As its primary means of data reduction, Veeam uses a performance-optimized combination of deduplication and compression, allowing a single backup proxy to process data streams at 1GB/s during both vSphere or Hyper-V backups and restores — up top 10x faster than the competition!

Veeam performs deduplication at the job level at both the source (i.e., at backup proxy) and the target (i.e., at backup repository):

• Source-side deduplication ensures that only unique data blocks not already present in the previous restore point are transferred across the network
• Target-side deduplication checks the received blocks against other virtual machine (VM) blocks already stored in the backup file, thus providing global deduplication across all VMs included in the backup job.

You can choose from four different deduplication options based on backup size and connection speed to backup target ― so you can optimize both backup size and job performance.

Veeam also offers five levels of compression that allow you to find the right balance of storage consumption and backup proxy load, which, in turn, can affect performance. This includes a dedupe-friendly compression designed for deduplication storage devices.

Veeam’s built-in deduplication and compression work hand in hand to provide the best possible ratio of storage savings to backup performance. Unlike compression, inline deduplication is known to be very taxing on backup performance. Veeam ensures performance by using fairly large block sizes for deduplication, which allows for much better compression levels. The resulting data-reduction ratio is comparable to the ratio achieved by most advanced deduplication technologies — but with a much smaller processing overhead!

Deduplication appliance-friendly compression — the best of both worlds More ways to save storage and reduce network traffic.

Veeam Backup & Replication performs its deduplication and compression both at the source by backup proxy (i.e., before the data has traversed the network) and at the backup repository (i.e., before the data is stored on disk). This provides the benefit of bandwidth savings without impacting backup performance, as well as additional storage savings on the backup target.

Through its “dedupe-friendly” option, which uses a RLE (Run-Length Encoding) algorithm, Veeam Backup & Replication can send Veeam-deduplicated data to a backup appliance, and then have the backup appliance apply its own native deduplication to achieve even further data reduction. This allows you to have the best of both worlds — reducing bandwidth usage due to at-source data reduction, along with additional advanced deduplication by the storage.

In addition to well-known deduplication and compression techniques, Veeam introduces BitLooker, a new and advanced procession capability that further reduces vSphere or Hyper-V backup file size, facilitating faster data transfers and optimizing storage usage by up to 20%:

• Excluding swap and hibernation files: Unlike competing solutions, which require moving swap files to a separate virtual disk and then excluding them from backup, Veeam Backup & Replication automatically skips virtual disk blocks backing swap files from processing, saving you additional gigabytes of backup repository space, per vSphere or Hyper-V VM. Swap exclusion also improves backup performance, especially for incremental backups, which no longer need to transfer constantly changing swap file blocks.
• Excluding deleted file blocks: While performing file deletion, most file systems do not zero out blocks that belong to deleted files. These “deleted” blocks may result in image-level backups being significantly larger in size than the amount of disk space consumed from the guest OS perspective. BitLooker analyzes the NTFS Master File Table (MFT) to identify blocks belonging to deleted files, and automatically skips those blocks from image-level processing instead of saving those zeroed blocks into the backup image, thus reducing its size.
• Excluding user-specified files and folders: This feature lets users exclude individual files and folders during backup and replication, thus further reducing the amount of data that needs to be transferred and stored. Users can choose specific folders (with support for environmental variables), files and file masks for both inclusion and exclusion. For example, a user can exclude the DFS cache folder on a Windows file server VM to prevent constant cache changes from ballooning the incremental backup size.