Choosing the Right Azure Storage Tier for Large File Migrations
The complete decision framework for storage tier selection during large file migrations — Azure Files tiers, Blob tiers, cost modelling, early deletion traps, lifecycle automation, and the 2026 changes that affect every migration running today.
The most expensive mistake I see on large Azure file migrations is not choosing the wrong tool or running the migration in the wrong order. It is choosing the wrong storage tier. Pick Premium SSD for workloads that would perform perfectly well on Standard HDD and you spend six times more than necessary. Pick Cool or Cold for data that turns out to be accessed more frequently than estimated and the retrieval fees exceed what Hot would have cost. Pick Archive for data that needs to be recalled within an hour and you discover the hard way that standard rehydration takes up to fifteen hours. Storage tier selection is not a configuration detail — it is a cost architecture decision with consequences that compound for years. This guide is the decision framework I use on every migration I run.
The full picture — all Azure storage tiers in one view
Most migration guides discuss Azure Files tiers or Azure Blob tiers — not both together. Large file migrations frequently involve both services simultaneously: Azure Files for active SMB shares that need to be mounted, Azure Blob for backups, archives, and object data that is accessed via API or REST. Understanding both tier families — and knowing which service is appropriate for each workload type — is the prerequisite to choosing the right tier.
The four Blob tiers — what each one actually costs and when to use it
Azure Blob Storage's tier system is elegant in principle — pay less per GB as you access data less frequently — and treacherous in practice, because the operation costs move in the opposite direction to the storage costs. The tier that looks cheapest per GB can easily become the most expensive option if your access pattern does not match the assumption behind the tier pricing.
Starting July 1, 2026 for new storage accounts (July 1, 2027 for all accounts), Cool, Cold, and Archive tiers apply a 128 KiB minimum billable object size. A 4 KB log file in Cool storage is billed as if it were 128 KiB — a 32× billing multiplier.
If your large file migration includes small files (log files, metadata files, small documents), and you plan to put those files in Cool, Cold, or Archive, you must calculate effective storage cost using the 128 KiB minimum — not the actual file size. For a dataset of one million 4 KB files in Cool storage, you would be billed for 128 GB instead of 4 GB.
The rule: Cool, Cold, and Archive are appropriate for large files (>128 KB) — for small-object workloads, Hot tier is often cheaper in total despite the higher per-GB storage cost.
The tier selection decision flowchart — step by step
Use this flowchart for every share and dataset type in your migration. Run it once per share — not once per migration. A 15TB migration across eight shares may correctly route three shares to Premium SSD, three to Standard Hot, and two to Blob Archive.
Step by step: mapping your shares to the right tier
Classify each share by access pattern and file size distribution
For every share in your migration, capture three data points: average access frequency per file per month (from NAS access logs or Azure Migrate telemetry), median file size (determines whether the 128 KiB minimum object size penalty applies), and whether the share needs to be mounted via SMB/NFS (determines whether Azure Files or Blob Storage is the right service). This three-question classification drives every tier decision downstream. Do not estimate access frequency — measure it. NAS access logs are the ground truth.
Apply the service selection rule: Azure Files vs Blob Storage
If the share is accessed via mounted drive letter, UNC path, or SMB/NFS protocol — it must go to Azure Files. If the share contains backup data, object data, media files, or anything accessed via API or REST — it should go to Azure Blob Storage. Azure Files is 3× more expensive per GB than Blob Hot tier ($0.06 vs $0.018), so putting backup data in Azure Files instead of Blob Storage means paying three times more than necessary for the same storage. This is a systematic over-spend that affects every organisation that migrates NAS backup data to Azure Files without reviewing the service choice.
Model total monthly cost — not just storage cost per GB
For each share assigned to a Blob tier, calculate total monthly cost as: (storage cost × GB) + (retrieval cost × GB accessed per month) + (operation cost × operations per month). The crossover analysis in Figure 2 shows where Cool and Cold become more expensive than Hot. For most enterprise file shares accessed more than three times per month per file, Blob Hot is the economically correct choice even though it has the highest per-GB storage cost. Run the model before committing to a tier — not after you receive the first invoice.
Identify data that may be moved or deleted before minimum retention periods expire
Cool (30 days), Cold (90 days), and Archive (180 days) tiers all impose minimum retention periods. Data deleted or moved before the minimum period triggers an early deletion charge prorated over the remaining days. For migration staging data — files that are uploaded temporarily as part of a migration workflow — Cool, Cold, and Archive are dangerous choices. Use Hot for all transient and staging data. Use Cool and colder tiers only for data you are confident will not be moved, overwritten, or deleted within the minimum retention window.
