Storage efficiency XFS/ReFS vs. S3

Hi ​@MDK 

As per Anton Gostew wrote in this thread in forum, that was more inspiring for me, if you not use immutability in S3 bucket there aren't differences in space usage. When you set the Object lock there is an additional usage.

https://forums.veeam.com/object-storage-as-backup-target-f52/s3-object-storage-space-efficiency-t94222.html

So, what is the essential differences? XFS/ReFS uses reflink (Fast Clone in Veeam). S3 uses metadata+versioning these 2 functionality adds some space to use. 

There are a lot of good reasons to use S3 instead XFS/ReFS but the topic is very broad.

I hope I've replied to your question. 

I would in general expect almost the same consumption for both, if same compression settings are used. Why is that? Blocks of the backup become objects in S3 and all blocks of one chain (!) are only stored once (AKA deduped, or better → “prevented-from-being-duped”). In a XFS/ReFS you also get all changed blocks in a VIB and the next VBK re-uses all blocks that are already in the chain and only really stores the contents that would have been in this days VIB. 

Still, there are some differences:

• XFS/ReFS are DAS repos and therefore temporarily need a working space: 
• S3 on the other hand follows the concept of block generation leading to more storage being consumed for some time:
  Block Generation - Veeam Backup & Replication User Guide

Has to be taken into account, but does not change the big picture. Hope this helps.