Table of Contents
VMware vMotion and Storage vMotion are two of vSphere’s best-known live-migration features. They sound similar and are often confused, but they move different things: vMotion moves a running VM’s compute between hosts, while Storage vMotion moves a running VM’s disks between datastores. This guide explains how each works, their requirements and limits, and where they fit — including under Broadcom’s current licensing.
First, a clarification, since the term is often muddled: live migration means moving a running virtual machine — its compute or its storage — from one location to another with no downtime, so the services on it keep running throughout.
What is vMotion?
vMotion performs the live migration of a running VM from one ESXi host to another, with no downtime. The VM’s disk stays where it is (on shared storage reachable by both hosts); what moves is the VM’s live memory and CPU state. It’s used for:
- Hardware maintenance — evacuate a host before patching or servicing it.
- Load balancing — shift VMs off busy hosts (automated by DRS, Distributed Resource Scheduler).
- Retiring hardware — move workloads off aging or underperforming servers.
How vMotion works
vMotion does not clone the VM. It copies the VM’s active memory to the destination host over a dedicated vMotion network in an iterative pre-copy: memory pages are transferred while the VM keeps running, then the small set of pages that changed during the copy is synced. Once source and destination are in sync, execution switches to the destination host in a fraction of a second — transparent to the guest, which keeps the same identity, IP, and open connections.
Requirements to plan for:
- Compatible CPUs between source and destination hosts (Enhanced vMotion Compatibility, EVC, helps here).
- Shared storage reachable by both hosts (for classic vMotion, the disk doesn’t move).
- A dedicated vMotion VMkernel network, ideally 1 GbE or faster.
- Consistent networking — matching port-group/switch labels so connectivity is preserved.
The main challenges: vMotion needs a stable connection and sufficient resources on both hosts throughout the migration; if the link drops mid-migration, the move fails (the VM keeps running on the source). Large, busy VMs take longer and use more bandwidth.
What is Storage vMotion?
Storage vMotion performs the live migration of a VM’s disk files from one datastore to another while the VM keeps running — the VM stays on the same host, but its storage moves. It’s used for:
- Storage maintenance and array migration — move VMs off a datastore or array before upgrading or decommissioning it.
- Performance and tiering — relocate disks to faster or more appropriate storage.
- Reclaiming space and changing disk format (e.g. thin/thick) in the process.
How Storage vMotion works
Storage vMotion performs a single-pass copy of the virtual disk to the destination datastore. To keep source and destination consistent while the VM is still writing, it uses an I/O mirror: writes are mirrored to both the source and the new location during the copy. Once the copy completes and both are in sync, the VM is cut over to the disks on the destination datastore, and the source copy is removed.
The main limitation is time and bandwidth: because the entire disk contents move, migrating large disks can take a while and consume significant I/O — so it’s best scheduled thoughtfully in busy environments (Storage DRS can automate placement and balancing).
Enhanced and modern vMotion
Worth knowing, because the classic “shared storage required” rule no longer always applies:
- Enhanced (unified) vMotion combines vMotion and Storage vMotion in a single operation — moving both compute and storage at once, without requiring shared storage between hosts.
- Cross-vCenter and long-distance vMotion allow migrations between vCenter instances and across sites/networks.
- Encrypted vMotion protects the migration traffic in transit.
vMotion vs. Storage vMotion at a glance
| vMotion | Storage vMotion | |
|---|---|---|
| What moves | Running VM’s compute (memory + CPU state) | Running VM’s disk files |
| Between | Hosts (same cluster) | Datastores (same host) |
| Disk location | Stays on shared storage | Changes datastore |
| Classic requirement | Shared storage, vMotion network | Access to source + target datastore |
| Typical use | Maintenance, load balancing (DRS) | Storage maintenance, tiering, array migration |
| Speed | Fast (memory-sized transfer) | Slower (whole-disk transfer) |
| Automation | DRS | Storage DRS |
Enhanced vMotion combines both, removing the shared-storage requirement.
A note on licensing in 2026
Both are licensed vSphere features, and since Broadcom’s acquisition of VMware they come bundled within the subscription editions (such as vSphere Foundation and VMware Cloud Foundation) rather than sold à la carte. As many organizations reassess VMware costs, it’s worth noting that other platforms provide their own live migration — KVM live migration on OpenStack and Proxmox, for example — so these capabilities aren’t unique to vSphere if you’re evaluating a move.
How Storware fits
Live migration and backup are complementary, not interchangeable: vMotion and Storage vMotion keep workloads mobile and available; backup and recovery ensure you can recover them after failure, corruption, or ransomware. Storware Backup and Recovery covers the second role for VMware — and it adds a migration-adjacent capability of its own: during instant restore, you can enable live migration so the restored VM’s disks are automatically moved to your chosen storage after power-on. Storware can also restore VMs to a different location or, with cross-platform restore, a different hypervisor — useful if you’re moving workloads off VMware entirely.
Ready to protect your data?
Final thought
vMotion and Storage vMotion are complementary tools: vMotion moves a running VM’s compute between hosts, Storage vMotion moves its disks between datastores, and enhanced vMotion does both at once without shared storage. Together they reduce downtime, enable maintenance, and keep workloads balanced. Choose based on what you need to move — and, whichever you use, pair live migration with a solid backup strategy, because mobility isn’t recoverability.

