ICT developments for Elettra 2.0 and beyond Facts & Figures Elettra Highlights 2024–2025 170 During 2024 and 2025 Elettra has significantly consolidated and enhanced its ICT infrastructure to prepare for the forthcoming challenges of Elettra 2.0, aligning with FAIR principles and modern security standards. The Scientific Computing Activity has been split in Business Intelligence and Scientific and Quantum Computing to better support science, research and the whole scientific business. Among the most noteworthy advancements, the introduction of the Elettra Scientific Data Lake (EDL) stands out as a strategic evolution for managing the diverse, complex data generated by synchrotron facilities. The Elettra Data Lake (figure 1) embodies a tailored adaptation of the innovative data lakehouse architecture, effectively merging the flexibility of data lakes with the performance and governance typical of data warehouses. The architecture facilitates the seamless management of heterogeneous scientific data—ranging from TIFF images and crystallographic files (CIF) to proprietary data formats— maintaining full adherence to FAIR principles through the strategic use of HDF5 container formats and Digital Object Identifiers (DOIs). EDL operates entirely on-site, spanning edge computing devices at beamlines to high-performance computing (HPC) clusters, ensuring the required data sovereignty and real-time processing capabilities critical for experimental workflows. The infrastructure incorporates diverse computational resources, including GPU-accelerated nodes dedicated to machine learning and data-intensive reconstructions, supported by a sophisticated scheduling system. An IBM Spectrum Archive tape library, integrated with customdeveloped software featuring RESTful APIs and robust data integrity verification processes, complements the tiered storage strategy. Complementing these technical advancements, a suite of tailored webbased applications has been deployed. These include Enodo (enodo.elettra. eu), a platform inspired by Zenodo and WeTransfer for the association and publication of datasets with persistent DOIs, facilitating open scientific dissemination (figure 2). Other innovative tools like XRFitVis and 4DHorizon provide interactive, intuitive visualization for X-ray Fluorescence and AngleResolved Photoemission Spectroscopy data, respectively, significantly enhancing data interpretability and user engagement (figure 3). Parallel to these developments, the IT Group strengthened general ICT infrastructure by replacing obsolete network and server equipment, significantly improving performance and reliability. A substantial expansion of high-speed virtual storage and enhanced parallelization of tape library operations have bolstered both data backup and long-term scientific data archival processes. Furthermore, the high-performance computing (HPC) infrastructure has undergone comprehensive software updates to ensure compatibility with modern computational and simulation software, with GPU nodes now also supporting local deployment of advanced AI services, including Large Language Models (LLMs). A notable achievement was obtaining ISO27001 certification, underscoring a commitment to data security, compliance, and quality management. Concurrently, preparations for aligning the infrastructure with the new NIS2 directive have begun, reinforcing Elettra's proactive approach to cybersecurity and regulatory compliance. To further support the Elettra 2.0 project management and logistics, the Virtual Unified Office (VUO) has introduced specialized tools compliant with BIM specifications. Notably, the ACDAT system, developed in collaboration with the University of Padua, ensures detailed management and documentation of technical components, project documents, and procurement workflows, fostering enhanced traceability and efficiency. In the context of Elettra 2.0, a novel interactive mapping tool named PAnTHer (Particle Accelerator on THree.js) has been developed. This innovative tool leverages advanced web and touch technologies to create dynamic 2D and 3D maps of particle accelerators. These maps integrate seamlessly with real-time data sourced from accelerator control systems, simulation outputs, and external component databases, delivering an optimal balance of visual fidelity and performance, significantly enhancing user experience. In the same context the digital twin of Elettra 2.0 has been developed using the most advanced available tecnology and release for full use by machine physicists. Additionally, the IT Group actively supports the data acquisition processes across various Elettra and FERMI beamlines. The Tango control system, paired with specialized Python tools such as the DonkiOrchestra framework, facilitates a wide spectrum of experimental techniques, particularly in Imaging, Diffraction, and Absorption beamlines (figure 4). As an active participant in international collaborations MXCuBE and ISPyB, Elettra contributes to the development of common software tools for diffraction beamlines. For post-acquisition activities, the versatile Python-based tool RunScriptWeb has been introduced to efficiently manage real-time data processing and sharing phases. Furthermore, the old Beamline Control System (BCS) has been replaced by the new GeCo Elettra 2.0 Beamline Interlock System on eight beamlines, employing a robust and streamlined Siemens PLC infrastructure interfaced via Tango, ensuring seamless integration with end-station software. The DonkiWeb synoptic tool has also demonstrated significant robustness and efficacy, with newly introduced features based on user feedback. In conclusion, the introduction of the Elettra Data Lake, combined with comprehensive infrastructure upgrades and strategic applications promise to accelerate discovery, optimize experimental workflows, and robustly support the ambitious scientific goals of Elettra 2.0.
RkJQdWJsaXNoZXIy MjQ4NzI=