WO2016086899A1 - 一种存储设备元数据管理处理方法及系统 - Google Patents

一种存储设备元数据管理处理方法及系统 Download PDF

Info

Publication number
WO2016086899A1
WO2016086899A1 PCT/CN2015/096496 CN2015096496W WO2016086899A1 WO 2016086899 A1 WO2016086899 A1 WO 2016086899A1 CN 2015096496 W CN2015096496 W CN 2015096496W WO 2016086899 A1 WO2016086899 A1 WO 2016086899A1
Authority
WO
WIPO (PCT)
Prior art keywords
super block
address
storage device
sector address
mapping information
Prior art date
Application number
PCT/CN2015/096496
Other languages
English (en)
French (fr)
Inventor
周溱
阳学仕
Original Assignee
上海宝存信息科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海宝存信息科技有限公司 filed Critical 上海宝存信息科技有限公司
Priority to US15/528,839 priority Critical patent/US10445231B2/en
Publication of WO2016086899A1 publication Critical patent/WO2016086899A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00—Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02—Addressing or allocation; Relocation
    • G06F12/06—Addressing a physical block of locations, e.g. base addressing, module addressing, memory dedication
    • G06F12/0607—Interleaved addressing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/14—Error detection or correction of the data by redundancy in operation
    • G06F11/1402—Saving, restoring, recovering or retrying
    • G06F11/1415—Saving, restoring, recovering or retrying at system level
    • G06F11/1441—Resetting or repowering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00—Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02—Addressing or allocation; Relocation
    • G06F12/06—Addressing a physical block of locations, e.g. base addressing, module addressing, memory dedication
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00—Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02—Addressing or allocation; Relocation
    • G06F12/06—Addressing a physical block of locations, e.g. base addressing, module addressing, memory dedication
    • G06F12/0646—Configuration or reconfiguration

