There are many ways to store data. Just as you keep photos in an album, documents in a filing cabinet, and old items in a warehouse, Azure provides a variety of storage services depending on the type and purpose of your data. This guide explains all of Azure's storage services with analogies so that even first-time learners can understand them clearly.
What Is an Azure Storage Account?
An Azure Storage account is the top-level container you need before you can use any Azure storage service. Once you create a storage account, you can place various storage services inside it — Blob, Files, Queue, and Table. Think of it like a bank account: the account (Storage account) holds multiple products inside it — savings (Blob), term deposits (Files), foreign currency (Queue), and so on.
There are three key settings when creating a storage account. The first is the performance tier: Standard is HDD-based and lower cost, while Premium is SSD-based and high performance. The second is the redundancy option — you choose from LRS, ZRS, GRS, RA-GRS, or GZRS. The third is the account type, with General Purpose v2 being the most widely recommended.
Azure Blob Storage — The All-Purpose Warehouse
Blob (Binary Large Object) Storage is Azure's most widely used storage service. It is an all-purpose warehouse where you can store images, videos, documents, backup files, log files, and virtually any other type of data. There is almost no file size limit, and it can store billions of objects.
Blob has three types, each optimized for a different use case, so you choose based on your situation.
Block Blob — The Standard Box
The most common type. Stores images, videos, documents, backups, and similar files. The file is divided into blocks for upload, so if an upload fails partway through, only the failed block needs to be re-uploaded. Most general-purpose file storage needs use Block Blob.
Page Blob — The Random-Access Drawer
Used when random (non-sequential) read and write operations are needed. Azure VM virtual hard disk (VHD) files are stored as Page Blobs. The ability to quickly read and write at specific positions makes it well suited for disk use. Supports up to 8 TB.
Append Blob — The Append-Only Notebook
Data can only be added to the end — modifying existing content is not allowed. This makes it ideal for log files and audit data records. Once written, content cannot be modified or deleted; only appended. It is ideal for data where integrity matters most, like security audit logs and streaming log records.
| Blob Type | Characteristics | Use Cases | |----------|----------------|----------| | Block Blob | Block-based upload | Images, videos, documents, backups | | Page Blob | Random read/write | VM virtual disks (VHD) | | Append Blob | Append only | Log files, audit data |
Storage Access Tiers — Temperature Control for Your Warehouse
Blob Storage offers four access tiers based on how frequently you need to access your data. Just as you keep frequently used items within easy reach and rarely used items deep in storage to save costs, the access tier you choose affects both storage cost and access cost. The core principle is that storage cost and access cost are inversely proportional — the lower the storage cost, the higher the retrieval cost.
Hot Tier — The Refrigerator (Frequently Accessed)
Best for data you access daily or frequently. Storage cost is highest, but access cost is lowest. Examples: images on an active website, log files created today, data from actively used applications.
Cool Tier — The Warehouse Shelf (Occasionally Accessed)
Best for data you do not expect to access for at least 30 days. Storage cost is lower than Hot, but there is an access cost when you retrieve data. Examples: backup files older than one month, quarterly reports, monitoring data you check occasionally.
Cold Tier — The Underground Storage (Rarely Accessed)
Best for data that will not be accessed for at least 90 days. Storage cost is lower than Cool. A relatively recent addition, it serves as a middle option between Cool and Archive.
Archive Tier — The Long-Term Vault
The cheapest storage tier for data that will almost never be accessed for 180 days or more. Retrieving data (rehydration) can take up to several hours. Used for medical records, legal documents, and old media files that must be retained for compliance reasons. Setting rehydration priority to High (within one hour) speeds up retrieval at an additional cost.
| Tier | Storage Cost | Access Cost | Minimum Retention | Best For | |------|-------------|------------|------------------|---------| | Hot | High | Low | None | Frequently accessed data | | Cool | Medium | Medium | 30 days | Occasionally accessed data | | Cold | Low | High | 90 days | Rarely accessed data | | Archive | Very low | Very high (rehydration) | 180 days | Long-term retention, almost never accessed |
!Blob storage access tiers
Azure Files — Cloud Shared Drive
Azure Files is a fully managed file sharing service that you mount using SMB (Windows shared folder protocol) or NFS protocol. It is like moving the shared network drive (for example, the Z: drive) from a company server to the cloud. Multiple VMs or computers can access the same files simultaneously.
