A metadata access method, memory, and device
By using a target leaf node that is larger than the minimum read/write unit of the memory in the metadata index, and simplifying the location through append writes and mapping relationships, the problem of low metadata writing efficiency is solved, and the performance of the data storage system is improved.
CN122086940APending Publication Date: 2026-05-26HUAWEI TECH CO LTD
View PDF 0 Cites 0 Cited by
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUAWEI TECH CO LTD
- Filing Date
- 2024-11-25
- Publication Date
- 2026-05-26
AI Technical Summary
Technical Problem
In existing technologies, metadata writing efficiency is low, especially during batch writing due to severe write amplification problems caused by tree-like index structures such as B+ trees, which affects the performance of data storage systems.
Method used
By using a target leaf node that is larger than the minimum read/write unit of memory in the metadata index, the number of leaf nodes is reduced, metadata is written in an append-only manner, and the mapping relationship is used to simplify the location of metadata storage location, thus simplifying the tree index structure.
Benefits of technology
It improves metadata writing efficiency, reduces the number of read, modify, and write operations, and enhances the performance of the data storage system.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122086940A_ABST
Abstract
This application discloses a metadata access method, memory, and device. In this application, a storage device receives a first instruction message, which instructs the storage device to write first metadata into a metadata index. The storage device queries the metadata index based on the identifier of the first metadata to obtain a target leaf node in the metadata index. The size of the target leaf node is larger than the minimum read / write unit of the memory. The storage device writes the first metadata into the target leaf node. The larger size of the target leaf node in the metadata index allows for the storage of more metadata within that node, reducing the need for numerous leaf nodes in the metadata index and simplifying its structure, thus effectively improving metadata access efficiency. When writing the first metadata into the metadata index, only a small number of target leaf nodes need to be operated on, reducing the number of read / write operations and improving metadata writing efficiency.
Need to check novelty before this filing date? Find Prior Art