Differentiated Protection and Hot/Cold-Aware Data Placement Policies through k-Means Clustering Analysis for 3D-NAND SSDs

3D-NAND flash memory provides high capacity per unit area by stacking 2D-NAND cells having a planar structure. However, because of the nature lamination process, frequency error occurrence varies depending on each layer or physical cell location. This phenomenon becomes more pronounced as number write/erase (Program/Erasure) operations increases. Error correction code (ECC) is used for in majority flash-based storage devices, such SSDs (Solid State Drive). As this method constant level data protection all-flash pages, there limitation memory, where rate Consequently, paper, pages and layers with varying rates are classified into clusters using k-means machine-learning algorithm, cluster assigned different strength. We classify based occurrences measured at end endurance test, areas vulnerable to errors, it shown an example providing differentiated strength adding parity stripe. Furthermore, retention errors identified rates, bit significantly reduced through our hot/cold-aware placement policy. show that proposed differential policies improve reliability lifespan compared existing ECC- RAID-type scheme.