It depends on the OS.  I have seen that Veeam can determine how much space is used with Windows disks and will read the used space only. With Linix the whole diskspace is read. But in both cases the used space only is transmitted after compression.

 So, the backup will be less than 40GB, because of compression (and deduplication, if there are more than one VM in the job).

In the default config, the 100GB is what gets processed, but only the 40GB of needed data is backed up, it’s slightly more complicated than that but give this page a read to understand what can impact this:

https://helpcenter.veeam.com/docs/backup/vsphere/backup_job_advanced_storage_vm.html?ver=110
 

In essence though, unless you change the settings, Veeam will read the “entire” disk on the first backup, it will ignore deleted blocks, it will also ignore swap files & hibernation files. Of the remaining blocks it will deduplicate & compress so if you’re using only 40GB of your 100GB VM disk size, it wouldn’t be uncommon to still see a 2:1 reduction on your 40GB source data for a 20GB backup, but that all depends on the content!

The above is all true for the initial backup, then for your subsequent incrementals, unless settings are changed, Veeam will used Changed Block Tracking (CBT) to fetch only the changed blocks, then perform the above processing again to ensure only the necessary data is backed up for subsequent restores, these files will rely on the rest of the backup files in the “backup chain”.

The only thing I haven’t gone into detail on here is block sizes, but that’s a further optimisation story that doesn’t relate directly to your question. Read the storage optimisation section here for more detail on that if interested: https://helpcenter.veeam.com/docs/backup/vsphere/compression_deduplication.html?ver=110#optimization

I see @JMeixner beat me to it and described it better in fewer words

@JMeixner 

I believe, you are referring to a VBR feature called Bitlooker. It is available for VM Backups and works only for NTFS partitions. :)

With this feature, deleted blocks will be ignored and only used space will be analyzed.

Deleted File Blocks (BitLooker) - User Guide for VMware vSphere (veeam.com)

Maybe

I just observed this behaviour and could not name it. So, thank you for teaching me what feature this is.

Good comments here - thank you @MicoolPaul  @Mildur  @JMeixner  - one thing I’d also add is you can configure the compression & deduplication options in the backup job. You can go with the default settings; which are suitable for routine deployments however if you want you can go to “high” or “extreme” to trade CPU cycles for backup storage consumption reduction. It very much depends on the source data here, I’ve seen someone have encrypted videos get no compression (any setting) and some file servers or structured data types go very small after the backup with the extra compression.