Flagship future physical science research infrastructures, such as CERN’s Large Hadron Collider (HL-LHC) and the SKA Observatory (SKAO), will generate massive quantities of raw data, measured in Exabytes (1018 bytes) per year, far greater than all the data generated by a global service such as Facebook. This flood of data paves the way for major scientific breakthroughs, but poses a crucial challenge: current digital technologies are no longer able to process these massive data streams, and major discoveries may go unnoticed. The ODISSEE project tackles this crucial issue by proposing innovative digital technological and methodological solutions to reveal new scientific knowledge.

ODISSEE is a €9.5 million research project funded with €8.3 million from the European Union, including three emblematic infrastructures of the European Strategic Forum for Research Infrastructure (ESFRI) program: HL-LHC, SKAO and SLICES-RI. Coordinated by Observatoire de Paris, the three-year ODISSEE consortium brings together 14 partners and 2 associate partners from academia and industry. By involving such a diverse pool of partners, the project will be able to design technological solutions for the benefit of research that can be directly applied in the industrial world, thereby helping to boost European industry competitiveness and European digital sovereignty. The project was launched at a kick-off meeting at Observatoire de Paris, in Meudon, France on 13 and 14 January.

ODISSEE proposes an innovative methodological and technological approach to meet the needs of HL-LHC and SKAO scientific communities:
· On-the-fly data processing using eco-responsbible artificial intelligence for efficient selection of relevant data and trustworthy preservation of actionable data products;
· Development of dedicated hardware and open, non-proprietary software solutions that are both energy-efficient and flexible, to facilitate their adaptability to unknown situations;
· Exploit and enhance the European High Performance Computing (HPC) ecosystem, anticipating the needs of next-generation Exascale supercomputers.
These challenges can be more efficiently addressed thanks to the capabilities of SLICES RI, which offers a unique experimental platform for component configuration, testing calculations and user training.

To demonstrate the effectiveness of these new technologies, ODISSEE will work on specifiying a pilot combining the use of both SKAO and HL-LHC: the search for the nature of dark matter. Identifying the nature of this fundamental mystery of the universe, which accounts for a large proportion of its mass but remains invisible, requires to combine the capabilities of both infrastructures, setting new standards in terms of data handling and operational models.

“We are entering into the era of the data deluge, where the volumes generated by our new research infrastructures such as the SKAO or CERN’s HL LHC exceed anything we could have imagined just a few years ago. To take full advantage of these planetary infrastructures, we need to propose and validate new solutions, combining innovations in hardware and software, enabling more flexible data processing as well as trustworthy preservation of relevant and actionable data products.”, said Damien Gratadour, Research Scientist at CNRS, Paris Observatory and coordinator of the ODISSEE project.

« CERN looks forward to partnering with SKAO, and our ODISSEE collaborators, to unlock the full scientific potential of exabytes of data. By leading the way in AI and data classification techniques, developing specialised hardware and software solutions, and exploiting the full capabilities of European HPC resources, we will not only shape our own research domains, but also drive innovation across the broader scientific community.”, said Dr. Maria Girone, Head of CERN openlab.

« Thanks to its flexibility and diversity, SLICES will provide an experimental platform that can be tailored to the needs of ODISSEE. The solutions developed within the project could thus be studied and validated before entering into production.« , said Christian Perez, Inria.

“At SiPearl, we are proud to be part of ODISSEE consortium. With our Rhea-based server, we will contribute to codesign energy-efficient building blocks interoperable with ODISSEE infrastructure. We will also contribute to improve the software ecosystem to better exploit heterogeneity and parallelism of hardware resources for HPC and AI codes”, said Philippe Notton, CEO & Founder of SiPearl.

« At NextSilicon, we are honored to contribute to the ODISSEE project, which addresses extreme-scale data challenges like those exemplified by CERN’s HL-LHC and SKAO. Our Maverick-2 Intelligent Compute Accelerator (ICA) architecture is purpose-built for these demanding, high-precision compute workloads, enabling researchers to extract accurate, actionable insights without compromise. We are excited to collaborate with leading partners to advance Europe’s digital leadership and support the next generation of scientific breakthroughs through intelligent and efficient computation.« , said Elad Raz, Founder and CEO, NextSilicon.

Jacques Tissot, Senior Communication Expert at Neovia Innovation
jacques.tissot[at]neovia.innovation.eu