Configure lifecycle management policies — let Azure manage tier transitions automatically
Azure Blob Storage supports lifecycle management policies that automatically move blobs between tiers based on age, last-modified date, or last-accessed date. A well-designed lifecycle policy starts data in Hot tier, moves it to Cool after 30 days, to Cold after 90 days, and to Archive after 180 days — without requiring manual intervention. This is the correct operating model for large Blob datasets: start all migrated data in Hot, let the policy age it through tiers automatically, and review the cost impact of the policy quarterly.
If using Archive tier — plan your rehydration strategy before data enters Archive
Archive tier blobs are completely offline. Reading archived data requires rehydrating the blob to Hot or Cool tier first — a process that takes up to 15 hours at standard priority, or under 1 hour at high priority (which costs $0.03/GB in addition to the standard retrieval fee). Plan your rehydration strategy before data enters Archive: which data might need emergency access, what the acceptable retrieval time is, and whether high-priority rehydration costs are budgeted for unplanned access events. If you cannot answer these questions with confidence, Cold tier is a safer choice than Archive.
Lifecycle policy: automating tier transitions after migration
The ideal tier selection at migration time is Hot. Then let lifecycle policy move data through tiers automatically as it ages. This approach avoids the early deletion penalty risk of choosing a cold tier at upload time, while still achieving the long-term cost savings of colder tiers as data ages.
"rules": [{
"name": "migration-tier-policy",
"type": "Lifecycle",
"definition": {
"filters": { "blobTypes": ["blockBlob"] },
"actions": {
"baseBlob": {
// Move to Cool after 30 days of no modification
"tierToCool": { "daysAfterModificationGreaterThan": 30 },
// Move to Cold after 90 days
"tierToCold": { "daysAfterModificationGreaterThan": 90 },
// Archive after 180 days — plan rehydration before this
"tierToArchive": { "daysAfterModificationGreaterThan": 180 },
// Optional: delete after 2 years for compliance-bound data
"delete": { "daysAfterModificationGreaterThan": 730 }
}
}
}
}]
}
Applying the framework: tier decision worksheet for a 15TB migration
| Share / Dataset | Access freq | SMB needed? | Avg file size | Service | Tier | Reason |
|---|---|---|---|---|---|---|
| Finance (3.2TB) | Daily · heavy | Yes (mapped drive) | >100KB | Azure Files | Premium SSD | Finance apps require <2ms latency · 3K+ concurrent ops |
| Projects (4.1TB) | Daily · moderate | Yes (mapped drive) | >100KB | Azure Files | Standard Hot | Active but latency <10ms acceptable · no IOPS pressure |
| HR (0.8TB) | Daily | Yes (mapped drive) | Mixed | Azure Files | Standard Hot | Sensitive data · active daily use · Hot cost justified |
| Shared (1.9TB) | Daily · low intensity | Yes (mapped drive) | >100KB | Azure Files | Standard Cool | Read-mostly, 1–2×/week per file · Cool saves vs Hot |
| Media (0.9TB) | Weekly | Yes (media server mount) | Large (video) | Azure Files | Standard Cool | Weekly access · large files → no 128KB penalty · Cool saves |
| Backups (0.3TB) | <Monthly | No (automated scripts) | >1MB each | Blob Storage | Blob Cold | API access only · large files · accessed <quarterly |
| Archive2022 (2.1TB) | <Annually | No | >100KB | Blob Storage | Blob Archive | Compliance hold · never read unless audit · 15hr wait acceptable |
| Archive2023 (1.7TB) | <Annually | No | >100KB | Blob Storage | Blob Archive | Same as Archive2022 · will apply lifecycle policy for further aging |
On our 15TB migration, the original plan would have put all eight shares in Azure Files Standard Hot — the path of least resistance. The tier analysis above shows that Finance should actually go to Premium SSD (better performance, slightly higher cost justified by the workload), Archive2022 and Archive2023 should go to Blob Archive (not Azure Files at all — saving approximately $60/month versus Azure Files Cool), and Backups should go to Blob Cold (not Azure Files, saving approximately $15/month).
The total monthly cost difference between the unanalyzed plan (everything in Azure Files Standard Hot) and the correctly tiered plan is approximately $85/month — $1,020/year — on a 15TB migration. On a 50TB migration with similar data distribution, the difference would be $3,400/year. The tier analysis takes two hours. It pays for itself in month one.
"Access tiers, retrieval charges, and early deletion penalties routinely add 30–70% to a theoretical storage bill. Teams that choose a tier based on per-GB cost alone often end up paying more than the Hot tier would have cost."
— Finout 2026 Cloud Storage Cost AnalysisGet the Azure field notes — every week
Storage tier decisions, cost modelling, migration guides. Written by Francis Avorgbedor.