Definitions

  • the present invention relates to a computer storage technology, and in particular, to a storage device metadata management method and system.
  • Flash memory NAND Flash
  • HDD hard disk drive
  • FTL Flash Translation Layer
  • LBA Logical Block Address
  • PBA Physical Block Address
  • the dynamic mapping table forms part of the metadata to implement the function of flash LBA to PBA conversion.
  • SSDs Solid-state drives
  • the metadata that needs to be maintained includes the following parts: 1. The basic format of the SSD, including the type of flash memory used, the number, and the organizational structure. 2, those blocks are bad blocks, that is, blocks that cannot be used. 3. How the logical sector address (LBA) maps to the physical sector address (PBA). 4. If the same LBA corresponds to multiple PBAs, which one corresponds to the latest version. 5, the number of erasures per block. 6, other information that needs to be retained, such as various statistics. If you need to save metadata on flash media, you need to consider the following: 1. Where is it? Each erase block can be bad or bad, and how to find the metadata the next time you turn it on. 2, when to write. Metadata is constantly changing, and flash is not rewritten in its original location.
  • an object of the present invention is to provide a storage device metadata management method and system for solving the problem of low reliability of storage and recovery of metadata on a flash memory medium in the prior art.
  • the present invention provides a storage device metadata management method, the storage device metadata management method comprising: composing a super block by at least one erase block; the super block retaining a storage a space of the super block related metadata; the super block related metadata includes a sequence number allocated to the super block and address mapping information in the super block; the address mapping information stores a physical sector address to a mapping relationship of logical sector addresses; when writing data to the super block, the corresponding address mapping information of the data is also written into the super block; when the system is restored, according to each of the super blocks
  • the address mapping information recovers a page table of the storage system; the page table identifies a mapping relationship between a logical sector address and a physical sector address.
  • the address mapping information in each of the super blocks is reversed to restore the page table of the storage system.
  • the space for storing the super block related metadata is immediately following the data storage space of the super block.
  • the super block related metadata further includes an erasure number of the super block and/or at least one type of statistics of the super block.
  • the physical sector address corresponding to the logical sector address in the page table is a superblock to which the serial number is assigned.
  • the corresponding physical sector address in the address mapping information is a superblock to which the serial number is assigned.
  • the physical sector address corresponding to the logical sector address in the page table is the post-write of the super block.
  • the corresponding physical sector address in the address mapping information is the post-write of the super block.
  • the storage device metadata management method further includes: when the system is restored, performing a full scan on the uncompleted write super block to read the logical address and the physical address for the super block that has not completed writing. Mapping the relationship and restoring the page table in the uncompleted write superblock.
  • the static data and/or bad block information in the metadata is stored in a starting erase block of the storage device.
  • the present invention provides a storage device metadata management system, the storage device metadata management system comprising: a super block management module, configured to form at least one erase block into one super block; a space of superblock related metadata; the super block related metadata includes a sequence number assigned to the super block and address mapping information within the super block; the address mapping information holds a physical sector address to a logical fan a mapping relationship of the area address; when the data is written to the super block, the corresponding address mapping information of the data is also written into the super block; the system recovery module is configured to Address mapping information in the super block recovers the storage system Page table; the page table identifies a mapping relationship between a logical sector address and a physical sector address.
  • the system will reverse the address mapping information in each of the superblocks upon system recovery to restore the page table of the storage system.
  • the space for storing the super block related metadata is immediately following the data storage space of the super block.
  • the super block related metadata further includes an erasure number of the super block and/or at least one type of statistics of the super block.
  • the physical sector address corresponding to the logical sector address in the page table is the serial number assignment.
  • the physical sector address corresponding to the logical sector address in the page table is the super block.
  • system recovery module is further configured to: perform a full scan on the uncompleted write super block to read a mapping relationship between a logical address and a physical address for the super block that has not completed writing, and restore the The page table in the super block that was not written.
  • the present invention further provides a storage device including the storage device metadata management system as described above.
  • the storage device is a flash-based solid state hard disk.
  • the storage device metadata management method and system of the present invention have the following beneficial effects: taking into account various aspects such as fault tolerance, performance cost, and implementation complexity, the page table can be completely restored at boot time, and the SSD is started. Can work normally.
  • the present invention can satisfy the following requirements: 1. It is not affected by dynamically generated bad blocks. 2, not affected by sudden power outages, system crashes. 3. Even if some data cannot be recovered due to extreme fault conditions, the impact should be limited. 4, the boot time is as short as possible. 5, the shutdown time is as short as possible. 6, the performance impact on normal work is as small as possible. 7, space requirements are as small as possible. And the technical solution of the invention can achieve the following performances: 1. The shutdown time is less than 10ms, and at most two flash chips must be written in two pages. 2. At most, you need to read one thousandth of the total capacity when booting. 3, occupying one thousandth of the total flash memory capacity. 4, under normal work, there is basically no impact on performance.
  • FIG. 1 is a schematic flowchart diagram of an embodiment of a storage device metadata management method according to the present invention.
  • FIG. 2 is a block diagram showing an embodiment of a storage device metadata management system of the present invention.
  • FIG. 3 is a block diagram showing an embodiment of a storage device of the present invention.
  • FIG. 4 is a block diagram showing an embodiment of a storage device of the present invention.
  • the invention provides a storage device metadata management method.
  • the storage device metadata management method includes:
  • Step S1 a super block is composed of at least one erase block; the super block retains a space for storing the super block related metadata; the super block related metadata includes a sequence number assigned to the super block and Address mapping information in the super block; the address mapping information holds a mapping relationship between a physical sector address and a logical sector address.
  • the metadata of the storage device includes: a basic format of a storage device (such as an SSD, a Solid State Drive, a solid state hard disk), including a type of flash (Flash) used, a quantity, an organizational structure, and the like, and the data is reformatted, Otherwise unchanged, it can also be called static data; bad block information, which blocks are bad blocks, that is, unusable blocks, bad blocks may increase with the passage of time, but the addition speed is very slow; logical sector address (LBA, Logical Block Address, also called logical block address) how to map to physical sector address (PBA, physical block address, also called physical block address); if the same LBA corresponds to multiple PBA, then which One corresponds to the latest version; the number of erasures for each erase block; other information that needs to be retained, such as various statistics.
  • a basic format of a storage device such as an SSD, a Solid State Drive, a solid state hard disk
  • flash flash
  • the space storing the super block related metadata is immediately following the data storage space of the super block. Therefore, we can refer to the space in which the super block related metadata is stored as an epilog.
  • static data and/or bad block information in the metadata is stored in the storage device The starting erase block.
  • the bad block information may be a bad block table.
  • the super block related metadata further includes an erasure count of the super block and/or at least one statistical data of the super block.
