Nonvolatile Memory Logical Page Sorting for FTL Table Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing nonvolatile memory devices, such as NAND flash memory, face inefficiencies in managing logical addresses to physical addresses, leading to increased storage requirements for mapping tables due to the need for precise address conversion and sorting.

Innovation Solution

Implementing a method where logical pages are written to nonvolatile memory devices in ascending order of their logical addresses, utilizing a tree structure with nodes representing subsets of logical addresses, allowing for efficient access and reducing the number of bits required for mapping by sorting pages within physical clusters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If logical pages are written to nonvolatile memory devices without sorting by logical address, then write operations can be performed faster, but the flash translation layer table size increases requiring more bits for address mapping

Engineering Contradiction:
Improvewrite operation speedVSAvoidFTL table size
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The patent applies preliminary action by sorting logical pages by their logical addresses before writing them to physical locations in the nonvolatile memory device. This pre-sorting operation prepares the data in advance, allowing the FTL to maintain a more compact translation table while still enabling efficient sequential writes. The sorting is performed on the logical page addresses before the actual write operation, thus preventing table bloat from unsorted address mappings.

Inventive Principle:
Principle #10Preliminary action

2Quantity of substance

If logical pages are sorted by logical address before writing, then FTL table size is reduced, but an additional sorting operation is required before writing

Engineering Contradiction:
ImproveFTL table sizeVSAvoidwriting process complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent merges the sorting operation with the existing write process by integrating the logical address sorting step directly into the write workflow. Instead of treating sorting as a separate complex stage, the system combines sorting and writing into a coordinated process where logical pages are sorted by their addresses and then immediately written to the nonvolatile memory device in the sorted order. This integration reduces overall system complexity despite adding a sorting step, as the sorting operates on already-structured data and prepares it for sequential storage.

Inventive Principle:
Principle #5Merging (Combining)

3Measurement precision

If pages are stored in ascending order of logical addresses, then address mapping becomes more efficient with fewer bits required, but random access performance may be impacted

Engineering Contradiction:
Improveaddress mapping efficiencyVSAvoidrandom access speed
Core Design Contradiction:
Measurement precisionVSSpeed

Solution Approach 1:

The patent applies segmentation by organizing the nonvolatile memory device into multiple blocks, each capable of storing sorted logical pages. This segmentation allows the system to maintain sorted order within each block for efficient FTL table management while enabling parallel or independent access to different blocks for random read operations. The logical address sorting is performed at the block level, creating manageable segments that balance mapping efficiency with access performance.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS9483413B2Nonvolatile memory devices and methods of controlling the same
Publication Date: 2016.11.01 SAMSUNG ELECTRONICS CO LTD
  • US9483413B2 patent drawing
  • US9483413B2 patent drawing
  • US9483413B2 patent drawing

AI summary

At least one example embodiment discloses a method of controlling a nonvolatile memory device including a plurality of blocks, each block including a plurality of physical pages. The method includes receiving a plurality of logical pages associated with a first plurality of logical addresses, respectively, and writing the first plurality of logical pages to the plurality physical addresses according to an ascending order of the logical addresses of the first plurality of logical pages.