All-flash storage device supplier Pure Storage expects the capacity of its Direct Flash Module (DFM) solid-state drives to grow sixfold in the next few years, to 300 TB capacity. The company expects such dramatic capacity increases to be possible due to the increased areal density of 3D NAND through increased layer count and other factors. His SSDs for client PCs are also expected to grow in capacity over the next few years, but less dramatically.
Pure Storage CTO Alex McMullan said in the following interview: blocks and files“Today we are shipping 24 TB and 48 TB. [DFM] drive.you can expect […] We announced our ambition to deliver 300 TB drive capability by 2026 here and made many announcements at the Accelerate conference on larger and larger drive sizes. ”
Pure Storage’s proprietary DirectFlash Module (DFM) drives are used exclusively in its FlashArray and FlashBlade storage systems. DFMs are similar to ruler form factor SSDs and also use standard U.2 NVMe connectivity, but are designed specifically for specific machines. The module features full custom firmware and a 3D TLC or 3D QLC NAND memory device, but uses an enterprise-grade SSD controller, although the company uses a custom package to increase the capacity of its flash chips. I don’t know if there are any.
There are ways to increase Pure’s DFM capacity by 6x in 3 years, but it’s not going to be easy. The most obvious way is to use more sophisticated 3D NAND ICs. Currently, the company uses 112-160 layers of his 3D NAND devices, but plans to increase the number of active layers to 400-500 in the next five years, which will increase the capacity of 3D NAND ICs. increase.
“Every fab vendor has a plan and a path to get to 400-500 tiers in the next five years,” McMullan said. “Of course, that in itself will help us.”
3D NAND manufacturers introduce new generation devices (i.e., increasing the number of active layers) approximately every two years. Manufacturers are expected to begin production of ~200-layer 3D NAND this year, so expect ~300-layer 3D NAND in 2025 and ~400-layer 3D NAND in 2027. To significantly increase DFM capacity by 2026, Pure Storage will have to rely on more than just increasing the number of active layers.
Existing DFMs can accommodate a limited number of 3D NAND packages. Therefore, Pure Storage either uses packages that pack more 3D NAND devices (which would require a controller to work with such packages) or increases the number of packages and adopts a larger footprint for DFM. can do.
Perhaps Pure Storage’s plans to increase the capacity of its DFM drives by a factor of six in three years seem ambitious (although we like the idea of a premier flash drive with 300TB capacity! On the other hand, 3D NAND Flash and memory packaging technology is evolving rapidly, and I wouldn’t be surprised if Pure’s DFM-based storage systems exceed the storage densities offered by storage systems using HAMR hard drives within a few years.
