Manage storage and file services
File systems and deduplication
NTFS remains the general-purpose choice; ReFS adds integrity streams and block cloning but cannot host the boot volume and does not support Data Deduplication. Data Deduplication runs only on NTFS volumes, works per volume, honors a minimum file age and exclusion list, and offers optimization types tuned for general file shares, VDI, or backup workloads. Deduplication is a background job, not an inline feature of the file system.
Access and governance
Share permissions and NTFS permissions combine, and the most restrictive effective result wins. The practical pattern is broad share permissions with precise NTFS control. Access-based enumeration hides folders a user cannot read. File Server Resource Manager supplies hard and soft quotas, auto-apply quota templates, file screening, storage reports, and classification.
Namespaces and replication
DFS Namespaces present one path to many targets and choose referrals by target priority and site cost; clients can fail back. DFS Replication moves data between members using remote differential compression, a staging folder, and conflict-and-deleted folders for losers of a conflict. Namespace servers do not require DFS Replication.
Hybrid tiering
Azure File Sync uses a Storage Sync Service, a sync group, one cloud endpoint (an Azure file share), and server endpoints. Cloud tiering keeps hot files local and replaces cold files with reparse points, so recalling tiered data requires connectivity. A server endpoint cannot be the system volume.
Traps
ReFS plus deduplication is invalid. DFS Replication is for file data, not for presenting a namespace. BitLocker, quotas, and deduplication settings are per volume and sometimes per folder.
Rule of thumb
If the requirement is one friendly path, use DFS Namespaces; if it is keeping copies in sync, use DFS Replication; if it is capacity relief, use Azure File Sync cloud tiering.