Organizations who truly want to ensure a resilient and effective data protection solution want to be able to recover both from a disaster within a single site (single or potentially more than one virtual machine failure) as well as be able to withstand a total site failure (all resources in the site are down). There are many different aspects of data protection that can help to withstand any number of different scenarios as they relate to recovering from either virtual machine failure or failure of an entire site. Here is where we start thinking about the differences and benefits of both backups of virtual machines and replication of virtual machines. What role do each play in the overall data protection solution and how do we architect a solution to utilize the best of both worlds to achieve our data protection solution objectives?
What are Virtual Machine Backups and What Role do they play?
Virtual machine backups provide a compressed and deduplicated backup of a virtual machine running in virtual infrastructure. Generally speaking, most modern backup solutions today have the ability to make use of changed block tracking (CBT) in VMware vSphere environments and resilient change tracking (RCT) in Microsoft Hyper-V environments. This allows for a very efficient backup mechanism as after the first full backup, only the changes that have been made since the full backup will be copied with each backup iteration.
Typically, virtual machine backups are carried out at the host level and allow capturing the virtual machine data as well as the virtual machine hardware settings and configuration. The configuration settings are required to configure the virtual hardware on a restored virtual machine. So virtual machine backups allow restoring the entire virtual machine as well as specific files inside the guest operating system of the virtual machine. This allows for quickly recovering from a failure of an entire VM but also the ability to restore files within the VM that either were corrupted or accidentally deleted by an end user. Additionally, if damage occurred as a result of a ransomware infection, files could be restored either in total as a VM restore or by restored files that were affected by the ransomware infection.
Virtual machine backups that are application aware also give administrators the ability to perform backups that are consistent from a transaction standpoint if backing up databases, etc. This allows for no additional steps to be performed to bring a database back to a certain consistency in a point in time. With application-aware backups as well, certain housekeeping functions can be carried out such as truncating log files as regards to database driven applications such as Microsoft SQL Server and Microsoft Exchange Server. This keeps drive space from filling up and allows for an all in one solution for both backing up a database from an application standpoint as well as maintaining a healthy environment from a low-level disk perspective.
How Does Replication Jobs Fit into the Data Protection Plan?
Replication jobs give organizations the ability to create an exact copy or replica in a separate location of a virtual machine running in production. With a modern backup solution that makes full use of changed block tracking, replication can make use of the CBT process as well with each replication iteration applying the latest changed blocks only which allows for very efficient copies with each replication interval. Replication is not creating a backup copy in a backup repository in another location but rather creates a “restored” virtual machine in another VMware vSphere or Microsoft Hyper-V environment.
Replica virtual machines that are created in the secondary environment have the ability to keep restore points in the form of snapshots in VMware vSphere or checkpoints in the case of Microsoft Hyper-V. Once the configured limit of restore points is reached, the oldest snapshot or checkpoint is rolled off the virtual machine.
Typically, when we are thinking about designing a data protection plan that would allow withstanding an entire site failure, as mentioned, you would want to make sure the replica environment is located outside the production site. This would require a separate replicated environment running either VMware vSphere or Microsoft Hyper-V being targeted for replication. Additionally, site-to-site connectivity would also be required to replicate the virtual machines from production to the secondary DR facility.
With the above mentioned additional infrastructure costs, a secondary DR facility does represent a substantial capital investment in the facility, hardware required for the secondary environment, and the WAN circuit required to establish network connectivity. This results in many organizations trying to cut costs by simply not considering the value of replication as something that has much of a return on the investment.
However, organizations not creating replica virtual machines in a different location may find that if they experience a complete site failure, the recovery time involved with bringing all production resources back online from backup, can damage and bring untold loss to business reputation and customer confidence.
By having a secondary DR location, connectivity can simply be “failed over” to the secondary location if connectivity to the primary site is down due to a disaster recovery event. DNS records can quickly be changed and pending convergence of DNS propagation, connectivity will transition to the new site/network location of the DR environment.
Replication certainly bolsters the ability of an organization to withstand a major disaster where an entire site is down due to various events. Having the replica virtual machine already “restored” via the replication process drastically shortcuts the time it would take to first restore the virtual machines before failing services over to the new environment.
Utilizing both Backups and Replication in Data Protection
By using both backups and replication together, organizations can maintain very effective and efficient RTO an RPO objectives. Instead of competing with one another, backups and replication are complementing technologies in providing very resilient data protection solutions. Depending on RTO and RPO objectives, organizations can choose how replication is scheduled in with the backup schedule, as well as which virtual machines if not all need to be replicated to a secondary location.
There are definitely benefits to using both backups and replication in a data protection solution. Instead of being competing technologies or use cases, they instead complement one another. While backups are able to protect an organization and business-critical applications by providing the ability to restore single or even multiple VMs in a site, replication gives you the ability to withstand an entire site failure where restoring backups would simply not be feasible. Modern backup solutions such as Vembu BDR Suite allow organizations to effectively backup a VMware vSphere or Microsoft Hyper-V environment but also create replica virtual machines on a secondary location or DR site. Vembu BDR Suite is also able to effectively leverage public cloud for providing secondary locations for storing backup data, bolstering the overall data protection plan and agility that organizations can make use of for business-critical application resiliency.