DKRZ, Germany’s Climate Archive, Rebuilt for the Exabyte Era

The German Climate Computing Center (DKRZ, Deutsches Klimarechenzentrum) is Germany’s national supercomputing center for climate research. Founded in 1987 by Nobel laureate Klaus Hasselmann and headquartered in Hamburg, DKRZ provides high-performance computing and data services to more than 5,000 climate and earth-system researchers.

Every major climate model, simulation, and dataset produced by Germany’s leading research institutions passes through DKRZ’s systems. Its tape archive holds 254 PB of climate research data, grows by roughly 30 PB every year, and spans two sites, two library vendors, and five generations of LTO media.

Climate science is increasingly a data-reuse discipline. Researchers now revisit decades of simulation output to reproduce results, feed AI-driven weather and climate workflows, and run higher-resolution models that produce data at greater volume, velocity, and file count than any previous system was designed to handle. To support this shift, DKRZ needed an archive that could operate not as a passive preservation tier, but as an active scientific resource. One that is searchable, policy-driven, open, and ready for the next generation of climate research.

DKRZ ran a formal public tender with a demanding list of technical requirements shaped by hard-won operational experience. Versity ScoutAM cleared every mandatory requirement and gave DKRZ what previous systems couldn’t: 100% local administrative control, direct tape access in an open format, and a flexible policy engine engineered from day one for exabyte-scale.

Rewriting 254 PB across five media generations would have been a multi-year, high-risk undertaking. Instead, ScoutAM ingested the existing metadata and treats the legacy archive as a read-only tier. New writes land in the open Versity format on current-generation media, while legacy data remains fully accessible in place, transparent to users.

ScoutAM’s orchestration engine also solves one of the fastest-growing headaches in climate research: small files. Modern formats like Zarr favor huge numbers of small files, a pattern notoriously inefficient on tape. Containerization was a hard requirement in DKRZ’s procurement, and ScoutAM addresses it directly by automatically packing small files into larger objects on tape.

Furthermore, DKRZ’s users now have a direct, self-service layer for accessing the archive. On legacy systems, retrieval of big data sets often meant asking an administrator to run a query and stage files first. Here, scientists can search, locate, and retrieve their own data with automatic metadata extraction and custom tagging built in. The archive shifts from a passive long-term store to an active, queryable knowledge base that plugs directly into modern HPC and AI workflows.