Speaker
Description
The High Energy Photon Source (HEPS) is a fourth-generation synchrotron radiation facility scheduled to enter formal operation in June 2026. Enabled by high brightness and advanced detector technologies, HEPS is expected to generate more than 200 PB of experimental data annually across 14 Phase-I beamlines. The wide range of experimental techniques, including imaging, diffraction, scattering, and spectroscopy, produces highly heterogeneous data in throughput, volume, and processing complexity, posing significant challenges for large-scale scientific data processing.
To address these challenges, we developed DAISY (Data Analysis Integrated Software System), a general-purpose scientific data processing framework. DAISY provides high-throughput data I/O, unified access to heterogeneous data sources, and support for elastic and heterogeneous computing. Based on DAISY, multiple domain-specific scientific applications have been developed and deployed on HEPS beamlines, where they support beamline commissioning, beam tuning, and process validation through online and near-real-time data processing. Joint data analysis approaches across multiple experimental methodologies are being actively explored to support increasingly complex scientific use cases, and future work will further incorporate artificial intelligence methods to enhance the efficiency, accuracy, and automation of scientific data processing.
This contribution reports on the design and implementation of the DAISY framework, the application of DAISY-based scientific software across multiple HEPS beamlines, and data processing practices and experiences toward the formal operation of HEPS.
