Skip to main content
Virtual machine storage is a core concept when deploying resilient and performant Azure resources. This article explains the VM disk types in Azure, performance considerations (IOPS/throughput/latency), and management options so you can choose the right storage for your workload.

​
Azure VM disk types

There are three primary disk types you should know when architecting or operating Azure VMs:
  • OS disk
    The OS disk contains the operating system and boots the VM. It is a VHD stored as a page blob when using unmanaged disks or as a managed disk (an Azure-managed resource backed by platform storage). OS disks are persistent and durable and should contain only the OS and system-level configuration.
  • Temporary disk
    The temporary disk is directly attached to the VM host and is intended for ephemeral data: swap/page files, temporary caches, or transient scratch space. Data on the temporary disk can be lost during host maintenance, reboots, or VM deallocation — do not use it for persistent data. Note that not all VM sizes include a temporary disk; check the VM size documentation before relying on it.
Do not store important or persistent data on the temporary disk. Use data disks, managed disks, or external storage services (like Azure Files or Azure Blob Storage) for persistent storage and backups.
  • Data disks
    Data disks are additional persistent disks you attach to a VM for application and data storage. The number of attachable data disks and the aggregate IOPS/throughput available depend on the VM size and series. Always verify disk counts and performance limits for your chosen VM size.
A slide titled "Virtual Machine Storage" showing a diagram of a virtual machine with OS, temporary, and data disks connected to Azure Blob Storage. Two side panels list "Performance Tiers" (disk tier, HDD/SSD, IOPS/cost) and "Management" (choose disk type, managed vs unmanaged).

​
Performance tiers and choosing the right disk

Choose a disk tier based on workload I/O profile, latency sensitivity, and cost. Consider these general recommendations: When evaluating disk tiers, weigh IOPS, throughput (MB/s), latency, and cost. Also account for VM-level maximums: many VM SKUs limit aggregate IOPS/throughput across attached disks, so pick a VM size that supports your storage performance requirements.

​
Managed disks vs unmanaged disks

Two management approaches exist for VM disks in Azure:
  • Managed disks
    Managed disks are an Azure-managed abstraction over the underlying storage. Azure handles provisioning, placement, scaling, and the storage account boundaries for you. Managed disks simplify operations and integrate with features such as snapshots, Azure Backup, Azure Site Recovery, and availability constructs (availability sets and availability zones). They support Standard HDD, Standard SSD, Premium SSD, and Ultra Disk tiers. Use managed disks for most deployments to avoid storage account IOPS limits and reduce operational overhead.
Managed disks simplify operations and are the recommended choice for most deployments. They support snapshots, backup, and native integration with availability sets and zones.
  • Unmanaged disks
    Unmanaged disks are VHD files stored as page blobs in a storage account you create and manage. This gives you low-level control over storage accounts, but you must manage scalability, distribution, and avoid exceeding storage account IOPS/throughput limits. Unmanaged disks are generally only used when you have a specific requirement to manage the storage accounts directly.

​
Quick comparison

​
Practical CLI examples

Create a managed disk and attach it to an existing VM using the Azure CLI (example): Create a managed disk:
Attach the managed disk to a VM:
Check VM disk limits and select an appropriate VM size by consulting the VM sizes documentation for your region and series:

​
Additional considerations

  • Snapshots and backups: Managed disks support on-demand snapshots and integrate with Azure Backup; use snapshots for point-in-time backups and rapid cloning.
  • Availability: For high availability, combine VM placement strategies (availability sets or availability zones) with managed disks to distribute disks and VMs across fault and update domains.
  • Disk limits: The maximum number of attachable data disks and total IOPS/throughput depend on VM size. Always consult the VM size documentation before provisioning production workloads.
  • Data protection: Use backups, replication (for example Azure Site Recovery), or external storage for critical data protection strategies.

​
Next steps

Now that you understand VM storage types, performance tiers, and management options, follow the next article to learn how to create and configure VMs in the Azure portal. That guide will cover OS disk sizing, attaching and formatting data disks, disk tier selection, and setting availability options.

Watch Video