Workload, not vertical branding
A sector label tells you nothing. File size, ingest rate and who set the retention period tell you almost everything.
The same four architectural layers get assembled differently for a hospital, a pipeline operator and a broadcaster, because their data behaves differently.
These pages describe sector workloads and constraints. They do not claim named client deployments in each sector.
There is no such thing as a healthcare storage array. There are arrays that cope with millions of small DICOM objects and a radiologist waiting for a prior study, which is a workload description that happens to occur most often in hospitals. Sector pages are useful only when they say something concrete about that workload. The seven below are written that way: what the data looks like, how fast it arrives, how quickly it must come back, and who has the authority to say how long it is kept.
How workload shapes the architecture
Four properties drive most of the design, and they cut across industries rather than following them.
- Object size and count. Billions of small files and a handful of very large ones stress completely different parts of a system. Metadata handling usually fails before capacity does.
- Ingest shape. Continuous streams, overnight batches and unpredictable bursts each imply a different cache and drive count.
- Recall expectation. Whether a retrieval taking two minutes is routine or a service failure decides how much stays on disk.
- Retention authority and residency. Whether the clock is set by a regulator, a contract or internal policy, and whether the data may leave the country. Interpreting that obligation is work for your legal advisers.
What differs between sectors
| Sector | The property that dominates |
|---|---|
| Government | Long-horizon retention set by a records authority, plus strict residency expectations |
| Healthcare and medical imaging | Huge object counts with clinician-facing recall times |
| Banking and financial services | Evidential integrity and auditor-facing proof of enforcement |
| Energy, oil and gas | Very large datasets kept for field life, often at remote sites |
| Telecommunications | Sustained high-volume record generation and rapid growth |
| Media and video surveillance | Continuous write with occasional urgent retrieval of a single clip |
| Research, AI and HPC | Reproducibility, dataset provenance and unpredictable re-reads |
Where your workload sits between these matters more than the label on your organisation. An assessment establishes it, and the solutions section covers the risks each pattern creates.
Explore industries
Industry
Government & Public Sector
Aban Smart designs archive infrastructure for public bodies where data location, classification handling and disposal authority constrain the design before capacity does.
Learn moreIndustry
Healthcare & Medical Imaging
Aban Smart designs imaging archive tiers around what your scanners actually generate, so cache, capacity and retention are sized from modality data rather than a single average.
Learn moreIndustry
Banking & Financial Services
Aban Smart builds retention infrastructure for financial institutions where the archive must survive challenge, not just retrieval, with the evidence trail an examiner will ask for.
Learn moreIndustry
Energy, Oil & Gas
Aban Smart builds archive tiers for subsurface data where the working unit is enormous, the retention horizon outlasts several hardware generations, and old data returns to active use.
Learn moreIndustry
Telecommunications
Telecom archives are not a few enormous files. They are an unimaginable number of very small ones, and that inversion decides almost every design choice that follows.
Learn moreIndustry
Media & Video Surveillance
One is written constantly and almost never watched. The other is watched, cut and re-sold for decades. Designing one archive as if it were the other is how these projects go wrong.
Learn moreIndustry
Research, AI & HPC
The next person to open it is a reviewer, a doctoral student or a retraining job, working from a citation and whatever survived of the metadata.
Learn moreFrequently asked questions
We describe workloads honestly rather than implying a client list. One manufacturer reference we can point to is a QStar and DISC deployment at Nottingham City Transport for onboard video retention, which is the manufacturer's reference rather than an Aban Smart customer. Beyond that, judge us on whether the sector page describes your constraints accurately.
Related pages
Turn your requirement into a defensible architecture
Share the workload, capacity, retention, access, and resilience requirements. Aban Smart will identify the next discovery inputs and the appropriate engagement path.