  • Step S2 when writing data to the super block, the corresponding address mapping information of the data is also written into the super block. Specifically, when data is written to a certain super block, the write data immediately writes the corresponding address mapping information of the written data to the super block.
  • the system accesses the data through the page table, and the mapping relationship between the logical sector address and the physical sector address is identified in the page table, but the written address mapping information data is inverse to the data in the page table, that is, the address mapping The data in the information identifies the mapping relationship between the physical sector address and the logical sector address. With this method, it is equivalent to distributedly storing part of the metadata information (including the entire page table information) in each super block.
  • various statistical information is also written to the space of the superblock storing the superblock related metadata.
  • Step S3 When the system is restored, the address mapping information in each of the super blocks is reversed to restore a page table of the storage system; the page table identifies a mapping relationship between a logical sector address and a physical sector address. Since the small page table in each super block contains only the mappings in this super block, and this mapping is reversed, ie PBA->LBA, we need to flip it to get the LBA->PBA mapping. This is still an O(n) algorithm.
  • the mapping from PBA to LBA may be many-to-one, that is, the same LBA may correspond to multiple PBAs, and only one of them is the final version. If the allocation and record serial number is written for each PBA, the cost is relatively large.
  • the method employed is to assign a serial number to each superblock and record it in the postscript.
  • the first two PBAs First, when the same logical sector address exists in the address mapping information of different superblocks, the logical fan in the page table The physical sector address corresponding to the area address is the corresponding physical sector address in the address mapping information of the super block to which the serial number is assigned. That is, between different super blocks, the new block data is new, and the old block data is old.
  • the storage device metadata management method further includes: performing a full scan of the uncompleted write super block to read a logical address for a super block that has not completed writing when the system is restored. A mapping relationship with the physical address and restoring the page table in the uncompleted written super block.
  • the storage device metadata management method can be applied to an SSD, and can also be applied to Traditional hard disk-based arrays, or arrays of next-generation storage media.
  • the storage device metadata management system 1 includes a super block management module 11 and a system recovery module 12. among them:
  • the super block management module 11 is configured to group at least one erase block into one super block; the super block retains a space for storing the super block related metadata; the super block related metadata includes allocating the super block Serial number and address mapping information in the super block; the address mapping information holds a mapping relationship between a physical sector address and a logical sector address; when data is written to the super block, the data is corresponding The address mapping information is also written to the super block.
  • the metadata of the storage device includes: a basic format of a storage device (such as an SSD, a Solid State Drive, a solid state hard disk), including a type of flash (Flash) used, a quantity, an organizational structure, and the like, and the data is reformatted, Otherwise unchanged, it can also be called static data; bad block information, which blocks are bad blocks, that is, unusable blocks, bad blocks may increase with the passage of time, but the addition speed is very slow; logical sector address (LBA, Logical Block Address, also called logical block address) how to map to physical sector address (PBA, physical block address, also called physical block address); if the same LBA corresponds to multiple PBA, then which One corresponds to the latest version; the number of erasures per physical block; other information that needs to be retained, such as various statistics.
  • a basic format of a storage device such as an SSD, a Solid State Drive, a solid state hard disk
  • flash flash
  • the space storing the super block related metadata is immediately following the data storage space of the super block. Therefore, we can refer to the space in which the super block related metadata is stored as an epilog.
  • the static data and/or bad block information in the metadata is stored in a starting erase block of the storage device.
  • the bad block information may be a bad block table.
  • the superblock management module 11 when writing data to a certain superblock, writes the data corresponding to the corresponding address mapping information of the written data to the superblock.
  • the system accesses the data through the page table, and the mapping relationship between the logical sector address and the physical sector address is identified in the page table, but the written address mapping information data is inverse to the data in the page table, that is, the address mapping The data in the information identifies the mapping relationship between the physical sector address and the logical sector address.
  • the super block related metadata involved in the super block management module 11 further includes the number of erasures of the super block and/or various statistics of the super block. After writing the data to a certain super block, the super block management module 11 also writes at least one type of statistical information into the space of the super block storing the super block related metadata.
  • the system recovery module 12 is configured to reverse address mapping information in each of the super blocks to restore a page table of the storage system when the system is restored; the page table identifies a mapping of a logical sector address to a physical sector address Relationship; when the same logical sector address exists in the address mapping information of the different superblocks, the physical sector address corresponding to the logical sector address is corresponding to the address mapping information of the superblock to which the serial number is assigned Physical sector address.
  • the small page table in each super block only contains the mapping in the super block, and the mapping is reversed, that is, PBA->LBA, we need To flip it, get the mapping of LBA->PBA. This is still an O(n) algorithm.
  • the PBA to LBA mapping may be many-to-one, that is, the same LBA may correspond to multiple PBAs, of which only one is the final version. If the allocation and record serial number is written for each PBA, the cost is relatively large.
  • the method of the system recovery module 12 is to assign a serial number to each super block and record it in the postscript. Depending on this field, the system recovery module 12 can easily know who is the first two PBAs: First, when the same logical sector address exists in the address mapping information of different super blocks, the page table The physical sector address corresponding to the logical sector address is the corresponding physical sector address in the address mapping information of the super block to which the serial number is assigned.
  • the system recovery module 12 further includes: performing a full scan of the uncompleted write super block to read the logical address and the physical for the super block that has not completed writing when the system is restored.
  • the storage device metadata management system can be applied to an SSD, and can also be applied to a conventional hard disk based array, or an array of next generation storage media.
  • the storage device writes to the first few erase blocks only for the static data in the metadata, which may be referred to as a Master Boot Record (MBR). .
  • MBR Master Boot Record
  • the erase blocks from the at least one flash chip form a super block, as shown in FIG. 3, wherein the LUN (Logic Unit Number) is a separately operated flash unit, SB (Super Block).
  • _ is a super block.
  • the internal organization of each super block is shown in Figure 4.
  • a super block consists of 6 erase blocks from 6 LUNs, one of which is used to store the parity of the super block. .
  • Epilog is a postscript, that is, a space for storing the super block related metadata.
  • the storage device comprises a solid state drive (SSD, Solid State Drive).
  • the storage device metadata management method and system of the present invention have the following beneficial effects: taking into account various aspects such as fault tolerance, performance cost, and implementation complexity, the page table can be completely restored at boot time, and guaranteed.
  • the storage device works fine when it starts up.
  • the present invention can satisfy the following requirements: 1. It is not affected by dynamically generated bad blocks. 2, not affected by sudden power outages, system crashes. 3, even if due to extreme fault conditions, some data can not be recovered, the impact should be limited. 4, the boot time is as short as possible. 5, the shutdown time is as short as possible. 6, the performance impact on normal work is as small as possible. 7, space requirements are as small as possible. And the technical solution of the invention can achieve the following performances: 1.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Quality & Reliability (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Memory System (AREA)
  • Techniques For Improving Reliability Of Storages (AREA)

