Legacy backup solutions require that you periodically run full backups, that is, if your VM is 200 GB, a legacy solution will periodically back up all 200 GB over and over again, even if there only 20 MB of changes in the VM. Alternatively, legacy backup solutions can perform incremental backups but periodically require transforming increments into full backups. Such approaches are time consuming, resource intensive, and require an additional storage space. NAKIVO Backup & Replication offers a better approach. After an initial full backup, all jobs are forever incremental and transfer only changes to the backup repository. The data in the backup repository is stored using the full synthetic mode, which eliminates the need for backup transformation. As a result, NAKIVO Backup & Replication can back up and recover VMs faster, use less resources, save time, costs, and improve your recovery point objectives (RPO) and recovery time objectives (RTO).
After each job run, NAKIVO Backup & Replication creates a recovery point – a set of references to data blocks in the backup repository, which are required to reconstruct the entire VM as of a particular moment in time. Here is an example: You run the first backup of a VM on Sunday. For the sake of simplicity, let’s say that the VM consists of only 2 data blocks: A and B. Then on Monday, you run an incremental backup, which finds that the block A was deleted, but a new block C was added. Then on Tuesday, the incremental backup finds that the block B was deleted and a new block D was added. Here’s how the VM would look like during the three days:
And here’s how the data will be stored in the Backup Repository if the job is set to keep 3 or more recovery points:
As you can see from above, all unique data blocks in a backup repository are stored in a single pool, while recovery points are just references to the data blocks that are required to reconstruct a machine as of a particular moment in time. Since each recovery point "knows" what data blocks are needed for recovery, there is no need to run periodic full backups, apply increments one by one to get to a particular recovery point, or periodically transform backed up data blocks.
The full synthetic data storage approach provides a number of advantages: