When protecting virtual machines using Hyper-V replica, the first question that comes to mind is whether you should protect the replica VMs or the primary VMs. From what we have seen in the real world, three scenarios are found to be useful when protecting replica VMs:

• Reduce the impact of backup on the production workloads: In today's world, workloads need to run 24 hours a day, 7 days a week with high performance. If the workload will run on differencing virtual disks for a long period of time, this will impact the system during backup operations.
• Limited network bandwidth between the branch office and primary office: Network bandwidth is expensive, and it is redundant to send data to the head office twice, one for disaster recovery and a second for backup. A more efficient way is to support backup from a replica DR site so that you can manage the backup infrastructure from a single location.
• Enterprise to hoster scenario: Many customers do not want to build a second datacenter for disaster recovery and prefer to leverage a hosting service provider. Hosting service providers offer SLAs as a way around backing up VMs on a regular basis (for example, daily), and this can be easily achieved using Hyper-V replica VM backups.

There is some limitation with backing up replica VMs: you can only get crash-consistent backups. However, most customers are comfortable with this. Remember, crash-consistency doesn't mean inconsistency. It's equivalent to the state when the power plug is pulled. Applications know how to recover from this state.