Abstract

一种存储设备元数据管理方法及系统。所述存储设备元数据管理方法包括：由至少一个擦除块组成一个的超级块；所述超级块保留有一个存储所述超级块相关元数据的空间；所述超级块相关元数据包括为所述超级块分配的序列号以及所述超级块内的地址映射信息；所述地址映射信息保存了物理扇区地址到逻辑扇区地址的映射关系（S1）；当写入数据到所述超级块时，将所述数据相应的地址映射信息也写入到所述超级块中（S2）；系统恢复时，将根据各个所述超级块中的地址映射信息恢复所述存储系统的页表；所述页表标识了逻辑扇区地址到物理扇区地址的映射关系（S3）。上述技术方案能够保证页表在开机时完全恢复，保证存储设备启动时可以正常工作。

Description

一种存储设备元数据管理处理方法及系统 技术领域
本发明涉及一种计算机存储技术，特别是涉及一种存储设备元数据管理方法及系统。
背景技术
基于闪存(NAND Flash)的数据存储技术在过去十几年发展迅速，在很多应用中逐渐取代了传统的磁记录硬盘(hard disk drive,HDD)。闪存相对于传统的基于硬盘的最显著的区别是需要一个相当复杂的闪存转换层(FTL)。由于闪存的不可覆盖写特性，逻辑块地址(LBA:Logical Block Address)到物理块地址(PBA:Physical Block Address)是动态变化的，而FTL需要维护这一动态变化的映射表数据结构.这一动态映射表组成了元数据(metadata)的一部分，来实现闪存LBA到PBA相互转换的功能。在数据存储系统中，除了用户数据需要存储之外，一定数量的元数据(metadata)也是必不可少的。固态硬盘(SSD)为数据存储系统的性能带来了革命性的提升，同时也对元数据的存储和恢复带来了全新的挑战。由于系统重启，断电等事件不可避免，元数据不仅需要存在于易失性的DRAM中，以便于FTL的随时取用和更新，还需要保留在非易失性的存储介质中，而且必须保证百分之百的数据完整性，否则将带来严重的用户数据丢失，甚至整个SSD不再可用。
需要维护的元数据包括以下部分：1，SSD的基本格式，包括使用的闪存种类，数量和组织结构等。2，那些块是坏块，即不可使用的块。3，逻辑扇区地址(LBA)如何映射到物理扇区地址(PBA)。4，如果同一个LBA对应多个PBA，那么哪一个对应最新版本。5，每一块的擦除次数。6，其他需要保留的信息，如各种统计等。如果需要将元数据保存在闪存介质上，还需要考虑以下方面：1，存在哪里。每一块擦除块都有可能是坏的，也有可能变坏，下次开机的时候如何找到元数据。2，什么时候写入。元数据是不停变化的，而闪存是不可在原位置重新写入的。当前，业界的SSD主要分为采取两种方式：一种是使用另外的可靠介质存储元数据。另一种是仅在关机或断电时整体写入元数据。但不管那种方式都有严重问题：因为在元数据的大小和需要的写入性能要求下不存在真正意义的可靠介质。同时，由于系统崩溃，突然断电等可能性存在，整体写入的可靠性不高。
鉴于此，如何提高闪存介质上元数据的存储和恢复的可靠性成为本领域人员亟待解决的问题。
发明内容
鉴于以上所述现有技术的缺点，本发明的目的在于提供一种存储设备元数据管理方法及系统，用于解决现有技术中闪存介质上元数据的存储和恢复的可靠性不高的问题。
为实现上述目的及其他相关目的，本发明提供一种存储设备元数据管理方法，所述存储设备元数据管理方法包括：由至少一个擦除块组成一个超级块；所述超级块保留有一个存储所述超级块相关元数据的空间；所述超级块相关元数据包括为所述超级块分配的序列号以及所述超级块内的地址映射信息；所述地址映射信息保存了物理扇区地址到逻辑扇区地址的映射关系；当写入数据到所述超级块时，将所述数据相应的地址映射信息也写入到所述超级块中；系统恢复时，将根据各个所述超级块中的地址映射信息恢复所述存储系统的页表；所述页表标识了逻辑扇区地址到物理扇区地址的映射关系。
可选地，当系统恢复时，将根据各个所述超级块中的地址映射信息逆转以恢复所述存储系统的页表。
可选地，所述存储所述超级块相关元数据的空间紧接着所述超级块的数据存储空间。
可选地，所述超级块相关元数据还包括所述超级块的擦除次数和/或所述超级块的至少一种统计数据。
可选地，当不同的超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为序列号分配在后的超级块的地址映射信息中对应的物理扇区地址。
可选地，当同一超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为所述超级块的后写入的地址映射信息中对应的物理扇区地址。
可选地，所述存储设备元数据管理方法还包括：系统恢复时，针对未完成写入的超级块，通过对所述未完成写入的超级块进行完全扫描以读出逻辑地址与物理地址的映射关系并恢复所述未完成写入的超级块中的页表。
可选地，所述元数据中的静态数据和/或坏块信息存储于所述存储设备的起始擦除块中。
本发明提供一种存储设备元数据管理系统，所述存储设备元数据管理系统包括：超级块管理模块，用于将至少一个擦除块组成一个超级块；所述超级块保留有一个存储所述超级块相关元数据的空间；所述超级块相关元数据包括为所述超级块分配的序列号以及所述超级块内的地址映射信息；所述地址映射信息保存了物理扇区地址到逻辑扇区地址的映射关系；当写入数据到所述超级块时，将所述数据相应的地址映射信息也写入到所述超级块中；系统恢复模块，用于在系统恢复时，将根据各个所述超级块中的地址映射信息恢复所述存储系统的 页表；所述页表标识了逻辑扇区地址到物理扇区地址的映射关系。
可选地，所述系统将在系统恢复时将各个所述超级块中的地址映射信息逆转以恢复所述存储系统的页表。
可选地，所述存储所述超级块相关元数据的空间紧接着所述超级块的数据存储空间。
可选地，所述超级块相关元数据还包括所述超级块的擦除次数和/或所述超级块的至少一种统计数据。
可选地，在系统恢复时，当不同的超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为序列号分配在后的超级块的地址映射信息中对应的物理扇区地址。
可选地，在系统恢复时，当同一超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为所述超级块的后写入的地址映射信息中对应的物理扇区地址。
