Title:	TIME TO DEVELOP A DIGITAL TWIN FOR THE HPC DATA CENTER
Author(s):	Luigi Telesca, Davide De Chiara, Marco Antonini, Giovanni Ponti and Marta Chinnici
ISBN:	978-989-8704-71-9
Editors:	Paula Miranda and Pedro Isaías
Year:	2025
Edition:	Single
Keywords:	Digital Twin, Data Center, HPC, Energy Efficiency, Data Science, AI
Type:	Short Paper
First Page:	270
Last Page:	274
Language:	English
Cover:
Full Contents:	if you are a member please login
Paper Abstract:	High-performance computing (HPC) data centers are facing increasing energy consumption, even though there is a pressing need for greater efficiency. By 2030, these facilities are projected to account for 13% of global energy demand, and this could rise to as much as 21%. Furthermore, HPC data centers are expected to contribute 8% of global carbon emissions, which can result in substantial operational costs, threats to power security, and environmental challenges. Traditional methods such as heuristics, statistics, and engineering approaches have proven ineffective in enhancing energy efficiency due to the vast number of configurations, complex non-linear parameter interactions, and the enormous volume of operational data involved. Therefore, optimizing data centers—particularly HPC clusters—has become a critical issue. The integration of the Internet of Things (IoT), sensors, and intelligent devices has significantly increased the generation of operational management data from various aspects of the data center industry. By effectively modeling and processing this data, it is possible to improve energy efficiency, ensure reliability, reduce operating costs, and sustainably manage data centers. The authors propose a holistic approach that utilizes a Blockchain Digital Twin framework, enhanced by Artificial Intelligence (AI) processed data, to represent the physical complexities of data centers in virtual and tokenized models. The goal of this approach is to achieve real-time prediction, optimization, monitoring, control, and improved decision-making. Specifically, the methodology, architecture, and steps taken to develop a scalable analytical dashboard using Blockchain technology are detailed. This pioneering Blockchain Digital Twin approach has been tested in a real-world setting, specifically at the ENEA HPC Cluster known as CRESCO7. This system learns from actual operational data and enables real-time management, monitoring, and control, allowing for comprehensive visibility into the status of the data center and optimizing its functionalities. The authors also aim to demonstrate the benefits of the Blockchain Digital Twin approach for various stakeholders involved in the data center ecosystem.
	Go Back