smartData Protection & Archiving

Five risks, five architectures

Pick the failure you are trying to avoid, and the technology argument becomes much shorter.

Immutability, longevity, retention, recovery and legacy replacement are separate problems. Most estates need two or three of them, rarely all five at once.

These five pages describe what each approach does not cover as well as what it does.

Enterprise data-protection architectureData sources flow through a policy and management layer to storage tiers spanning disk, object, tape, optical, and cloud, producing governed and protected copies.DATA SOURCESApplicationsFile & NASS3 / ObjectBackup jobsPOLICY LAYERClassifyApply retentionMove & tierAudit & recallSTORAGE TIERSDisk / Flash cacheObject storageLTO tapeOptical WORMWORMCloud / S3
Conceptual architecture. Final topology depends on verified product compatibility.
Storage tier ladderFive storage tiers stepping down from disk and flash through object storage, LTO tape, optical WORM and cloud. Access latency increases down the ladder and retention suitability lengthens from days to decades. LTO tape is the lowest cost per terabyte; optical WORM costs more per terabyte but is the immutable, longest-retention tier.STORAGE TIERCOST / TBRETENTION FITACCESS LATENCY: FAST → SLOWDisk & flashSub-millisecond recall · always onlineHighestDays–weeksObject storageMilliseconds · always onlineHighMonths–yearsLTO tapeSeconds–minutes · offline capableLowest10–30 yearsOptical WORMSeconds–minutes · offline, write-onceIMMUTABLEHigh50–100 yearsCloud / S3Network-bound · egress fees applyOp-exPolicy-setDown the ladder recall takes longer and retention lengthens. Tape is the lowest cost per terabyte; optical buys lifespan, not cheap capacity.
Conceptual storage tier comparison. The cost and retention bar lengths are illustrative of relative position only, not measured pricing. Final topology depends on verified product compatibility.

Where each tier earns its place.

Archive requirements are usually described in terms of a product category, which is why so many of them get solved badly. "We need WORM" and "we need an archive" can each mean at least three different things depending on which failure the buyer is actually worried about. Naming the risk first narrows the field faster than any feature comparison. What follows sorts our work into five risks. Each links to a page that goes into the mechanisms, the trade-offs and the residual exposure in detail.

Start from the risk, not the product

The risk Where to read next
Data can be altered or deleted early, including by a privileged account Immutable and WORM storage
Data survives but nothing can read it in fifteen years Long-term data archiving
A retention rule exists on paper and nothing enforces the clock Regulatory data retention
Copies exist but recovery misses the time the business was promised Backup and recovery storage
The platform holding the archive is out of support or unreadable Archive migration and modernisation

How the five areas fit together

They are layers of the same estate rather than alternatives. Enforcement decides what cannot be removed. Longevity decides what remains readable. Retention decides how long the clock runs and who set it. Recovery decides how fast a copy returns. Migration is what happens when one of the first four was solved on a platform that has now aged out.

Working from left to right in that order avoids the most expensive mistake, which is buying capacity that enforces nothing. If you are not sure which of the five applies, an assessment sorts it, and the technologies section covers the components underneath.

Explore solutions

Frequently asked questions

Whichever one an auditor, an incident or a support expiry has already raised. Where nothing external is forcing the order, enforcement and recovery come first, because both are cheap to specify at design time and expensive to retrofit. Longevity and migration matter most for estates that already hold years of data on ageing media.

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.