可选地，所述系统恢复模块还用于针对未完成写入的超级块，通过对所述未完成写入的超级块进行完全扫描以读出逻辑地址与物理地址的映射关系并恢复所述未完成写入的超级块中的页表。
本发明再提供一种存储设备，所述存储设备包括如上所述的存储设备元数据管理系统。
可选地，所述存储设备为基于闪存的固态硬盘。
如上所述，本发明的存储设备元数据管理方法及系统，具有以下有益效果：兼顾了容错性，性能代价和实现复杂度等诸多方面，能够保证页表在开机时完全恢复，保证SSD启动时可以正常工作。
本发明能够满足以下要求：1，不受动态产生的坏块影响。2，不受突然断电，系统崩溃的影响。3，即使由于极端故障条件的出现，某些数据无法恢复，受到的影响也应是受限的。4，开机时间尽可能短。5，关机时间尽可能短。6，对正常工作的性能影响尽可能小。7，空间需求尽可能小。并且采用本发明的技术方案能达到以下性能：1，关机时间小于10ms，至多每一个闪存芯片须写入两页。2，开机时至多需要读取总容量的千分之一的数据。3，占用闪存总容量的千分之一。4，正常工作下对性能基本无影响。
附图说明
图1显示为本发明的存储设备元数据管理方法的一实施例的流程示意图。
图2显示为本发明的存储设备元数据管理系统的一实施例的模块示意图。
图3显示为本发明的存储设备的一实施例的结构示意图。
图4显示为本发明的存储设备的一实施例的结构示意图。
元件标号说明
1                   存储设备元数据管理系统
11                  超级块管理模块
12                  系统恢复模块
S1～S3              步骤
具体实施方式
以下通过特定的具体实例说明本发明的实施方式，本领域技术人员可由本说明书所揭露的内容轻易地了解本发明的其他优点与功效。本发明还可以通过另外不同的具体实施方式加以实施或应用，本说明书中的各项细节也可以基于不同观点与应用，在没有背离本发明的精神下进行各种修饰或改变。
需要说明的是，本实施例中所提供的图示仅以示意方式说明本发明的基本构想，遂图式中仅显示与本发明中有关的组件而非按照实际实施时的组件数目、形状及尺寸绘制，其实际实施时各组件的型态、数量及比例可为一种随意的改变，且其组件布局型态也可能更为复杂。
本发明提供一种存储设备元数据管理方法。在一个实施例中，如图1所示，所述存储设备元数据管理方法包括：
步骤S1，由至少一个擦除块组成一个超级块；所述超级块保留有一个存储所述超级块相关元数据的空间；所述超级块相关元数据包括为所述超级块分配的序列号以及所述超级块内的地址映射信息；所述地址映射信息保存了物理扇区地址到逻辑扇区地址的映射关系。具体地，存储设备的元数据包括：存储设备(如SSD，Solid State Drive，固态硬盘)的基本格式，包括使用的闪存(Flash)种类，数量和组织结构等，这部分数据除非重新格式化，否则不变，也可以称作静态数据；坏块信息，哪些块是坏块，即不可使用的块，随着使用时间的推移，坏块可能会增加，但添加速度很慢；逻辑扇区地址(LBA，Logical Block Address，也称作逻辑区块地址)如何映射到物理扇区地址(PBA，Physical Block Address，也称作物理区块地址)；如果同一个LBA对应有多个PBA，那么哪一个对应最新版本；每一擦除块的擦除次数；其他需要保留的信息，如各种统计等。所述存储所述超级块相关元数据的空间紧接着所述超级块的数据存储空间。因此，我们可以把所述存储所述超级块相关元数据的空间称作后记(Epilog)。在一个实施例中，所述元数据中的静态数据和/或坏块信息存储于所述存储设备 的起始擦除块中。所述坏块信息可以为坏块表。在一个实施例中，所述超级块相关元数据还包括所述超级块的擦除次数和/或所述超级块的至少一种统计数据。
步骤S2，当写入数据到所述超级块时，将所述数据相应的地址映射信息也写入到所述超级块中。具体地，当写入数据到某个超级块时，紧接着写入数据将所写入的数据相应的地址映射信息也写入到所述超级块中。系统通过页表来访问数据，页表中标识的是逻辑扇区地址到物理扇区地址的映射关系，但写入的地址映射信息数据与所述页表中的数据是逆向的，即地址映射信息中的数据标识的是物理扇区地址到逻辑扇区地址的映射关系。采用本方法，相当于将部分元数据信息(包括整个页表信息)分布式地存储在各个超级块中。在一个实施例中，当写入数据到某个超级块后，还将各种统计信息写入到所述超级块的存储所述超级块相关元数据的空间中。
步骤S3，系统恢复时，将各个所述超级块中的地址映射信息逆转以恢复所述存储系统的页表；所述页表标识了逻辑扇区地址到物理扇区地址的映射关系。由于每一个超级块内的小页表仅包含本超级块内的映射，而且此映射是逆向的，即PBA->LBA，我们需要将其翻转才得到LBA->PBA的映射。这仍是一个O(n)的算法。PBA到LBA的映射有可能是多对一，即同一个LBA有可能对应多个PBA，其中只有一个是最终的版本。如果为每一个PBA写入分配和记录序列号的话，代价比较大。在一个实施例中，采用的方法是为每一个超级块分配一个序列号，并记录在后记中。依赖于此字段，我们很容易就可以知道任两个PBA谁先谁后：一，当不同的超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为序列号分配在后的超级块的地址映射信息中对应的物理扇区地址。即不同超级块之间，新块数据新，旧块数据旧。二，当同一超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为所述超级块的后写入的地址映射信息中对应的物理扇区地址。即同一超级块内前面写入的PBA旧，后面写入的新。
在一个实施例中，所述存储设备元数据管理方法还包括：在系统恢复时，针对未完成写入的超级块，通过对所述未完成写入的超级块进行完全扫描以读出逻辑地址与物理地址的映射关系并恢复所述未完成写入的超级块中的页表。SSD关机的时候有可能有一个或多个超级块未包含完全的后记，我们称此超级块是未封口的。针对未完成写入的超级块，其后记不可用，我们需要从头扫描此超级块。每一个PBA在写入的时候，它所包含的LBA也在写入的数据内，所以我们仍然可以得到PBA->LBA映射。由于未封口超级块的个数严格受限，所以这不影响全算法的时间复杂度。所述存储设备元数据管理方法可以应用在SSD中，还可以运用到 传统基于硬盘的阵列，或下一代存储介质的阵列上。
