VAST Data warns petabyte SSDs will break RAID in parallel file systems
VAST Data presented a simulation at SC26 of its 900 PB storage system showing that as SSDs scale to petabyte capacities, current RAID schemes used in parallel file systems face a 1-in-11 chance of data loss over 10 years, because larger drives take longer to rebuild after failures. Principal technical strategist Glenn Lockwood said global declustered RAID with extra parity reduces that risk to about 1 in 54 million, while VAST's own erasure-coding scheme with at least three parity strips reaches roughly 1 in 62 billion.
GoKawiil's interpretation of the reporting above, not reported fact.
The findings suggest that as flash capacity grows to save power—since storage power scales with bandwidth rather than drive size—existing RAID architectures used by rivals like DDN and HPE Cray could become a reliability liability at scale. VAST's framing positions its own erasure-coding approach as better suited to a future of power-constrained AI data centers using massive SSDs, though the comparison comes from VAST's own simulation rather than independent testing.
- VAST Data's simulation finds a 1-in-11 risk of data loss over 10 years for a 900 PB system using today's RAID in parallel file systems with petabyte-scale SSDs.
- VAST's recommended fix—system-wide erasure coding with at least three parity strips—reportedly cuts that risk to about 1 in 62 billion.
- Larger SSDs are attractive because storage power scales with bandwidth, not capacity, letting 122 TB drives use a fraction of the power of 15 TB drives for the same storage volume.
Source: tomshardware.com — Shane Downing, 2026-11-20
Published there as: “VAST says petabyte SSDs will outgrow today’s parallel file system RAID”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.