Impact on Deduplication? - Local backup copy to Azure Blob

[dtwiley]

Trying to understand if there is a deduplication process at each stage of a backups movement.

If we backup say 20 VMs to a NTFS based repository, with the repo set to create a backup chain per VM, I'm assuming this then means the deduplication is unique to the chain and there is no saving across chains as there is no way to share between them?

When this backup is copied to Azure Blob Storage with a backup copy job, is the deduplication worked out again when its copied? ie. Will we then get further savings from deduplication because we're now putting ALL of these VMs into block storage or will it still deduplicate "per VM" or across all data stored?

[Mildur]

Hi Dave

If we backup say 20 VMs to a NTFS based repository, with the repo set to create a backup chain per VM, I'm assuming this then means the deduplication is unique to the chain and there is no saving across chains as there is no way to share between them?

Correct, not on NTFS. If you want to save storage, migrate your backups from NTFS to a ReFS volume (or a linux server with XFS) and leverage our FastClone feature. With FastClone, your synthetic full backups will be almost space less.

When this backup is copied to Azure Blob Storage with a backup copy job, is the deduplication worked out again when its copied? ie. Will we then get further savings from deduplication because we're now putting ALL of these VMs into block storage or will it still deduplicate "per VM" or across all data stored?

Did you mean object storage? You write Azure Blob and block storage.
For Backup Copy in v12 and later per-machine backup files are mandatory. This is also a requirement for object storage repositories.
Veeam's deduplication is "per backup chain".
With direct copy to object storage you write forever forward incremental backups. Only changed or new data block will be copied to Azure Blob. If active full is enabled, we will write the size of an entire full backup to Azure Blob.

Best,
Fabian