Seagate Nvme Ssds Data Integrity Reliability Tb100 1 1606us

Extrait du fichier (au format texte) :

Seagate Nytro® NVMe SSDs:
More Than Speeds and Feeds for
Business-Critical Needs
Technical Brief

Overview
Non-Volatile Memory Express (NVMe) accelerates SSD performance beyond the limits of legacy bus architectures like SATA and SAS. This new protocol uses the PCIe interface and delivers optimizations to lift business-class storage performance to new heights. However, SMBs and large organizations demand more than fast speeds,
which is why the Seagate Nytro XF1440 (2.5-inch), Nytro XM1440 (M.2) and Nytro
XP7102 (HHHL) enterprise-grade SSDs integrate several features critical to protecting data integrity and extending drive reliability.

Error Detection and Correction
Seagate Nytro SSDs include several data protection methods already common throughout the NAND world, including cyclic redundancy check (CRC) code for data passing between the host and storage device as well as error-correcting code (ECC)
applied to data written to and read from NAND. Seagate NVMe SSDs take protection much further, however, with three key technologies:
Low-density parity-check (LDPC) code
Nytro SSDs feature LDPC as their first line of defense against media error correction. This methodology is particularly suited to moving data through noisy channels with no integrity loss. Also found in 10Gbit Ethernet, high-performance
Wi-Fi technologies and hard drives, Seagate implements LDPC alongside user data within each NAND memory page. The LDPC technology featured in
Nytro SSDs combines hard and soft decision decoding to achieve significantly better error correction than SSDs using conventional BCH codes. Applied to
SSDs, LDPC can help to extend NAND's longevity and deliver enterprise-class endurance by reducing the potential for corrupted data as NAND reaches its end of life.
End-to-End Data Protection
The Nytro SSD's controller hardware appends an embedded logical block address
(LBA), a CRC and an ECC field to each block of host data. The controller checks these fields prior to data being returned to the host after it has been processed through the SSD and stored on the NAND. This provides protection against silent data corruption across all memory in the data path. Additionally, end-to-end data protection allows the controller and firmware to safeguard user data without assistance from the host system, which in turn frees up host processing cycles.

Seagate Nytro® NVMe SSDs: More Than Speeds and
Feeds for Business-Critical Needs
RAISE: RAID-like NAND Protection
Seagate's engineers applied the data protection processes of conventional RAID to the die of NAND devices, creating an implementation called a Redundant Array of Independent Silicon Elements (RAISE!"). The Seagate RAISE technology protects against rare errors that conventional ECC protection might miss. In a 32-die implementation, for example, the user data on any given die is evenly spread across the other 31 die, ensuring that no data exists uniquely at any given location. In exchange for this redundant protection which improves the drive's Uncorrectable Bit Error Rate (UBER) by a factor of almost one quadrillion
 RAISE carries a capacity cost of only about three percent. Out of every 32 die, 31 are user-accessible (97% of total NAND
storage). The result is a remarkably small capacity sacrifice in exchange for a tremendous leap in data defense.