The Volume Replicator (Volume Replicator) option offered by InfoScale Storage and InfoScale Enterprise provides a means to replicate virtual machine (VM) data.

While there are host-based technologies for replicating data across distances, they're usually expensive, requiring not only more storage, but also exactly the same hardware at both ends. They can also be limited in their ability to provide a solution that accounts not only for the data, but also for the applications that access it.

Volume Replicator runs at the host level, making it possible to replicate data volumes across distances to provide a means to extended disaster recovery without requiring that the hardware be exactly the same at both ends, and generally requiring less storage. But, while this works fine to protect application data in the guest, allowing recovery at the DR site, running in the guest does nothing to protect the VM.

By installing the required InfoScale product in the parent partition, volumes that contain VHD files used for VMs and/or application data can be selectively replicated to DR sites, either synchronously or asynchronously, over an IP network. Volume Replicator uses a replicator log to store all writes to the volumes grouped together in what is known as a replicated volume group in the correct order and replicate them to the DR site, maintaining write order fidelity, and thereby, consistency. The replicated VMs remain offline at the DR site until required to be brought online, either due to an intentional migration of services from the primary site, or due to an outage at the primary site, requiring the DR site to take over services.

For planned outages at the primary site, perhaps for a maintenance window, the primary role can be migrated to a DR site, allowing the VMs that have been replicated to be brought online and the applications that run on them to access the data that has also been replicated to the DR site.

For unplanned outages at the primary site, operations can be moved to a DR site by a takeover operation, which turns the DR site into an active primary, allowing VMs to be brought online. Depending on the mode of replication, they can either be completely up to date or behind the previous primary. In either event, consistency is maintained and applications are able to successfully attach to their data. The primary can be migrated back to the original site when it becomes available.