Glossary

Storage tiering

Storage tiering places data on different types of storage according to how often it is used: fast media for active data, cheaper media for the rest. Policies move data between tiers as its use changes, so that most capacity sits on low-cost media.

What are the tiers?

Layouts vary, but a common ladder has four steps. Flash holds the active data, as described under hot storage. High-capacity hard disk holds data read less often. Dense capacity systems hold the large bulk of data that is kept for reference. At the bottom sit tape and cloud archive for data that is rarely touched. Each step down costs less per terabyte and responds more slowly.

How does data move between them?

Policies decide. A rule might demote data that has not been read for thirty days, or move objects after a set age. Storage lifecycle management describes the rules and the schedule. Some systems move blocks transparently inside a single volume, and others move whole files or objects between separate systems, leaving a stub or an index entry behind.

The scale of the saving depends on how much data is idle. In most estates a large share of capacity has not been touched in a long time, which is why the price gap between the top and bottom tiers matters at petabyte scale.

What does tiering down not account for?

The way back. Archive tiers add retrieval delays, egress fees, per-object charges and minimum storage periods. Small files are hit hardest, since per-object fees do not shrink with file size. A tier that looks cheap on the monthly invoice can cost more than the tier above it the first time a large amount of data is called back.

For that reason, data that an AI project or a restore might need often belongs on low-cost online disk rather than an offline archive. Cold storage covers the recall costs in detail.

Is tiering the same as caching?

No. Tiering moves the data, so one copy exists at a time, in the tier suited to its use. A flash cache copies data into fast media while the original stays where it is. The two combine in a hybrid flash array.

Frequently asked questions

What triggers a move between tiers?

Access frequency, age, size or an explicit policy. Most systems use time since last access.

Can tiering be automatic?

Yes. Policy engines or system software move data on a schedule. Some leave the application unaware, and some require applications to follow a stub.

Does tiering reduce the amount of data stored?

No. It changes where data sits. Reducing volume is the job of compression and deduplication.