ZIH-Info No. 196, March 2026
Table of contents
Editor: Jacqueline Papperitz
Data storage systems for NFDI at the TUD
The ZIH has secured funding of one million euros under the call "Data storage systems to strengthen the National Research Infrastructure (NFDI)". The project aims to procure and operate tape-based storage systems for long-term archiving, as well as disk-based object storage for research data repositories closely connected to the NHR Center’s HPC systems. The new storage systems will be available to the NFDI consortia GHGA, NFDI4Earth, PUNCH4NFDI, NFDI4Culture, NFDI4DataScience and NFDI4Chem. Commissioning is planned for this year. Funding is provided in equal parts by the federal and state governments. (Contact: Dr. Matthias Lieber)
ZIH further develops IT Services with AI
New functions have been introduced in the TUD's Self-Service Portal (SSP), marking the first use of artificial intelligence (AI) in ZIH IT services. Users can now generate access tokens (API tokens) directly in the SSP for the use of ScaDS.AI models. This enables a wide range of AI applications to connect to models operated at the LZR in a data protection-friendly environment. In addition, a new change tracking function has been introduced. Messages for new software releases are now generated automatically with AI support. These features are also the first in the SSP to be developed primarily using AI. Unlike earlier experiments that focused on prototypes, the AI-generated code has now passed the standard quality assurance and approval processes. AI-supported development and the future use of agentic AI functions sare expected to accelerate the implementation of new features, reduce costs, automate recurring tasks, and further support digitalization at TU Dresden. At the same time, the goal is to ensure sovereign and privacy-friendly use of AI services on the university’s own infrastructure. The use of AI within the services is made transparent to users. Self Service Portal: https://selfservice.tu-dresden.de/. (Contact:Dr. Daniel Hackenberg)
TUD:AI – AI services at the TUD
ScaDS.AI Dresden/Leipzig and the ZIH jointly provide various AI services for employees, students and guests at TUD under the name TUD:AI. The former ScaDS.AI LLM Inference Service will be renamed TUD:AI. This service includes applications for text and image generation, image recognition and speech processing. IIt can be accessed via a web interface or programming interfaces (APIs) and is primarily aimed at researchers. The chat web interface is integrated into TUD's group and organizational structure, with login via Shibboleth. Processing is based on powerful open-weight models such as Llama, Qwen, DeepSeek, and Kimi. A knowledge integration function (Retrieval Augmented Generation, RAG) is also available to incorporate users’ own data. All data processing takes place on TUD infrastructure, enabling data protection-compliant use. An extension for applications with higher protection requirements is currently being planned. API keys can be generated via the Self-Service Portal. Further information: https://llm.scads.ai/. (Contact: Taras Lazariv, )
IT monitoring with Checkmk
The ZIH has deployed a new infrastructure for monitoring IT services and server systems. The introduction of Checkmk replaces the previous monitoring solution and extends its functionality. The new system enables more detailed and flexible recording of system statuses and performance data. This allows faults to be detected more quickly and helps to further improve service stability. Central monitoring at the ZIH has already been migrated, and decentralized organizational units are currently being connected step by step. (Contact: Dr. Michael Kluge)
Modernized DLR supercomputer at the LZR
The German Aerospace Center (DLR) has comprehensively modernized its CARA high-performance computing system at the Dresden site and officially handed it over on 1 April 2026. The system, which is operated at the TUD's Lehmann Center (LZR) and supported by the ZIH, has been in operation here since 2020 and has now been technically upgraded to a new performance level. This is DLR's response to increasing demands in research and development, for example in the fields of aerospace, energy, transportation, artificial intelligence and quantum technologies. Among other things, the modernization includes a significantly expanded storage capacity and continues to enable highly complex simulations and data-intensive analyses. The integration into the data center infrastructure at the LZR provides particular added value: CARA's efficient hot water cooling enables particularly energy-efficient operation. The waste heat generated is reused both in the LZR's local heating network to heat a DLR building and in the city of Dresden's district heating network. (Contact: Tel.: HA -40000)
New DFG project Mod4Comp starts
In June 2026, the DFG research group Mod4Comp will start investigating energy-efficient software. The aim is to systematically model software and hardware properties in order to be able to evaluate and optimize the energy efficiency of applications without costly measurements. The project thus makes a contribution to sustainable computing and addresses the increasing energy requirements of modern computing infrastructures, for example in the context of AI applications and new hardware architectures. In addition to the ZIH and other researchers at TU Dresden, Friedrich-Alexander-Universität Erlangen-Nürnberg and Forschungszentrum Jülich are also involved in the project. Further information: https://www.mod4comp.de/ (Contact: Dr. Robert Schöne)
HPC introduction at the ZIH now available as a video
The ZIH introductory course in High Performance Computing (HPC) is now also available as a video recording. The content has been divided into thematic chapters and is available online. The videos provide an overview of key topics for getting started with HPC systems, including Linux basics, access, batch systems, monitoring and software and project management. The video recordings complement the ZIH's existing training and consulting services. This includes regular consulting appointments with HPC and NHR experts as well as a wide range of training courses, including courses on performance topics and machine learning. Further training options are also available via the NHR's HPC calendar. Training courses in the field of AI are provided by ScaDS.AI Dresden/Leipzig. Further information: https://doc.zih.tu-dresden.de/#support-and-consultation. (Contact: Tel.: HA -40000)
ZIH Colloquium
On Thursday, April 23 at 15:00 in Willersbau A317 a ZIH colloquium with Prof. Dr. Siegfried Raasch (Chair of Meteorology and Climatology, Leibniz University Hannover; Pecanode GmbH) will take place under the title "The PALM Model System Release 25.10 - An LES Code for Basic and Environmental Research - Capabilities and Recent Applications". The presentation will give an introduction to the treatment of atmospheric turbulence using large-eddy simulations (LES) as well as an overview of the possibilities of the PALM software in environmental and climate research. Further information: https://tu-dresden.de/zih/kolloquium. (Contact: Dr. Hartmut Mix)
Open events
- Softphone consultation hour: 13/20 and 27/04/2026 10:00-11:00 a.m. each day (online)
- Open Q&A session for users of the NHR@TUD computing cluster: 13/20 and 27/04/2026 13:30–14:30 each day (online)
- 16.04.2026, 09:20–10:50: Opal basic course (online)
- 23.04.2026, 09:00–15:00: Girls' Day
- 23.04.2026, 09:20–10:50: ONYX basic course (online)
- 24.04.2026, 16:40–18:10: Pillars and Umbrellas | #5 Reshape Reality - Questioning Claims of Neutrality for Inclusive TechnoFutures
- 30.04.2026, 09:20–10:50: Opal basic course (online)