Network Attached Storage (NAS) Sizing: Mixed Drives, SHR Parity, and Filesystem Overheads
Purchasing and configuring a Network Attached Storage (NAS) appliance—whether from Synology, QNAP, Asustor, TrueNAS, or a custom unRAID server—represents a major investment in data consolidation and backup resilience. Yet, many first-time NAS owners experience immediate confusion upon completing volume initialization: four 8TB hard drives (32TB raw) frequently yield only approximately 21.4TB of actual usable storage for shared folders. Understanding the layered architecture of modern NAS operating systems clarifies where that storage goes and prevents costly capacity shortages.
Why Usable NAS Storage is Less Than Raw Drive Capacity
The gap between the physical terabytes written on drive retail packaging and the actual space available in your SMB/NFS network folders is caused by four successive overhead reductions:
1. Binary Conversion: 1 TB Decimal = 0.909 TiB Binary2. RAID Parity: 1 or 2 Drives Reserved for Redundancy3. NAS OS Partition: ~3.5 GB per Drive (System Root & Swap Mirror)4. Filesystem & Snapshots: 2% to 4% Metadata + Configured Snapshot Buffer
First, drive manufacturers measure capacity in decimal base-10 units (1 TB = 1,000,000,000,000 bytes), whereas modern storage operating systems calculate capacity in binary base-2 gibibytes and tebibytes (1 TiB = 1,099,511,627,776 bytes). This conversion alone immediately reduces perceived capacity by approximately 9.1%. Second, parity arrays require reserving the capacity equivalent of one or two full drives for fault tolerance.
Third, NAS operating systems (like Synology DSM or TrueNAS) format a hidden RAID 1 mirror across the first few gigabytes of every connected drive to house the Linux operating system kernel, service daemons, and swap partitions. This guarantees that the NAS can boot even if multiple storage drives fail. Finally, modern copy-on-write filesystems such as Btrfs and OpenZFS duplicate metadata trees and reserve slop space to ensure transactional consistency and atomic snapshot creation.
The Magic of Synology Hybrid RAID (SHR) and Mixed-Drive Slicing
In traditional RAID 5 or RAID 6, adding disks of mismatched capacities restricts every drive to the capacity of the smallest disk. For instance, if you combine two 8TB drives and two 16TB drives in traditional RAID 5, each drive is treated as an 8TB disk, leaving 16TB of capacity on the larger drives completely unusable.
Proprietary volume managers—most notably Synology Hybrid RAID (SHR) and unRAID—overcome this restriction by dynamically slicing drives into horizontal sub-arrays. SHR creates an initial RAID 5 array across the 8TB baseline of all four drives, providing 24TB of usable space. It then creates a secondary RAID 1 mirror across the remaining 8TB on the two larger drives, providing an additional 8TB of usable storage. As a result, SHR delivers 32TB of usable capacity where traditional RAID would deliver only 24TB, while still tolerating any single drive failure.