本发明提供一种存储设备元数据管理系统。在一个实施例中，如图2所示，所述存储设备元数据管理系统1包括超级块管理模块11和系统恢复模块12。其中：
超级块管理模块11用于将至少一个擦除块组成一个超级块；所述超级块保留有一个存储所述超级块相关元数据的空间；所述超级块相关元数据包括为所述超级块分配的序列号以及所述超级块内的地址映射信息；所述地址映射信息保存了物理扇区地址到逻辑扇区地址的映射关系；当写入数据到所述超级块时，将所述数据相应的地址映射信息也写入到所述超级块中。具体地，存储设备的元数据包括：存储设备(如SSD，Solid State Drive，固态硬盘)的基本格式，包括使用的闪存(Flash)种类，数量和组织结构等，这部分数据除非重新格式化，否则不变，也可以称作静态数据；坏块信息，哪些块是坏块，即不可使用的块，随着使用时间的推移，坏块可能会增加，但添加速度很慢；逻辑扇区地址(LBA，Logical Block Address，也称作逻辑区块地址)如何映射到物理扇区地址(PBA，Physical Block Address，也称作物理区块地址)；如果同一个LBA对应有多个PBA，那么哪一个对应最新版本；每一物理块的擦除次数；其他需要保留的信息，如各种统计等。所述存储所述超级块相关元数据的空间紧接着所述超级块的数据存储空间。因此，我们可以把所述存储所述超级块相关元数据的空间称作后记(Epilog)。在一个实施例中，所述元数据中的静态数据和/或坏块信息存储于所述存储设备的起始擦除块中。所述坏块信息可以为坏块表。在一个实施例中，当写入数据到某个超级块时，超级块管理模块11紧接着写入数据将所写入的数据相应的地址映射信息也写入到所述超级块中。系统通过页表来访问数据，页表中标识的是逻辑扇区地址到物理扇区地址的映射关系，但写入的地址映射信息数据与所述页表中的数据是逆向的，即地址映射信息中的数据标识的是物理扇区地址到逻辑扇区地址的映射关系。采用本系统，相当于将部分元数据信息(包括整个页表信息)分布式地存储在各个超级块中。在一个实施例中，超级块管理模块11所涉及的所述超级块相关元数据还包括所述超级块的擦除次数和/或所述超级块的各种统计数据。当写入数据到某个超级块后，超级块管理模块11还将至少一种统计信息写入到所述超级块的存储所述超级块相关元数据的空间中。
系统恢复模块12用于在系统恢复时，将各个所述超级块中的地址映射信息逆转以恢复所述存储系统的页表；所述页表标识了逻辑扇区地址到物理扇区地址的映射关系；当不同的超级块的地址映射信息中存在相同的逻辑扇区地址时，所述逻辑扇区地址所对应的物理扇区地址为序列号分配在后的超级块的地址映射信息中对应的物理扇区地址。具体地，由于每一个超级块内的小页表仅包含本超级块内的映射，而且此映射是逆向的，即PBA->LBA，我们需 要将其翻转才得到LBA->PBA的映射。这仍是一个O(n)的算法。在一个实施例中，PBA到LBA的映射有可能是多对一，即同一个LBA有可能对应多个PBA，其中只有一个是最终的版本。如果为每一个PBA写入分配和记录序列号的话，代价比较大。系统恢复模块12的方法是为每一个超级块分配一个序列号，并记录在后记中。依赖于此字段，系统恢复模块12很容易就可以知道任两个PBA谁先谁后：一，当不同的超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为序列号分配在后的超级块的地址映射信息中对应的物理扇区地址。即不同超级块之间，新块数据新，旧块数据旧。二，当同一超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为所述超级块的后写入的地址映射信息中对应的物理扇区地址。即同一超级块内前面的PBA旧，后面的新。在一个实施例中，所述系统恢复模块12还包括：在系统恢复时，针对未完成写入的超级块，通过对所述未完成写入的超级块进行完全扫描以读出逻辑地址与物理地址的映射关系并恢复所述未完成写入的超级块中的页表。所述存储设备元数据管理系统可以应用在SSD中，还可以运用到传统基于硬盘的阵列，或下一代存储介质的阵列上。
一种存储设备，所述存储设备包括如上所述的存储设备元数据管理系统1。在一个实施例中，所述存储设备对于元数据中的静态数据，仅在格式化的时候写入到前几个擦除块内，这部分可以称为主引导分区(MBR，Master Boot Record)。对于坏块信息，我们在坏块出现的时候顺序写入到MBR的剩余空间内。前几个擦除块之后，来自于所述至少一个闪存芯片的擦除块组成一个超级块，如图3所示，其中LUN(Logic Unit Number)为单独操作的闪存单元，SB(Super Block)_为一超级块。每一个超级块内部的组织结构如图4所示，在这一实例中一个超级块由6个来自于6个LUN的擦除块组成，其中一个LUN用于存放该超级块的奇偶校验码。在这一超级块内，来自于每一个LUN的同一个页面(Page)形成一个奇偶条带(Stripe)。图4中，Epilog为后记，即存储所述超级块相关元数据的空间。在一个实施例中，所述存储设备包括固态硬盘(SSD，Solid State Drive)。
综上所述，本发明的一种存储设备元数据管理方法及系统，具有以下有益效果：兼顾了容错性，性能代价和实现复杂度等诸多方面，能够保证页表在开机时完全恢复，保证存储设备启动时可以正常工作。本发明能够满足以下要求：1，不受动态产生的坏块影响。2，不受突然断电，系统崩溃的影响。3，就算由于极端故障条件的出现，某些数据无法恢复，受到的影响也应是受限的。4，开机时间尽可能短。5，关机时间尽可能短。6，对正常工作的性能影响尽可能小。7，空间需求尽可能小。并且采用本发明的技术方案能达到以下性能：1，关机 时间很小，至多每一个闪存芯片须写入两个页面。2，开机时至多需要读取的元数据总容量的千分之一的数据。3，元数据存储占用存储总容量的约千分之一。4，正常工作下对性能基本无影响。所以，本发明有效克服了现有技术中的种种缺点而具高度产业利用价值。
上述实施例仅例示性说明本发明的原理及其功效，而非用于限制本发明。任何熟悉此技术的人士皆可在不违背本发明的精神及范畴下，对上述实施例进行修饰或改变。因此，举凡所属技术领域中具有通常知识者在未脱离本发明所揭示的精神与技术思想下所完成的一切等效修饰或改变，仍应由本发明的权利要求所涵盖。

