Skip to main content
StashRaid

Storage Parameters

Usable Capacity

12 TB

Fault Tolerance

4

Array Resilience

Architecture

RAID 5 • Parity

Fault Tolerance

1 Concurrent Failures Allowed

Current Configuration

RAID 5 (Parity)

12 TB Usable

Total RAW: 16 TB

Disk 01DATA
Disk 02DATA
Disk 03DATA
Disk 04PARITY

Total Drives

4 x 4 TB

Efficiency

75% Utilized

Usable: 12 TBWasted to redundancy: 4 TB

Pros & Cons

  • PRO: Good balance of capacity/speed/safety

  • PRO: Cost-effective redundancy

  • CON: Slower writes due to parity calculation

    This setup prioritizes redundancy but has associated performance or capacity costs.

Performance Methodology

Last updated: May 2026

Performance estimates on StashRaid use the Arcserve write-penalty framework, the industry-standard model for comparing RAID-level throughput. All multipliers, penalties, and mathematical baselines are detailed below to ensure our calculations are fully transparent, auditable, and academically grounded.

Core Mathematical Model

Our calculator derives RAID group capabilities from the baseline specifications of a single enterprise-grade mechanical spindle or solid-state cell:

Metric ComponentBaseline CapacityCalculation Logic / Rationale
Drive Throughput Base250 MB/s seq / 80 random IOPSRepresents a typical modern enterprise 7,200 RPM mechanical HDD (SATA/SAS).
Group Read Speed MultiplierN × baselineAll available drive spindles read data concurrently across the stripe width.
Group Write Speed MultiplierN / Write PenaltyWrite operations suffer from parity generation and LVM/filesystem overhead.

Write Penalties by RAID Level

The write penalty represents the number of physical disk I/O operations required to complete a single logical write. The table below lists the multipliers used in our calculation engine:

RAID LevelWrite PenaltyOperational Overhead & RationaleStandard Reference Source
RAID 01No redundancy overhead. Pure striping.Standard RAID spec
RAID 11Mirrors written simultaneously; equals single-drive speed.Standard RAID spec
RAID 54Requires 4 I/O actions: read old data, read old parity, write new data, write new parity (RMW).Arcserve Framework
RAID 66Requires 6 I/O actions: read old data, read 2× old parity, write new data, write 2× new parity.Arcserve Framework
RAID 102Every logical write is mirrored to two physical disks.Arcserve Framework
RAID 508Stripes across multiple nested RAID 5 sets.Nested parity overhead
RAID 6012Stripes across multiple nested RAID 6 sets.Nested parity overhead
JBOD1Spanned configuration. No striping or parity.Standard JBOD spec
ZFS RAID-Z13Copy-on-Write (CoW) + variable stripe width completely eliminates full-stripe read-modify-write.Ars Technica Benchmarks
ZFS RAID-Z25Double parity under CoW: 2.5–4.2× faster sequential writes than hardware RAID 6 in tests.Ars Technica Benchmarks
ZFS RAID-Z37Triple parity under CoW; avoids write-hole penalties.OpenZFS Community Model
Synology SHR4.5RAID 5 mdadm base (4) + ~12% LVM spanning and mapping layer overhead.Empirical mdraid+LVM estimates
Synology SHR-26.5RAID 6 mdadm base (6) + ~8% LVM spanning and multi-array lookup overhead.Empirical mdraid+LVM estimates

Academic & Empirical Sources

  • Arcserve Storage Reference: Establishes the core I/O write penalties (4 for RAID 5, 6 for RAID 6, 2 for RAID 10) universally used by manufacturers like Dell EMC, HPE, and NetApp.
  • Ars Technica Benchmarks (Jim Salter): Empirical benchmarks on 8× Ironwolf HDDs comparing RAID-Z2 vs traditional RAID 6. Demonstrates the write-throughput advantages of OpenZFS Copy-on-Write.
  • SNIA (Storage Networking Industry Association): Mathematical modeling guidelines for estimating storage array IOPS and sequential performance multipliers.

Important Caveats & Limitations

While these calculations are accurate representations of storage mechanics, real-world performance varies due to:

  • Workload Pattern: The calculations assume contiguous sequential throughput. Random 4K read/write loads suffer seek time bottlenecks on mechanical HDDs, yielding much lower performance.
  • RAM and SSD Caching: ZFS relies heavily on the Adaptive Replacement Cache (ARC) in RAM. Synology systems use LVM caching or NVMe SSD cache pools which mask the physical disk write penalty for active data sets.
  • Stripe Alignment: Partial-stripe writes require reading unmodified blocks to generate parity, resulting in massive performance drops. Full-stripe writes avoid this read overhead.
  • Hardware Offloading: Dedicated hardware RAID controllers cache writes in flash, allowing the system to verify writes immediately without waiting for disk sync.