Why use Azure Files? Use it to migrate on-premises file servers to the cloud, when multiple VMs need to share common configuration files or data, or when a legacy application exchanges data using the file sharing method.
Supports SMB 2.1, SMB 3.0, and NFS 4.1 protocols Mountable on Windows, Linux, and macOS Azure File Sync: automatically synchronizes between on-premises file servers and Azure Files, using the on-premises server as a cache while centrally storing data in Azure Available in both Standard (HDD) and Premium (SSD) performance tiers
Azure Queue Storage — The Numbered Ticket Queue
Queue Storage is a messaging queue service that stores and processes messages in order. Think of it like the numbered ticket system at a bank or restaurant. A customer (application A) takes a number (inserts a message) and a teller (application B) processes them in order. Two applications can exchange messages without being directly connected, which reduces coupling between services and smooths out processing speed differences.
A practical example: imagine a photo upload website. When a user uploads a photo (message inserted), a resizing and thumbnail generation service (message processor) retrieves the photo from the queue and processes it. Even if uploads arrive all at once, the queue acts as a buffer so the processing service does not get overloaded.
Stores messages up to 64 KB in size Retains messages for up to 7 days Ideal for asynchronous task processing and communication between microservices Visibility timeout can be configured to prevent a message from being visible again before processing completes
Azure Table Storage — Simple NoSQL Table
Table Storage is a simple key-value NoSQL data store. Like an Excel spreadsheet, it is organized into rows and columns, but the schema is not fixed — each row can have different columns. It is much cheaper than a relational database and well suited for storing large amounts of unstructured data. For complex queries or relational data, Cosmos DB is recommended.
Every item is uniquely identified by the combination of a Partition Key and a Row Key. Items sharing the same Partition Key are stored on the same physical server, enabling fast lookups.
Azure Managed Disks — The VM's Hard Drive
Managed Disks are virtual hard disks that attach to Azure VMs. A VM comes with one OS disk by default, but you can attach additional Managed Disks for extra data storage. "Managed" means Azure automatically handles the physical management of the disk — placing it in a storage account, replicating it, distributing it across availability zones, and so on. Previously you had to create and link storage accounts yourself; now Azure handles all of that.
Performance type characteristics: Ultra Disk: Extreme IOPS and throughput, for high-performance databases like SAP HANA and large SQL instances Premium SSD v2: High performance, low latency, with individually adjustable IOPS and throughput Premium SSD: SSD-based, the standard for high-performance production workloads Standard SSD: Balanced performance and cost, suitable for web servers and development environments Standard HDD: Cheapest, for backups and non-critical data
Storage Redundancy Options — Managing Data Copies
Redundancy is the strategy of copying data to multiple locations so you do not lose it if a failure occurs. Like making multiple copies of an important document and storing them in different places, higher redundancy means broader protection — and higher cost.
LRS (Locally Redundant Storage) — Three Copies in the Same Building
Stores three copies of data within the same data center (single physical location). The cheapest option, but if the entire data center fails (fire, flood), data could be lost. Use for reproducible data or cost-first environments.
ZRS (Zone-Redundant Storage) — Three Different Buildings in the Same City
Copies data to three different availability zones within the same Azure region. Even if an entire data center (building) goes down due to fire or power failure, data remains accessible from another availability zone. Recommended for production environments where high availability is important.
GRS (Geo-Redundant Storage) — A Backup Warehouse in Another City
Stores three copies with LRS in the primary region, then replicates three more copies to a secondary region hundreds of kilometers away — six copies total. Protects data even during region-wide disasters like earthquakes or widespread floods. The secondary region is not accessible until Azure initiates a failover.
RA-GRS (Read-Access Geo-Redundant Storage) — Readable from the Backup Warehouse at Any Time
Same geographic repl