2.1.3 Virtual Machine Replication
Replication of virtual machines to the target data center is key for the disaster recovery solution. With a disaster recovery solution, there are typically two types of replication available:
• Storage-based replication – This type of replication enables the storage layer to replicate the data directly to another storage subsystem located on the target site. This typically requires the same specification of storage hardware at each side of replication.
• Virtual machine replication – VMware vSphere® Replication™ enabled through the vSphere hypervisor allows the provider to offer storage subsystem-agnostic replication between environments. The vSphere environments on each side must be vSphere Replication compatible.
With this solution provided to end customers who might have different storage architectures than those offered by the cloud provider, the best option is to leverage a vSphere Replication based architecture. In this reference architecture, Rainpole Inc. will use vSphere Replication as the replication service between the on-premises site and the VMware Cloud Provider.
The key replication services are as follows:
• VM protection granularity – The end-user can self-service configure replication of any virtual machine within their source VMware vCenter Server®.
• RPO definition – The replication engine (vSphere Replication) allows the end customer to self-service the required RPOs between 15 minutes and 24 hours for their source virtual machines.
• Point-in-time recovery – PITR allows the end customer to take periodic snapshots of virtual machines which can be recovered. To use this option, you must manually revert the machine back to the snapshot after recovery.
Note RPO policy compliance is dependent on the bandwidth available between the end customer source data center and the VMware Cloud Provider Program DRaaS hosting provider.