Lesson 1: Every Azure Storage Service on One Page

Why start with a map

When a team says "we need storage on Azure", the first mistake usually isn't a wrong setting. It's choosing from the wrong group. Someone asks for a file share and gets a blob container. Someone needs low-latency block storage for a database and is handed an NFS share. Before we compare tiers, limits or prices in later levels, we need one page that shows every service, what kind of storage it is, and what it's called now.

Think of it like the floor plan of a hardware store. Before you compare drills, you need to know which aisle has drills.

The one-page map

Gotcha: "archive" is not a product. On Azure, archive storage is a tier on Blob Storage. If a requirement says "cheap long-term archive", you're still in the object aisle. You just pick a different tier.

Inside the storage account vs separate resources

This split matters in practice. Anything inside a storage account shares that account's redundancy choice, network rules, encryption settings and endpoint name. Separate resources have their own resource provider, their own networking model (often VNet-only) and their own billing unit.

A storage account is a building with five doors, one per data service:

https://<account>.blob.core.windows.net    # Blob Storage
https://<account>.dfs.core.windows.net     # Data Lake Storage endpoint (hierarchical namespace)
https://<account>.file.core.windows.net    # Azure Files
https://<account>.queue.core.windows.net   # Queue Storage
https://<account>.table.core.windows.net   # Table Storage

You can see this split yourself. Run the following in Cloud Shell:

# 1. A resource group to experiment in
az group create -n rg-storage-map -l uksouth

# 2. One storage account with hierarchical namespace switched on.
#    That flag is all "Data Lake Storage" is: no separate account type.
az storage account create \
  -n stmapdemo$RANDOM -g rg-storage-map -l uksouth \
  --sku Standard_LRS --kind StorageV2 \
  --enable-hierarchical-namespace true

# 3. List the resource types in the group
az resource list -g rg-storage-map --query "[].{name:name,type:type}" -o table

Expected output:

Name              Type
----------------  ---------------------------------
stmapdemo21873    Microsoft.Storage/storageAccounts

There's one resource and no "Data Lake" resource type, because Data Lake Storage is a set of capabilities on Blob Storage. If you created a managed disk or a NetApp account here, each would appear as its own type (Microsoft.Compute/disks, Microsoft.NetApp/netAppAccounts).

Trade-off to remember: you can't switch hierarchical namespace off again later. Treat it as a design decision, not a toggle. We cover the details in the Data Lake lesson.

# Clean up so you pay nothing
az group delete -n rg-storage-map --yes --no-wait

Names change, so read the banner

Microsoft renames and retires services, and old names live on in blog posts, Terraform modules and job specs. Some you'll still see:

• Data Lake Storage Gen2 is now just Azure Data Lake Storage. "Gen2" still shows up in API and SDK names, and even Storage Mover's docs say it uses the "ADLS Gen2 REST API".
• Data Lake Storage Gen1 is gone. It was retired on 29 February 2024.
• Data Box Heavy (about 770 TB usable) is retired and can't be ordered. The current devices are Data Box 120 and Data Box 525, which became generally available in March 2025. The older 80 TB Data Box is being retired in each region as the new devices arrive there.
• Azure HPC Cache retired on 30 September 2025. For HPC scratch and AI training data, look at Azure Managed Lustre instead.

A quick routine for checking any service on the map:

1. Go to its overview page on learn.microsoft.com (for example learn.microsoft.com/azure/databox/data-box-heavy-overview).
2. Read the Note or Important banner at the top. Retirements are announced there first, as on the Data Box Heavy page.
3. Check the page's last-updated date, then search Azure Updates for "retirement" plus the service name.
4. Write the service name and the date you checked into your design notes.

Best practice: the Learn overview "Introduction to Azure Storage" lists the core data services and the specialised ones (NetApp Files, Managed Lustre) on one page. Bookmark it. It's the closest thing to an official version of this map.

Key takeaways

• Map the need to a group first: object, block, file, high-performance file, archive, transfer or backup. Pick the service second.
• Blob, Data Lake, Files, Queue and Table live inside a storage account. Managed Disks, Elastic SAN, NetApp Files and Managed Lustre are separate resources with their own providers.
• Data Lake Storage is Blob Storage with hierarchical namespace switched on. It was formerly called Gen2. Gen1 retired in February 2024.
• Archive is a Blob tier, not a separate service.
• Data Box Heavy and HPC Cache are retired. Order Data Box 120/525, and consider Managed Lustre for HPC file workloads.
• Always check the banner on the service's Learn page before you trust a name or an option.