion cooling. Data centre operators are advised to consider long-term infrastructure planning, keeping in mind emerging technologies and trends that may require different cooling approaches. By implementing infrastructure that is adaptable and forward-looking, data centres can remain efficient and relevant in the face of evolving demands. By addressing local impact, adopting modular solutions, optimising system design, and planning for future scalability, data centre operators can navigate the challenges of cooling and achieve greater efficiency in their operations.
Embracing sustainability: The panel highlights the need for the data centre industry to prioritise sustainability. As the sector continues to grow, there is a concern that current practices may not align with sustainable principles. The panel agrees on the importance of learning from past experiences and implementing lessons to drive positive change. This includes a focus on sustainability from a global perspective, with specific attention to regions like Africa.
Harnessing heat recovery: One of the key strategies discussed is the significant potential of heat recovery in data centres. The panel emphasises the importance of exploring heat recovery applications, particularly in regions such as the colder Nordic countries and other parts of Europe. While the market demand for heat recovery may vary, there is a consensus on its value in reducing energy waste. By capturing and utilising excess heat, data centres can achieve both energy efficiency and sustainability goals.
Addressing water and power scarcity: Water and power scarcity emerge as critical considerations in data centre design. The panel recognises the need to account for these challenges when developing cooling systems. To mitigate water scarcity, alternatives to traditional cooling methods, such as refrigerant usage reduction, are explored. Additionally, the discussion touches on the importance of considering power scarcity and the potential for data centres to integrate with local energy grids, benefiting both the centres and surrounding communities.
Exploring efficient cooling technologies: The panel acknowledges the importance of adopting efficient cooling technologies to minimise energy consumption. The panellists highlight the need to reduce reliance on refrigerants and explore alternative cooling approaches. Examples include direct and indirect air cooling, liquid cooling, and the integration of renewable energy sources. By leveraging advanced cooling technologies, data centres can optimise energy usage, reduce carbon emissions, and enhance overall sustainability.
Huawei’s Zhou on data centre trends, challenges, and development direction
While the remainder of this feature deals with the cooling aspect of data centres, RACA Journal had the opportunity at the Pan African DataCentres Exhibition & Conference, to talk to Huawei director of IT Consulting & System Integration Department, Zhou Yilin, about the data centre landscape and future trends in 2023.
“Digital technologies have ushered in a new wave of technological revolution and industry transformation. Increasing requirements for data storage and processing the rapid development of data centres. At the same time, this brings new challenges to data centres, including different policy environments in regional customer requirements, and low-carbon sustainable development requirements. Globally, data centres are facing more complex and changing situations,” says Zhou.
“In the future, data centres in Africa will focus more on environmental protection, data security, and privacy protection. In addition, data centres in Africa will face more opportunities and challenges from the following aspects:
- Development of cloud computing and edge computing: With the increasing popularity of cloud computing and edge computing, data centres in Africa will face more challenges and opportunities. Data centres and service capabilities need to evolve and enrich to meet users’ requirements for intelligent, efficient, and secure services;
- Construction of green data centres: African governments have started to promote the construction of green data centres to reduce their impact on the environment. In the future, Africa’s data centres will focus more on environmental protection and sustainable development, and adopt more energy-saving and environmentally friendly technologies and equipment;
- Data security and privacy protection: With the increasing number of data breaches and cyber-attacks, data centres in Africa will pay more attention to data security and privacy protection. They need to adopt more secure and reliable technologies and solutions to enhance data encryption and access control to protect user data security and privacy.
In the context of global digital transformation, Zhou describes the direction of Huawei’s core competitiveness in the future:
“As the key foundation for carrying digital economy services, data centres are the cornerstone of digital transformation in various industries. In the digital era, customer service scenarios in various industries are changing from single product integration to full-stack data centre integration. Huawei’s full-stack data centre integration service solution provides full-stack service solutions, including data centre facility integration, IT integration and auxiliary operation, migration and disaster recovery, scenario-based intelligence, and campus digital platform integration. It covers the entire process of planning, construction, and operation, meeting the requirements of comprehensive, efficient, and intelligent construction in the architecture design, service rollout, and system operation of the data centre.
“Huawei data centre integration service uses low carbon, energy saving, and prefabricated assembly to build green and simplified scenario-based typical configuration solutions,” says Zhou. “For example, for integrated DC scenarios, L1 + L2 are integrated to build a full-stack liquid cooling solution. We also build an efficient air-cooled and water-cooled system to support scenario-based energysaving solutions for hub clusters, helping customers reduce the PUE from more than two to less than 1.5. PUE is the ratio of the total amount of power used by a computer data centre facility to the power delivered to computing equipment.
However, frequent natural disasters may cause power failures, device damage, and network interruptions in data centres, undermining the running and data security of data centres. Zhou explains that Huawei improves data centre resilience and ensures service continuity by providing automatic disaster warnings, automatic recovery solutions, one-click service switchovers, and unified ransomware detection and recovery management, helping customers improve service continuity and reduce disaster recovery (DR) management costs.
“To meet the high reliability requirements of industry systems, Huawei provides industry-leading DR consulting, DR integration, DR drill services, and intelligent DR management platforms for industries such as finance, government, healthcare, electric power, and energy. It provides service solutions such as intra-city active-active, remote DR, two-site three-centre, and multi-centre, helping customers improve service continuity,” says Zhou.