What Are Backup Storage Tiers?

Backup storage tiers are the different classes of storage a backup copy can live on, each trading cost per terabyte against how quickly and cheaply the data can be read back. Tiering is how organizations hold years of retention without paying primary-storage prices for all of it.

The distinction that matters most is this: tiers are usually chosen on cost per terabyte stored, but they are paid for on cost and time to read back. A tier that looks inexpensive on the invoice can be the reason a recovery takes three days, and that trade only becomes visible during an incident.

The common tiers

Terminology varies between vendors, but the ladder is consistent.

• Hot / performance. Fast, always-online storage for recent backups and anything with a short recovery time objective. Highest cost per terabyte, lowest cost to read.
• Warm / standard. Online and reasonably fast, used for the bulk of operational retention over weeks and months.
• Cool / infrequent access. Cheaper per terabyte, but with retrieval fees and often a minimum storage duration that penalizes early deletion.
• Cold / archive. Lowest storage price, with retrieval measured in minutes to hours and per-gigabyte egress charges. Suited to compliance retention, not operational recovery.
• Offline / tape. Very low cost per terabyte and inherently disconnected, but slow, manual, and awkward to test.

What cold tiers really cost

Archive pricing is quoted per terabyte per month, which is the number that gets compared. The costs that matter during a recovery sit elsewhere.

• Retrieval latency. A copy that takes hours to become readable adds that time to your recovery before any data moves, which is why restore throughput numbers are meaningless if you exclude the wait.
• Egress and retrieval fees. Reading a large dataset out of an archive tier can cost more in one incident than the storage saved over a year.
• Minimum duration charges. Deleting or moving data early often incurs the full minimum retention charge anyway, which quietly undermines a retention policy built around shorter cycles.
• Untested by default. Cold copies are expensive to read, so they get excluded from restore testing — meaning the copy you rely on longest is the one least often proven.

Matching tiers to recovery targets

The workable approach is to let recovery requirements drive placement rather than picking a tier and hoping.

• Start from the RTO. Anything with a recovery target measured in hours cannot live in a tier with multi-hour retrieval latency.
• Keep the operational window hot. Recent restore points — the ones most incidents need — belong on fast storage where instant recovery is possible.
• Reserve cold for compliance, not recovery. Use archive tiers for data you must keep but do not expect to restore under time pressure.
• Keep the immutable copy reachable. The 3-2-1-1-0 backup rule requires an immutable copy, and it only helps if you can actually read it quickly after an attack.
• Model the incident, not the month. Compare tiers on the total cost and time of a realistic full recovery, not on monthly storage price alone.

Backup storage tiers and ARTESCA

Scality ARTESCA is backup-first object storage you run on standard servers, which changes the economics that make tiering necessary in the first place.

• Performance at retention prices. Running on commodity hardware means a fast, always-online tier for far more of your retention, rather than pushing data cold to control cost.
• No retrieval penalties. On-premises object storage has no egress or retrieval fees, so a large recovery costs the same as a small one.
• Immutability on the fast tier. S3 Object Lock applies to online copies, so the protected copy is also the quickly recoverable one.
• Tier when it makes sense. Standard S3 semantics mean ARTESCA works alongside cloud or tape for genuine long-term archive, without those tiers becoming your recovery path.

The result is fewer copies stranded in storage classes you cannot afford to read.

Backup storage tier FAQs

Is cold storage a bad choice for backups?

Not inherently — it is a poor choice for recovery. Archive tiers are well suited to long-term compliance retention where you rarely read the data. Problems arise when they hold the copies you would need during an outage, because retrieval latency and fees turn a recovery into a slow, expensive process.

How many tiers should we use?

As few as your requirements allow. Every tier adds policy complexity, another place for a copy to be misplaced, and another retrieval behavior to remember under pressure. Two — a fast operational tier and a genuine long-term archive — is enough for most environments.

Does tiering affect immutability?

It can. Immutability guarantees and lock modes differ between storage classes and providers, and moving an object between tiers may reset or fail to preserve its lock. Verify that your immutable copy keeps its protection wherever it physically lands.

What is the most common tiering mistake?

Choosing tiers on monthly storage price alone. The invoice compares cost per terabyte stored, while the incident is decided by cost and time per terabyte read. Modeling a full recovery against each tier usually changes the decision.