Choose the right VM replication solution for you

by Patrick Redknap on February 25, 2014

One of the first decisions that must be made as part of DR planning is determining how critical data will be replicated from site to site. The below examples describe the methods available when using the PHD Virtual product offerings and advantages of each for VMware environments.

Host based replication

How it works:
For host-based replication, the ReliableDR replication engine is embedded within the ReliableDR server and managed from within the ReliableDR web interface. All of the replication tasks are performed by making remote calls to the production environment over the network. Leveraging the vSphere VIX APIs and Changed Block Tracking, the ReliableDR server accesses the production vCenter or ESXi hosts to snapshot and pull the changed data over the network to the DR site. The data is then written directly to the VM(s) virtual disks and then tested using automated recovery verification by actually powering up all included virtual machines in a sandbox. If testing is successful, a Certified Recovery Point will be created by taking a vSphere snapshot of the certified VM’s. This method is ideal for environments needing to protect smaller datasets or also have higher RPO’s.

Host-based replication with ReliableDR
In general, offsite backups are viewed as a protection policy. However, it is essential that you consider how you intend to use those backups for recovery purposes.


Advantages of host-based replication

  • Simplicity – It is completely integrated with ReliableDR and is very easy to use. You can quickly set up a replication job and then select how you wish to verify the applications replicated.
  • Storage Agnostic – Because it is software based, host-based replication allows users to leverage lower-cost, heterogeneous storage for the VMware environment at the secondary site.
  • Zero-Production footprint – The ReliableDR server will live at the secondary site, requiring no installation components at the production site.
  • Efficient, planned failover/failback – Users can execute planned failovers that properly shut down production systems and replicate the latest data to a secondary DR site. Thus, when failback occurs, the production site can be quickly synchronized with the latest data that changed at the DR site during failover.

Storage-based replication

How it works:
For storage-based replication, ReliableDR integrates with a number of storage vendors that also support data replication. As data is replicated by the storage subsystem – i.e. the LUN/Volume level – ReliableDR residing at the DR site will automate registering those LUNs or Volumes as a datastore in VMware and bring those VM’s online for testing and failover operations. Each successful test is captured as a snapshot at the storage level to keep the desired number of rollback points.
ReliableDR does not manage or initiate the replication process itself. That is conducted through the storage platform separately. For certain storage vendors, namely NetApp, Datacore and Nimble, ReliableDR can also reverse replication for failback. Additional storage vendors will be supported in 2014 since ReliableDR can easily be integrated with new storage APIs.


Advantages of storage-based replication

  • Scalability – Storage replication allows for fast, efficient data transfer for very large virtual environments.
  • Efficiency – Storage snapshots are typically less impactful to production resources than hypervisor snapshots. Therefore, critical applications that are highly transactional can often be protected with less impact using storage replication. Additionally, many storage models provide advanced WAN optimization for increased efficiency of data transfer.
  • Shorter Recovery Point Objectives (RPOs) – Due to its efficiency and scalability, users typically realize shorter RPOs with storage replication solutions.

CertifiedReplica – PHD Virtual Backup and Replication with ReliableDR

How it works:

PHD Virtual Backup and Replication performs VM replication in a unique way compared to host-based or storage-based options. A Virtual Backup Appliance, or VBA, is used at the production location to take backups to local storage using hypervisor snapshots. The backup data is then used to perform VM replication by deploying at least one VBA at the DR site. The VBA at the DR site will mount the backup store at the primary site and pull the data across the WAN, rehydrating the data, and leaving you with a VM existing in a powered off state. ReliableDR can then be used to automate recovery testing and failover for those VBA replicated VM’s.
This functionality is known as “CertifiedReplica”, but ReliableDR also can utilize its integration with PHD Virtual Backup to schedule and initiate the replication process instead of scheduling replication at the VBA level for simplified management.


Advantages of VBA Replication with ReliableDR

  • Flexibility – All replication is done per VM, giving users complete control over what they want to replicate.
  • Storage Agnostic – Because it is software based, host-based replication allows users to leverage lower-cost, heterogeneous storage for the VMware environment at the DR site.
  • Scalability – Multiple VBAs can be configured in multiple sites to scale as needed.
  • WAN efficiency – Source-side compression moves less data over the WAN.
  • Security – Source-side encryption provides built-in security for data in transit to the DR site.
  • Less production storage impact – There is typically more production impact using hypervisor snapshots compared to storage snapshots. However, the use of a single snapshot for both backup and replication reduces that impact. Additionally, splitting the replication process into a backup and a replication task reduces the amount of time that a snapshot is left open on the production storage for replication – because data is replicated from backup storage rather than production storage.


Don’t just take my word that ReliableDR has these replication advantages! Try it out for yourself now.

Go to and find out how!

Leave a Comment

Previous post:

Next post: