Flash Memory Summit 2012
Last week, I attended the 7th annual Flash Memory Summit held in Santa Clara, CA. In addition to a 30% increase in the number of exhibitors and attendees, there were several hot topics discussed throughout the event in speaking sessions, hallways, and on the exhibit floor. I have selected five of the hottest areas of interest here:
- The shrinking size of NAND lithography – To reduce NAND power consumption and footprint, NAND providers continue to push for smaller NAND flash (we are now approaching 1X nm). As this occurs, Floating Gate geometry also shrinks causing cell endurance & data retention degradation as the potential for unintended bit flips and data corruption will increase. Currently, NAND flash uses floating gate technology to store data. (A floating gate is a CMOS- (complementary metal-oxide semiconductor) based transistor that is capable of holding an electrical charge.) As NAND sizes continue to get smaller, current technologies such as floating gates are approaching end of life (2013 estimated timeframe) and new methodologies such as 3D NAND cells (2014-2015 estimated transition timeframe) must be matured to expand the market for solid state storage.
- The increasing importance of and focus on data integrity and device endurance – As NAND lithography gets smaller and enterprise datacenters incorporate more SSDs into mission critical production environments (that store a company’s most valuable data), the need for advancements in power loss data protection, error detection and correction code, Low Density Parity-Checks (LDPC), signal and processing techniques are increasing.
- SSD Speed is great, but the consistency of the performance is also important – SSDs must provide consistent performance over their designated life span to meet datacenter goals. This means that all SSD background activities such as wear leveling, garbage collection, free space reclamation, and error correction code must be improved such as to not have an impact on performance. It also means that mixed and changing read and write workloads do not dip below IOPs level that 100% reads or 100% writes can achieve.
- NAND cost per bit (measured in $/GB) is trending downward – NAND Flash is the single most expensive item in an SSD’s bill of materials and drives SSD pricing to OEMs and end users. As NAND cost per bit is reduced, SSDs prices are expected to follow suite – enabling more companies to purchase and implement SSDs.
- The Enterprise SSD market is segmenting into three sub-markets – One size does not fit all. A single SSD does not meet the diverse needs of different applications with significantly different workload characteristics. The storage industry is beginning to see the emergence of three sub-markets within enterprise SSDs:
- Entry Level SSDs designed for read intensive workloads
- Mainstream SSDs designed for complex, mixed read/write workloads
- High Endurance SSDs designed for write intensive workloads
Seagate had a strong presence at the FMS with 13 speaking sessions – including a Keynote session by Dr. Jeff Burke, VP Market Intelligence at Seagate – a large booth, technology demos of the Seagate hybrid drive for ultra-thin laptops and a technology demo of 12Gb/s SAS working with both Seagate SSDs and HDDs. Click Here to see more information on all of the presentations from the various Flash Memory Summit speaking sessions.
