The story
The gap
As licensing costs on legacy virtualization platforms climbed, enterprises started moving workloads to KVM, Proxmox and oVirt. The hypervisor migration was the easy part.
Disaster recovery did not come with them. Teams that had been running seconds of recovery point objective arrived on KVM and found themselves back on nightly snapshots. Years of progress, undone by a platform change.
Why nobody had built it
Not because the market was too small. Because the architecture does not port.
Continuous data protection on legacy platforms depends on hypervisor-level hooks that simply do not exist in the KVM world. The established vendors are snapshot-based by design, so adding true CDP is not a feature they can ship in a release. It is a rewrite. That is why their KVM support stops at backup.
What we built
KVMDR is native to KVM rather than adapted to it. Replication is driven by QEMU dirty bitmaps and block mirroring, tracking writes as they happen at the block layer.
Nothing is installed inside the guest. The full lifecycle is covered, including failover, failback, point-in-time recovery and re-IP on the target network, so a test or a real invocation follows the same path.
Where it is now
KVMDR reached general availability in July 2026, with the current version at v1.8 supporting Proxmox, oVirt, OLVM and RHV.