How maxRAID works
One idea.
A better RAID.
maxRAID is a flash translation layer that runs in the host, above the RAID layer. It changes the write stream before it ever reaches your SSDs. Speed, capacity, endurance and cost all follow.
The problem
RAID was built for hard drives.
Parity RAID updates data in place. On SSDs, every small write becomes extra reads and extra writes — and the SSD’s own controller adds more hidden work on top. Speed drops. Drives wear out.
Traditional parity RAID
One small write. Four I/Os.
- Read old data
- Read old parity
- Write new data
- Write new parity
Every random write becomes a read‑modify‑write. Throughput collapses and the parity drive is hammered.
maxRAID
Many small writes. One stripe.
Writes gather in memory, compressed and aligned
Parity is computed in memory. The stripe is written once, in full, with no reads. The SSDs only ever see sequential writes.
The solution
Six steps. Every write.
- 01
Gather
Incoming writes — random, small, from any application — are collected in memory instead of being sent straight to the drives.
- 02
Compress
Each block is compressed in line. Less data written means more free space, fewer SSDs and less wear.
- 03
Linearize
Blocks are laid out sequentially and aligned to the SSDs’ internal geometry, so the SSD controllers avoid unnecessary work.
- 04
Commit
Parity is calculated in memory and the stripe is written once, in full, across every drive. No reads. No read‑modify‑write.
- 05
Collect
A host‑side garbage collector reclaims space with near‑theoretical efficiency — within 0.5% of the optimal model.
- 06
Trim
Blocks of all zeros or all ones consume no space. Trim and discard work everywhere, without client or hypervisor support.
Why above RAID
The FTL belongs
in the host.
Inside an SSD, the flash translation layer can only see one drive. Above RAID, it sees the whole array — so RAID only ever processes linear, full‑stripe writes.
| Amplification | Read | Write |
|---|---|---|
| maxRAID‑5 | None | 1.15× |
| maxRAID‑6 | None | 1.33× |
- What this means: for each 1 GB you write, maxRAID‑5 writes 1.15 GB to the SSDs. RAID‑10 writes 2 GB, because it mirrors everything.
- Longer SSD life: in our 120 TB model, maxRAID needs no SSD replacements in five years. Under the same workload, RAID‑5 wears its SSDs out in about 20 months.
Where it shines
Made for real workloads.
The advantages are largest where writes matter: moderate to heavy write loads — at least 0.3 drive writes per day — with some data compressibility.
Databases
Steady low latency under mixed reads and writes.
Virtualization
Dense VM hosts without the RAID‑10 tax.
AI & analytics
Massive datasets, sustained throughput.
Cloud & multi‑tenant
Predictable performance for every tenant.
See it on your own hardware.
Try maxRAID free for 120 days, on arrays up to 100 TB.