A team of Japanese researchers led by Ken Takeuchi, a professor at Chuo University, Tokyo has announced a hybrid SSD framework that combines a ReRAM buffer with standard NAND flash components. The resulting memory chip offers significantly improved performance and longevity, and could potentially be a model for manufacturers trying to overcome some of NAND flash’s scaling issues.
Long-term NAND scaling currently faces a number of problems. NAND flash constructed on smaller process nodes is less durable, and attempts to move from MLC (two bits per cell) to TLC (three bits per cell) have been slow thanks to the greatly reduced number of write cycles TLC flash can perform. A ReRAM hybrid could solve a number of these issues.
ReRAM (Resistive Random-Access Memory) is one of a vast number of projects under development by corporations looking for next-generation storage mediums to replace NAND flash. In this context, its best advantage over NAND flash is that ReRAM is both more durable and requires far less energy per read/write.
Takeuchi and his team propose a hybrid NAND flash drive that will combine 256GB of NAND with roughly 1GB of ReRAM for both cache and storage. Instead of writing directly to the NAND, drive writes would be cached within ReRAM until a certain point was reached. These would then be written in bulk, thereby lowering data fragmentation and greatly reducing the number of write cycles the NAND flash performs.
Data is buffered and handled according to three algorithms:
- First, an anti-fragmentation filter makes sure that writes are saved until they reach at least sixty percent of 16K depending on the ReRAM’s capacity.
- Second, data that’s overwritten NAND is copied back to the buffer (to prevent fragmentation and additional re-write cycles).
- Finally, the buffer keeps a log of the most recently accessed addresses. If data is being written and re-written rapidly, the updates occur in ReRAM, not the NAND.
One of the most outstanding features of Takeuchi’s proposal is that the amount of ReRAM required is relatively small compared to the overall storage capacity of the flash drive. That’s vital when remembering the impact of expansion cost advantages.
Manufacturers who are currently working on building out NAND semiconductor fabrication plant capacities are not going to do anything drastic to rework production for ReRAM unless they absolutely have to. Incorporating a small amount of ReRAM around high-capacity NAND preserves the value of the current investment without the need for a separate production initiative.
There’s also a good chance that the ReRAM buffer’s cost could be offset by the use of TLC flash at some point in the near future. Takeuchi’s team designed their prototype by assuming the use of TSVs (through-silicon vias) but think the design could be adapted to other approaches currently in mass production. TSVs have been tested and are currently under development, but aren’t yet being distributed in widespread numbers.
The ReRAM buffer will not solve the long-term need for a NAND replacement, but it would be a good temporary solution while the search for a suitable substitute continues. The researchers have yet to build an actual prototype of their design, so current conclusions, while positive, should be considered solely theoretical.
Even if the ReRAM idea never gains any solid ground, we can expect to see more of these kinds of buffers in the very near future. Theoretically, NAND flash manufacturers could design specialized blocks of reliable SLC flash cache surrounded by an MLC or TLC main drive. While the entire flash drive would still use NAND, rapid writes would first by performed on the SLC block, where its much higher write endurance would provide a similar benefit. The lower-durability TLC/MLC flash would be tread as a “Read/Write Occasionally” data storage area.
The types of projects we’ll see underway in the next 3-5 years will depend on how rapidly NAND alternatives mature and how successfully corporations can launch future process nodes. Such flash drives could take a few years to trickle down into consumer hands depending on cost scaling and enterprise suitability.
Buffer drives like ReRAM could form a fourth tier of storage alongside cache drive SSDs, standalone SSDs, and HHD (hybrid hard drives).
Analysts believe that the cache drive and HDD solutions will exponentially grow in popularity over the next year and a half. Buffered drives could offset the reliability concerns that fuel the cache drive and HDD deployments and reassure customers that they only need one product, not two.