Azure Storage Services
Azure offers multiple storage types — blobs for unstructured data, files for shared drives, tables for NoSQL, and queues for messaging. Plus redundancy options that keep your data safe across datacenters and regions.
Azure Storage — where your data lives
Think of Azure Storage as a self-storage facility with different locker types.
Blob storage = a giant warehouse for anything — photos, videos, backups, documents. Just throw it in. It’s like a massive box room with no structure.
File storage = a shared filing cabinet. Multiple people can open the same drawer and access the same files, just like a network drive at work.
Table storage = a simple spreadsheet. Rows and columns for basic data that doesn’t need a full database.
Queue storage = a post office. Messages wait in a queue until someone picks them up and processes them.
And the best part? Azure makes copies of your data automatically so it’s never lost.
Storage services compared
| Feature | Blob Storage | Azure Files | Table Storage | Queue Storage |
|---|---|---|---|---|
| What it stores | Unstructured data (any binary or text) | Files accessible via SMB/NFS | NoSQL key-value pairs | Messages (up to 64 KB each) |
| Access method | REST API, SDKs, Azure CLI | SMB (like a Windows file share) | REST API, OData | REST API |
| Use case | Images, videos, backups, logs | Shared drives, config files, lift-and-shift | Simple data, IoT telemetry | Decoupling app components |
| Analogy | Warehouse box room | Shared filing cabinet | Simple spreadsheet | Post office queue |
Blob Storage — the most common
Blob stands for Binary Large Object. It stores any type of unstructured data.
Three types of blobs:
| Type | Best For |
|---|---|
| Block blobs | Files up to ~190 TB (images, videos, documents) |
| Append blobs | Data that’s always added to (log files, audit trails) |
| Page blobs | Random read/write operations (VM disks) |
Peak Roasters uses Blob Storage for product images, menu PDFs, and nightly database backups.
Storage tiers — pay less for data you access rarely
Not all data is accessed equally. Azure offers access tiers to optimise cost:
| Tier | Access Frequency | Storage Cost | Access Cost | Use Case |
|---|---|---|---|---|
| Hot | Frequently accessed | Highest | Lowest | Active app data, images |
| Cool | Infrequently accessed (30+ days) | Lower | Higher | Short-term backups, reports |
| Cold | Rarely accessed (90+ days) | Even lower | Even higher | Compliance data, old logs |
| Archive | Almost never accessed (180+ days) | Lowest | Highest + rehydration delay | Long-term legal records |
Key exam concept: Hot tier = expensive to store, cheap to access. Archive tier = cheap to store, expensive and slow to access. Choose based on access patterns.
Archive tier rehydration
Data in the Archive tier is offline. You can’t read it directly. To access it, you must “rehydrate” — move it to Hot or Cool tier first. This takes hours (up to 15 hours for standard, or you can pay for priority at ~1 hour).
Exam tip: If a question mentions “immediate access required,” Archive is the wrong tier. Even Cold tier is better for data that might be needed occasionally.
Redundancy options — protecting your data
Azure automatically replicates your data. You choose HOW MANY copies and WHERE they’re stored:
| Feature | LRS | ZRS | GRS | GZRS |
|---|---|---|---|---|
| Full name | Locally Redundant | Zone-Redundant | Geo-Redundant | Geo-Zone-Redundant |
| Copies | 3 in one datacenter | 3 across 3 zones | 6 (3 local + 3 in paired region) | 6 (3 across zones + 3 in paired region) |
| Protects against | Server/drive failure | Datacenter failure | Regional disaster | Zone AND regional failure |
| Durability | 11 nines | 12 nines | 16 nines | 16 nines |
| Cost | Lowest | Higher | Higher | Highest |
Read-access variants: GRS and GZRS also have RA-GRS and RA-GZRS — these add read access to the secondary region copy. Without the RA prefix, the secondary copy is only for failover.
How to choose redundancy
| Scenario | Recommended Redundancy |
|---|---|
| Peak Roasters — small budget, basic protection | LRS (cheapest, protects against drive/server failure) |
| Summit Construction — needs zone protection | ZRS (protects against datacenter failure) |
| Harbour Health — regulatory requirement for data in two regions | GRS or GZRS (copies in paired region) |
Exam tip: Redundancy question patterns
Common exam patterns:
- “Protect against datacenter failure within a region” → ZRS
- “Protect against a regional disaster” → GRS or GZRS
- “Need read access to data in the secondary region” → RA-GRS or RA-GZRS
- “Lowest cost, basic protection” → LRS
Storage account types
| Account Type | Supported Services | Performance | Use Case |
|---|---|---|---|
| Standard general-purpose v2 | Blob, File, Queue, Table | Standard (HDD) | Most workloads |
| Premium block blobs | Blob only | Premium (SSD) | Low-latency apps |
| Premium file shares | File only | Premium (SSD) | Enterprise file shares |
| Premium page blobs | Page blob only | Premium (SSD) | VM disks |
Exam tip: Standard general-purpose v2 is the default recommendation for most scenarios. It supports all storage services, all tiers, and all redundancy options.
🎬 Video walkthrough
🎬 Video coming soon
Azure Storage Services — AZ-900
Azure Storage Services — AZ-900
~11 minFlashcards
Knowledge Check
Peak Roasters stores product images that are accessed hundreds of times daily. Which storage tier is MOST cost-effective?
Harbour Health needs to ensure their patient data survives a complete regional outage. Which redundancy option should they choose?
Next up: Data migration tools — AzCopy, Storage Explorer, Azure File Sync, Azure Migrate, and Azure Data Box.