IDCA News

All IDCA News

By Loading

2 Feb 2022

Share on social media:


EU project develops AI enhanced cooling of small data centers

The ECO-Qube project, funded by the EU, aims to apply artificial intelligence to data center cooling. The project focuses on smaller data centers because emerging digital trends like artificial intelligence (AI), autonomous driving, augmented reality (AR), 5G and the Internet of Things (IoT) require lower latency and wide-band connections, and are forcing data centers to be closer to the end users. In other words, there will be more but smaller data centers that serve as edge computing sites. The sites must be able to operate efficiently under a variety of climatic conditions. Using artificial intelligence, cooling settings can be adjusted based on local conditions, such as temperature and humidity, and take into account the cooling and energy performance of the IT systems at the site.

Precision cooling

Research from Uptime Institute and other sources shows that data centers still mainly use air for cooling. Only 14% of companies use liquid cooling, 56% use precision cooling, and 30% use basic room-level cooling. This data shows that air cooling systems are still used by the vast majority of data centers.

The majority of these air cooling systems are not capable of precision cooling. Conventional cooling systems operate within a strict temperature interval and do not evaluate measurable cooling performance. Using only sensor readings, this cooling approach is unable to measure cooling efficiency of IT components. Underperforming IT performance results from unmeasured cooling efficiency, increasing overall energy consumption.

Enhance cooling performance

ECO-Qube, or Artificial Intelligence-Enhanced Cooling System for Small Data Centres, is a holistic management system designed to enhance energy efficiency and cooling performance by orchestrating both hardware and software components in edge computing applications. ECO-Qube utilizes unused data from active data center components to harvest valuable information. Big data is being used by an artificial intelligence augmented system to detect cooling and energy requirements instantly.

ECO-Qube differentiates from conventional cooling systems which keep operating temperatures within a strict interval and do not evaluate measurable cooling performance. Underperformance of cooling results in poor airflow, thermal disequilibrium, and high energy consumption. This project is developing a zonal heat management system that uses Computational Fluid Dynamics (CFD) simulations to determine the best airflow and cooling system with the least energy consumption. Furthermore, ECO-Qube uses intelligent workload orchestration to keep CPUs at their most energy-efficient state and maintain thermal equilibrium to reduce overheating risks.

Energy management system

ECO-Qube's smart energy management system (EMS) aims to track the energy demand and operate the energy supply in cooperation with the building or district's EMS. With this synergy, renewable energy sources are maximized, and sources with a large carbon footprint are minimized.

The solution developed for this project will be tested in three different pilot projects under different climatic conditions to validate energy efficiency.


Photo credit: Tim Gouw


Follow us on social media: