HSM & Cluster FS 2025: EOC Conference on Intelligent Data Management
July 1, 2025
HSM & Cluster FS 2025: EOC Conference on Intelligent Data Management
From 23 to 25 June, EOC held the 11th "HSM & Cluster FS" Conference at Ulrichshusen Estate in the Mecklenburg Lakeland. Eighty experts used the opportunity to share their experience and the latest developments in hierarchic storage management and cluster file systems. Since its premiere in 2003 this conference has established itself as the most important platform of its type in the German language area.
The participants at HSM & Cluster FS 2025:
80 people from 40 different companies and research institutions.
In everyday life, huge amounts of data are collected and processed daily – often without our being aware of it. Complex systems behind many applications assure that data are reliably stored, quickly retrieved, and can be efficiently used. The requisite technology - the architecture and infrastructure – is usually invisible, although it plays a decisive role.
This is precisely where the "HSM & Cluster FS 2025" conference organized by EOC comes into play. Already for more than 20 years this conference series has brought together experts from science, industry and the service sector to discuss together challenges and solutions in the area of data storage and processing.
This year some 40 companies and research institutions came together for an exchange concerning two especially important topics: hierarchic storage management (HSM) and cluster file systems. Both technologies assist in the management of large amounts of data – often amounting to petabytes - so that they are available exactly when required, at the right location, and at the required speed. That is crucial, for example, when supercomputers analyse huge data sets, such as for training AI models or evaluating satellite images.
In specialist presentations participants discussed such questions as:
How can large amounts of data be stored long-term without their being difficult to access?
Which storage technologies are today and will be in the future especially efficient, safe and low-energy – especially for data that are required often ("hot"), sometimes ("warm"), or only rarely ("cold")?
How can huge data management systems be updated without disrupting ongoing operations?
What requirements must the interfaces of modern data management systems fill so that they are both technically of high performance and user-friendly?
Which file systems are especially suitable for supercomputers, such as for the processing of AI training data?
How should the modern management of research data be conceived in order to combine data quality with user-friendliness?
An indication of their creative data management solutions was given by TU Dresden, Leibniz Institute of Atmospheric Physics, Alfred Wegener Institute, Max Planck Institute for Psycholinguistics and EOC, among others.
But such systems are not only relevant in the context of research data; we profit from such technologies also in everyday life – for example, when we retrieve films via streaming services or when pension funds safely manage our data.
"Suitably Archiving Cold Data: from OHSM to ScoutAM" by Dr. Malte was selected as the best conference presentation. Dr. Malte reported on the introduction of new archiving software for the Alfred Wegener Institute's heterogenous research data. From the viewpoint of a systems administrator he compared the functionality and user-friendliness of the new solution with previous systems. He also pointed out the importance of selecting appropriate storage hardware for achieving maximal data throughput in systems.
EOC develops and operates modern infrastructure for researchers, which includes the German Satellite Data Archive and the terrabyte platform. These systems store, manage and process huge amounts of data, mainly relying on such technologies as HSM and cluster file systems. In this way it is assured that science data is available long-term, is well organized, and optimally useful.
