Weather and climate archives can grow by tens of petabytes every year while retaining data for decades. ScoutAM provides the scalable, hardware-independent archive infrastructure to keep that data accessible, readable, and usable through multiple generations of storage technology.
Exabyte-scale in a single namespace, built to move climate data at up to 50 GB/s
Decades of archived data recalled for AI forecasting over S3 or POSIX
Open, standard, tar format on every tape, readable without Versity software
Why is weather and climate data so hard to archive?
Growth never plateaus
Higher resolution and larger ensembles multiply output faster than storage budgets grow. ScoutAM moves finished output off expensive parallel scratch onto tape, keeping the costliest tier small without taking the data out of reach.
Data nobody can predict they’ll need
Climate model output is rarely deleted, because the questions it will answer haven’t been asked yet. A simulation run for one study becomes the comparison baseline for another. Retention is effectively permanent, so access matters as much as capacity.
Observations Cannot Be Regenerated
A simulation can be re-run. An atmospheric measurement taken at a specific place and moment cannot. Observational records, satellite passes, and field campaign data are irreplaceable in ways that most enterprise data is not, and they need preservation guarantees that match.
Archives that outlive their software
National meterological services and climate centres hold data under mandates that run far longer than any hardware refresh cycle. ScoutAM’s policy-driven retention, replication, and integrity checking meet those obligations while the open format ensures the data survives the software that wrote it.
ScoutAM can read existing archive formats, so sites convert by importing metadata instead of copying petabytes of data.
Learn more about ScoutAM →Quiesce it to suspend user access and stop new writes.
Capture file names, paths, permissions, and media locations using system-specific commands.
Built-in commands preserve file structures and permissions.
Existing archival media is automatically added to ScoutAM’s catalog as read-only, guaranteeing integrity and backward compatibility.
Users read and write via ScoutAM using S3 or POSIX. Everything written from that point forward is in open tar format.
A modern weather and climate data archive should ingest HPC output at full speed, keep decades of data reachable without manual staging, store everything in an open format, and serve users through standard protocols. ScoutAM is the only one that ticks all the boxes.
Archives stop being useful when they become the bottleneck between compute and research. ScoutAM delivers throughput of up to 50 GB/s, keeping data moving efficiently between high-performance storage and tape with scalable cache infrastructure.
Getting data back should be as easy as archiving it. Researchers search the archive’s metadata to pinpoint the exact files a study or training run needs, then recall them straight to high-performance storage, without staging whole directories or waiting on an administrator. Only the data you asked for comes back, so tape time, cache space and compute hours go to the work itself.
ScoutAM asynchronously replicates metadata, cache, and archive data to one or more secondary sites. High availability, robust error handling, and a rebuildable metadata index help keep data protected and operations running through hardware failures.
ScoutAM writes data in an open, standard tar format. Any system that reads tar can read your archive, with or without Versity. Our software works with tape libraries from Spectra Logic, IBM, BDT, Quantum, and other vendors, object storage, cloud and disk. You choose the hardware; the archive software doesn’t choose it for you.
Users don’t need special tools or training. ScoutAM integrates with the access methods research organizations already use — Globus, S3, NFS, Samba, scp and rsync — and with job schedulers through prologue and epilogue scripts. Support comes with direct access to engineers who know the platform.
Trusted by the World’s Climate and Weather Institutions
Climate and weather organizations rely on Versity to manage rapidly growing research data at scale, keeping decades of scientific data accessible while freeing high-performance storage for active workloads.
An Exabyte of Climate Simulation, Zero Data Migration
DKRZ supports climate research across Germany and international modeling programs. Its archive holds about 64 million files, growing by roughly 30 PB a year.
After more than a decade on proprietary HSM platforms, DKRZ selected ScoutAM, delivered with NEC through a public tender, to manage its entire tape infrastructure: ten tape libraries, 130 LTO drives and capacity for up to 100,000 cartridges. The system offloads from Lustre to tape at up to 30 GB/s.
Protecting a National Environmental Record
ESNZ runs Rapids, an archive of two Spectra Logic TFinity tape libraries managed by ScoutAM and delivered with Xenon. It keeps more than 25 years of weather and climate data available for training AI forecasting models.
ESNZ chose ScoutAM for its explicit archiving model, which separates the filesystem from tape so the team decides what gets archived and when. Open formats mean ESNZ keeps independent access to its data, and automatic replication between primary and secondary sites protects the national record.
Direct GPU-to-Tape Transfer
The Versity GPUDirect Tape Gateway adds S3 over RDMA through NVIDIA’s cuObject framework and pairs it with ScoutAM, so data moves from tape toward GPUs without the usual CPU bottlenecks.
Historical weather data supports forecasting, climate research, validation, and AI training long after it leaves primary storage. ScoutAM keeps that data protected, accessible, and ready to move when it is needed again. Talk to Versity about your next decade of data growth.
