Who needs replication anyway?

vRanger Pro offers the flexibility of both backup and replication in one product, but sometimes it’s unclear how vRanger’s replication works and when it should be used. In this post, I’ll explain how vRanger replication can fit in with your data protection strategy.

Software-based replication creates a replica, or a copy, of your source data at specified intervals without relying on specialized hardware such as storage controllers. This is often the right choice for meeting all but the strictest service level agreements because it offers the right amount of protection by providing a current replica of data at a cost-effective price. VMware also offers software-based replication, but only for vSphere 5, and only with VMware’s Site Recovery Manager product that comes at an extra cost. It is not snapshot based, and uses a filter level driver, which has different tradeoffs. These are important factors to consider when choosing your replication solution.

Just as with backups, vRanger’s software-based replication operates at the VM image-level, unlike other kinds of software-based replication which operate at the file or application level. For bandwidth efficiency, only changes since the last replication are sent over the wire. vRanger’s ability to efficiently replicate your VMs to dissimilar storage and server hardware enables you to fully leverage your VMware environment for disaster recovery by allowing you to cost-optimize your disaster recovery infrastructure. For example, you could replicate critical VMs to older infrastructure or to infrastructure that would normally be used for testing, but that can also be leveraged as temporary infrastructure in a disaster scenario.

Kicking off replication in vRanger is easy - to select a VM for replication, simply right click the VM from your inventory in vRanger and select “Replicate…” as shown in the screenshot below:

As you go through the Replication Wizard, you can select a recurrence schedule that meets your service level agreements as shown:

But backups use less storage through vRanger’s space saving technologies such as incremental backups, Active Block Mapping (ABM), and compression (and more to come soon…stay tuned for my next post), so why use replication for data protection instead of backups? It all comes down to your Recovery Time Objective (RTO), or the time you can afford to wait before you can actually use your VM.

To use your replica VM in a disaster recovery situation, you can simply click on the “Failover” button in vRanger. The following steps occur during a failover:

  • The production VM will be powered off
  • You can optionally perform a final synchronization between sites to ensure that no data is lost during a failover.
  • The replica VM will be powered on
  • The replication job will be reversed, with the replica VM now acting as the source. This is done so you can “Failback” when you want to use your original production environment again.

Failover is completely automated so you can be up and running with your replica VM quickly because a replica of your VM is already standing by, just waiting to be powered on – contrast this with what would happen if you needed to restore your VM from a backup in a disaster situation. You would have to select a repository, pick a savepoint to restore, create a restoration job with a name and description, choose a viable destination (host, datastores etc.), choose the restore network, and finally choose your restore method (LAN-free or network). And all this manual intervention has to happen in addition to the time it takes for vRanger to actually perform the restore, which would take even longer for incremental backups as the full image would have to be re-composited from other backups.

Replication is ideal for many disaster recovery scenarios, but backups should be an important part of your data protection strategy because they allow multiple savepoints – with replication, if you find something wrong with the replica VM once you power it on, you are out of luck. vRanger does provide a test facility to mitigate this issue, but with backups, you can go back to a previous savepoint to provide protection against a wider range of problems – for example, you can restore a VM from a savepoint before a virus infection. A comprehensive data protection strategy requires both backup and replication, and that’s what you get with vRanger Pro.

Just one more thing - in addition to the data protection use cases for vRanger’s replication, you can also use replication to migrate VMs across datacenters when vMotion will not work. So if you haven’t looked at vRanger’s replication yet, give it a try, because the person who needs replication just might be you.