By David Watkins, Solutions Director at VIRTUS Data Centres
David Watkins
In today’s digital era, data centres are the backbone of our connected world, supporting everything from cloud services to artificial intelligence (AI). However, their significant energy demands, and environmental impact pose challenges that can’t be ignored. As data processing and storage needs increase, particularly with the rise of AI and high-performance computing (HPC), data centres are under scrutiny for their energy consumption and sustainability practices. Addressing these issues is not only crucial for operational efficiency but also for aligning with global sustainability goals.
The Energy Challenge
Data centres are known for their high energy consumption requirements, with cooling systems being a major contributor. Traditional air-based cooling methods, which circulate cold air to manage the heat from server racks, can be inefficient and power-hungry for growing computational demands. As the scale and intensity of AI workloads increase, so does the need for effective thermal management. This, in turn, drives up energy use and exacerbates the environmental impact of technology and its infrastructure, which are already significant contributors to global carbon emissions.
To meet these challenges, the data centre industry is increasingly turning to more sustainable cooling solutions. Innovations in cooling technology offer a promising path forward, potentially transforming data centre operations and reducing their environmental footprint.
Liquid Cooling: A Sustainable Solution
Liquid cooling has emerged as a transformative alternative to traditional air-based systems. By using liquid to directly cool components, this method offers superior heat dissipation compared to air. This not only enhances cooling efficiency but also reduces the overall energy footprint of data centres. It can reduce the amount of energy used by a facility as it takes less electricity to cool a server than air cooling systems. It is important to understand that whilst liquid cooling supports far denser compute deployments, it is not detrimental to a facilities Water Usage Effectiveness (WUE) performance. Both direct to chip and immersion cooling operate separate, dedicated cooling loops that in turn are reliant on the main data centre cooling system for heat transfer.
Two primary liquid cooling technologies stand out: immersion cooling and direct-to-chip cooling. Immersion cooling involves submerging IT hardware, such as servers and graphics processing units (GPUs), in a specially designed dielectric fluid. This fluid, which can be mineral oil or a synthetic coolant, absorbs heat directly from the components, eliminating the need for extensive air conditioning. This approach significantly boosts energy efficiency and lowers operational costs, and is particularly well-suited for high-density AI workloads that generate substantial heat.
Direct-to-chip cooling, or microfluidic cooling, delivers coolant directly to the heat-generating components like central processing units (CPUs) and GPUs. This method maximises thermal conductivity by targeting heat dissipation precisely at the source, which enhances performance and reliability. By efficiently managing the heat from critical components, direct-to-chip cooling helps enable optimal operation of AI applications and reduces the risk of thermal throttling and hardware failures.
Both these methods offer considerable energy savings compared to traditional air-based systems – and for most modern data centres, a mix and match approach is optimal.
Heat Reuse and Renewable Energy
Another key strategy for improving sustainability in data centres is heat reuse. Capturing and repurposing waste heat from data centre operations can contribute to energy conservation and reduce environmental impact. For instance, waste heat can be redirected to nearby residential or commercial buildings for heating or hot water, supporting local energy infrastructure and reducing the need for additional energy sources.
Integrating renewable energy sources into data centre operations is also crucial. Solar power, for example, can be harnessed by installing panels on data centre rooftops or adjacent areas. This renewable energy can directly power cooling systems and other operations, reducing reliance on fossil fuels and decreasing greenhouse gas emissions. Similarly, wind power, obtained through on-site turbines or power purchase agreements with wind farms, can provide a source of renewable energy, complementing other energy sources and supporting carbon neutrality goals.
Beyond adopting advanced cooling technologies and renewable energy, data centres can further enhance sustainability through various energy efficiency measures. Optimising cooling systems with AI and machine learning can help adjust temperatures based on real-time data and workload demands, reducing energy waste and improving resource utilisation.
Overcoming Challenges
As data centres continue to evolve in response to increasing digital demands, embracing energy and sustainability solutions is essential. By integrating advanced cooling technologies, renewable energy sources, and energy efficiency measures, data centres can significantly reduce their environmental impact and operational costs.
