Live Migration is the Hyper-V feature that moves a running virtual machine from one host to another without the user noticing an interruption. The VM keeps responding while it moves: its memory, its state and its disk become available on the destination host, and the switch is fast enough that network connections stay up.
Microsoft’s documentation puts flexibility first. A running VM stops being tied to one physical machine. You can drain a host of all its virtual machines before pulling it from the rack or applying updates that need a reboot. The service running inside the VM never knows the move happened.
What it’s for and who needs it
If you run more than one Hyper-V host, Live Migration is what saves you from shutting down services every time you touch the hardware. Firmware maintenance, RAM swaps, replacing an aging server: in all of these you move the workloads to another host, do the work, and bring them back.
Paired with Windows Failover Clustering, Live Migration is the foundation for highly available and fault-tolerant systems. The cluster watches the hosts and, together with live migration, redistributes or relocates VMs as needed. Microsoft also pairs it with System Center Virtual Machine Manager for larger estates where the full VM lifecycle is managed.
What changed in Windows Server 2016
The documented change is that, starting with Windows Server 2016, there are fewer restrictions on deploying Live Migration. The feature now works without Failover Clustering. This is what’s known as shared nothing live migration: you can move a VM between two independent hosts that share no storage and belong to no common cluster.
The rest of the behavior stays the same as in previous releases. Microsoft keeps historical documentation for Windows Server 2008 R2 and Windows Server 2012 R2, where Live Migration already existed but with more constraints. If you work with those older versions, the configuration steps differ.
To run migration without a cluster, Microsoft documents two steps: set up the hosts for Live Migration without Failover Clustering, and then run the move itself. This is the usual route in labs and small fleets where a full cluster isn’t worth the effort.
How it fits into a virtualization setup
Live Migration covers workload mobility within the Hyper-V world, the same way other platforms solve the problem with their own mechanisms. If you come from Xen, the concept is close to the live migrate that XCP-ng offers in its 8.3 LTS branch, and the underlying idea is the same: decouple the workload from the iron running it.
In a real deployment, Live Migration rarely stands alone. It leans on shared storage when there’s a cluster, on Failover Clustering for high availability, and on management tools like System Center as the host count grows. What it brings is continuity: moving a VM should not mean stopping it.