Abstract

To meet the increasing demand for higher storage density, modern NAND flash-based SSDs employ Quad-Level Cell (QLC) technology, which stores four bits per memory cell. However, it exacerbates the read disturb issue, where repeated read operations gradually shift the threshold voltages of unselected NAND flash cells. This voltage drift leads to frequent read retries and accelerates device wear, ultimately degrading performance and endurance. To address this issue, we propose Localized Page Allocation (LPA), a novel data mapping scheme that allocates each page to a restricted region of the wordline. Unlike the conventional scheme, LPA assigns all four bits in a single QLC cell to the same page. This approach confines read operations to a limited portion of the bitlines, thereby reducing the number of NAND flash cells exposed to read disturb. However, this localized access increases the complexity of sensing operations. To address this challenge, we introduce Progressive Voltage Convergence (PVC) method that adaptively determines the next sensing voltage for each bitline based on its previous sensing result. This fine-grained control reduces the number of sensing steps during read operations. For additional optimization, we employ lossless compression, which reduces the number of sensing steps for the compressed pages. Experimental evaluations using real-world workloads demonstrate that our design improves I/O performance by up to 9% and extends endurance by 78% compared to the state-of-the-art techniques. LPA and PVC require minor hardware modifications to the conventional NAND flash chips and the SSD controller, which leads to small area and power overhead