Claims (17)

  1. 一种存储设备元数据管理方法，其特征在于，所述存储设备元数据管理方法包括：
    由至少一个擦除块组成一个超级块；所述超级块保留有一个存储所述超级块相关元数据的空间；所述超级块相关元数据包括为所述超级块分配的序列号以及所述超级块内的地址映射信息；所述地址映射信息保存了物理扇区地址到逻辑扇区地址的映射关系；
    当写入数据到所述超级块时，将所述数据相应的地址映射信息也写入到所述超级块中；
    系统恢复时，将根据各个所述超级块中的地址映射信息恢复所述存储系统的页表；所述页表标识了逻辑扇区地址到物理扇区地址的映射关系。
  2. 根据权利要求1所述的存储设备元数据管理方法，其特征在于在系统恢复时，将各个所述超级块中的地址映射信息逆转以恢复所述存储系统的逻辑扇区到物理扇区的映射关系。
  3. 根据权利要求1所述的存储设备元数据管理方法，其特征在于：所述存储所述超级块相关元数据的空间紧接着所述超级块的数据存储空间。
  4. 根据权利要求1所述的存储设备元数据管理方法，其特征在于：所述超级块相关元数据还包括所述超级块的擦除次数和/或所述超级块的至少一种统计数据。
  5. 根据权利要求1所述的存储设备元数据管理方法，其特征在于：当不同的超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为序列号分配在后的超级块的地址映射信息中对应的物理扇区地址。
  6. 根据权利要求1所述的存储设备元数据管理方法，其特征在于：当同一超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为所述超级块的后写入的地址映射信息中对应的物理扇区地址。
  7. 根据权利要求1所述的存储设备元数据管理方法，其特征在于：所述存储设备元数据管理方法还包括：系统恢复时，针对未完成写入的超级块，通过对所述未完成写入的超级块进行完全扫描以读出逻辑地址与物理地址的映射关系并恢复所述未完成写入的超级块中的 页表。
  8. 根据权利要求1所述的存储设备元数据管理方法，其特征在于：所述元数据中的静态数据和/或坏块信息存储于所述存储设备的起始擦除块中。
  9. 一种存储设备元数据管理系统，其特征在于：所述存储设备元数据管理系统包括：超级块管理模块，用于将至少一个擦除块组成一个超级块；所述超级块保留有一个存储所述超级块相关元数据的空间；所述超级块相关元数据包括为所述超级块分配的序列号以及所述超级块内的地址映射信息；所述地址映射信息保存了物理扇区地址到逻辑扇区地址的映射关系；当写入数据到所述超级块时，将所述数据相应的地址映射信息也写入到所述超级块中；系统恢复模块，用于在系统恢复时，将根据各个所述超级块中的地址映射信息恢复所述存储系统的页表；所述页表标识了逻辑扇区地址到物理扇区地址的映射关系。
  10. 根据权利要求9所述的存储设备元数据管理系统，其特征在于在系统恢复时，将各个所述超级块中的地址映射信息逆转以恢复所述存储系统的逻辑扇区到物理扇区的映射关系。
  11. 根据权利要求9所述的存储设备元数据管理系统，其特征在于：所述存储所述超级块相关元数据的空间紧接着所述超级块的数据存储空间。
  12. 根据权利要求9所述的存储设备元数据管理系统，其特征在于：所述超级块相关元数据还包括所述超级块的擦除次数和/或所述超级块的至少一种统计数据。
  13. 根据权利要求9所述的存储设备元数据管理系统，其特征在于：在系统恢复时，当不同的超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为序列号分配在后的超级块的地址映射信息中对应的物理扇区地址。
  14. 根据权利要求9所述的存储设备元数据管理系统，其特征在于：在系统恢复时，当同一超级块的地址映射信息中存在相同的逻辑扇区地址时，所述页表中所述逻辑扇区地址所对应的物理扇区地址为所述超级块的后写入的地址映射信息中对应的物理扇区地址。
  15. 根据权利要求9所述的存储设备元数据管理系统，其特征在于：所述系统恢复模块还用于针对未完成写入的超级块，通过对所述未完成写入的超级块进行完全扫描以读出逻辑地址与物理地址的映射关系并恢复所述未完成写入的超级块中的页表。
  16. 一种存储设备，其特征在于：所述存储设备包括如权利要求9至15中任一项所述的存储设备元数据管理系统。
  17. 根据权利要求16所述的存储设备，其特征在于：所述存储设备为基于闪存的固态硬盘。
PCT/CN2015/096496 2014-12-05 2015-12-04 一种存储设备元数据管理处理方法及系统 WO2016086899A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/528,839 US10445231B2 (en) 2014-12-05 2015-12-04 Method and system for storage device metadata management and processing

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410737270.0 2014-12-05
CN201410737270.0A CN104461391B (zh) 2014-12-05 2014-12-05 一种存储设备元数据管理处理方法及系统

Publications (1)

