Design data storage solutions · Topic 4 of 5
Unstructured Data and Storage Accounts
AZ-305 unstructured data: choose Blob, Data Lake Gen2 or Azure Files, access tiers and lifecycle rules, redundancy options, immutability and protection.
- 5 min
- Includes a diagram
- Hot has no minimum. Cool is 30 days, cold 90 days, archive 180 days and offline (rehydration takes hours).
- LRS 11 nines, ZRS 12 nines, GRS and GZRS 16 nines. RA variants add read access to the secondary.
- Archive supports only LRS, GRS and RA-GRS. Azure Files geo redundancy needs HDD shares.
What it is
Unstructured data (documents, images, video, logs, backups, big data files) lives in Azure Storage. A storage account offers Blob storage (objects), Data Lake Storage Gen2 (blob storage with a hierarchical namespace for analytics), Azure Files (managed file shares), queues and tables. The design decisions are the service, performance, access tier, redundancy and protection, balanced against cost.
How it works
Choosing the service
| Need | Choice |
|---|---|
| Store images, documents, backups, media | Blob storage |
| Big data analytics with directories and POSIX like ACLs | Data Lake Storage Gen2 (Blob with hierarchical namespace enabled at account creation) |
| SMB or NFS file shares, lift and shift of file servers | Azure Files (with Azure File Sync to cache on premises) |
| Very high performance enterprise NFS or SMB | Azure NetApp Files |
| Block storage for VMs | Managed disks (Premium SSD, Premium SSD v2, Ultra, Standard) and Elastic SAN for pooled iSCSI |
| Simple message queues | Storage queues |
The hierarchical namespace makes directory rename and delete atomic, and is chosen when the account is created. Use it for data lakes, and use plain Blob for ordinary object storage.
Access tiers and lifecycle
| Tier | Online? | Minimum retention | Use |
|---|---|---|---|
| Hot | Yes | None | Frequently accessed |
| Cool | Yes | 30 days | Infrequent access |
| Cold | Yes | 90 days | Rare access, still online |
| Archive | Offline | 180 days | Long term retention. Must be rehydrated (up to about 15 hours) before reading |
Moving or deleting data before the minimum triggers an early deletion charge. Storage costs fall and access costs rise as the tier gets colder. Use lifecycle management policies to move blobs between tiers or delete them automatically by age or last access. Archive supports only LRS, GRS and RA-GRS redundancy. The account's default tier is hot or cool, and each blob can have its own tier.
Redundancy
| Option | Copies | Durability (per year) | Protects against |
|---|---|---|---|
| LRS | 3 in one datacenter | 11 nines | Drive and rack failure |
| ZRS | 3 across availability zones | 12 nines | Datacenter (zone) failure |
| GRS | LRS plus async copy to a paired region | 16 nines | Regional disaster |
| GZRS | ZRS plus async copy to the paired region | 16 nines | Zone and regional failure |
| RA-GRS / RA-GZRS | GRS or GZRS with read access to the secondary | 16 nines | Same, plus reads during an outage |
In plain GRS or GZRS the secondary is not readable unless you fail over. RA-GZRS gives the highest availability and durability. Azure Files supports geo redundancy only for HDD shares, and SSD shares offer only LRS and ZRS. Managed disks support LRS and ZRS.
Data protection
| Feature | Purpose |
|---|---|
| Soft delete for blobs, containers and file shares | Recover accidental deletion |
| Versioning and point in time restore | Roll back overwrites |
| Immutable storage (WORM) | Time based retention or legal hold so data cannot be changed or deleted, for compliance |
| Encryption | At rest by default with Microsoft managed keys, with customer managed keys as an option |
| Private endpoints, firewall and Entra authorization | Network and identity security. Prefer Entra and managed identities over account keys |
| Defender for Storage | Threat detection, malware scanning |
| Resource lock | Stops deletion of the account |
| Object replication | Copy blobs asynchronously between accounts or regions |
For ransomware and compliance retention, use immutable storage with versioning and soft delete. Use shared access signatures with expiry for limited, delegated access, and user delegation SAS (backed by Entra) rather than account key SAS.
Choosing quickly
| Requirement | Answer |
|---|---|
| Data rarely read, kept 7 years, lowest cost | Archive tier, with a lifecycle policy |
| Survive loss of a datacenter in the region | ZRS |
| Survive a regional disaster and read from the secondary | RA-GZRS |
| Records must not be altered for 6 years | Immutable storage with a time based retention policy |
| Analytics engine with directory structure | Data Lake Storage Gen2 |
| Replace an on premises file server | Azure Files with Azure File Sync |
A worked scenario
A hospital stores scans that are read often for 30 days, rarely for a year, and must be kept unaltered for 10 years. The architect uses a StorageV2 account with RA-GZRS, a lifecycle policy that moves blobs from hot to cool at 30 days, to cold at 90 days and to archive after a year, and immutable storage with a 10 year time based retention. Soft delete, versioning and a private endpoint complete the design.
What to remember for the exam
- Hot has no minimum. Cool is 30 days, cold 90 days, archive 180 days and offline (rehydration takes hours).
- LRS 11 nines, ZRS 12 nines, GRS and GZRS 16 nines. RA variants add read access to the secondary.
- Archive supports only LRS, GRS and RA-GRS. Azure Files geo redundancy needs HDD shares.
- Hierarchical namespace is set at creation for Data Lake Gen2.
- Immutable storage (time based retention or legal hold) provides WORM for compliance.
"Rarely accessed, keep for years, lowest cost" is archive. If it must still be readable immediately, use cool or cold.
How the exam asks about this
- "Keep data for 7 years at the lowest storage cost." Archive tier with lifecycle management.
- "Survive an availability zone failure." Zone redundant storage.
- "Prevent deletion or modification for a set period for regulators." Immutable storage with a retention policy.
- "A data lake with fast directory rename." Data Lake Storage Gen2 with hierarchical namespace.
Common mistakes
Archive for data that is read often. Rehydration is slow and costly.
Plain GRS for read access during an outage. Use RA-GRS.
Account keys everywhere. Use Entra authorization.
Forgetting early deletion charges.
Check yourself
What is the minimum retention in the archive tier?
180 days.
Which redundancy option spans availability zones in one region?
ZRS.
What feature stores data so it cannot be changed for a period?
Immutable (WORM) storage.
Topic 4 of 5 in Design data storage solutions
Test yourself on Design data storage solutions
The full Microsoft AZ-305 question bank has 755 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by Microsoft. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.