Publication Number Publication Date
WO2016086899A1 true WO2016086899A1 (zh) 2016-06-09

Family

ID=52907520

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/096496 WO2016086899A1 (zh) 2014-12-05 2015-12-04 一种存储设备元数据管理处理方法及系统

Country Status (3)

Country Link
US (1) US10445231B2 (zh)
CN (1) CN104461391B (zh)
WO (1) WO2016086899A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019089838A1 (en) * 2017-10-31 2019-05-09 Pure Storage, Inc. Overlapping raid groups
CN114691418A (zh) * 2022-05-30 2022-07-01 阿里巴巴（中国）有限公司 存储设备的信息恢复方法、装置、电子设备及存储介质
CN116880781A (zh) * 2023-09-08 2023-10-13 合肥康芯威存储技术有限公司 一种存储设备及其控制方法

Families Citing this family (28)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104461391B (zh) * 2014-12-05 2019-08-02 上海宝存信息科技有限公司 一种存储设备元数据管理处理方法及系统
CN106339179B (zh) * 2015-07-06 2020-11-17 上海宝存信息科技有限公司 主机装置、存取系统、以及存取方法
US10191841B2 (en) 2015-07-06 2019-01-29 Shannon Systems Ltd. Host device, access system, and access method
CN105808378B (zh) 2016-02-26 2018-10-30 华为技术有限公司 元数据修复方法及装置
KR102529696B1 (ko) * 2016-07-14 2023-05-10 에스케이하이닉스 주식회사 메모리 시스템 및 메모리 시스템의 동작 방법
CN107301073B (zh) * 2017-06-23 2021-06-15 郑州云海信息技术有限公司 一种基于固态硬盘系统的配置信息检索方法及其装置
CN107329909B (zh) * 2017-06-27 2020-07-07 郑州云海信息技术有限公司 一种数据管理方法及装置
US11204841B2 (en) * 2018-04-06 2021-12-21 Micron Technology, Inc. Meta data protection against unexpected power loss in a memory system
KR102620727B1 (ko) * 2018-04-18 2024-01-03 에스케이하이닉스 주식회사 전자 장치
CN108716916B (zh) * 2018-05-31 2021-05-25 北京航空航天大学 一种基于超级块的分布式并行星点质心提取方法及fpga实现装置
KR102611566B1 (ko) * 2018-07-06 2023-12-07 삼성전자주식회사 솔리드 스테이트 드라이브 및 그의 메모리 할당 방법
CN109101437A (zh) * 2018-07-20 2018-12-28 浪潮电子信息产业股份有限公司 一种数据存储方法与终端
CN109086222B (zh) * 2018-07-24 2023-08-25 浪潮电子信息产业股份有限公司 固态硬盘的数据恢复方法以及固态硬盘
US10901889B2 (en) * 2018-07-25 2021-01-26 ScaleFlux, Inc. Using hybrid-software/hardware based logical-to-physical address mapping to improve the data write throughput of solid-state data storage devices
CN109324930A (zh) * 2018-09-28 2019-02-12 深圳忆联信息系统有限公司 一种固态硬盘快速引导数据的处理方法及装置
CN109582239B (zh) * 2018-12-03 2022-02-18 郑州云海信息技术有限公司 一种ssd坏块表的保存方法、装置、设备以及存储介质
CN111506255B (zh) * 2019-01-31 2023-09-26 山东存储之翼电子科技有限公司 基于nvm的固态硬盘元数据管理方法及系统
US20200409601A1 (en) * 2019-06-28 2020-12-31 Western Digital Technologies, Inc. Hold of Write Commands in Zoned Namespaces
US11295031B2 (en) 2019-10-08 2022-04-05 International Business Machines Corporation Event log tamper resistance
US11392348B2 (en) * 2020-02-13 2022-07-19 International Business Machines Corporation Ordering records for timed meta-data generation in a blocked record environment
CN111581022B (zh) * 2020-04-30 2022-07-08 江苏芯盛智能科技有限公司 一种数据恢复方法及系统
US10990526B1 (en) * 2020-04-30 2021-04-27 Micron Technology, Inc. Handling asynchronous power loss in a memory sub-system that programs sequentially
CN111752487B (zh) * 2020-06-18 2024-01-12 深圳大普微电子科技有限公司 一种数据恢复方法、装置及固态硬盘
CN111813709B (zh) * 2020-07-21 2023-08-08 北京计算机技术及应用研究所 一种基于fpga存算一体架构的高速并行存储方法
CN113590505B (zh) * 2021-06-30 2024-07-19 深圳大普微电子股份有限公司 地址映射方法、固态硬盘控制器及固态硬盘
CN115048046B (zh) * 2022-05-26 2022-11-25 北京华昱卓程软件有限公司 日志文件系统以及数据管理方法
CN115629720B (zh) * 2022-12-20 2023-07-28 鹏钛存储技术(南京)有限公司 基于闪存为介质的存储设备上的非对称条带化方法
CN117806572A (zh) * 2024-03-01 2024-04-02 合肥康芯威存储技术有限公司 一种存储设备以及内存管理方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1936867A (zh) * 2005-09-25 2007-03-28 深圳市朗科科技有限公司 闪存介质数据管理方法
CN102866956A (zh) * 2012-09-14 2013-01-09 上海宝存信息科技有限公司 基于固态存储介质的数据实时跟踪存储系统及方法
US20140089630A1 (en) * 2012-09-24 2014-03-27 Violin Memory Inc Virtual addressing
CN103744795A (zh) * 2013-12-19 2014-04-23 记忆科技(深圳)有限公司 固态硬盘快速启动的方法及其固态硬盘
CN104461391A (zh) * 2014-12-05 2015-03-25 上海宝存信息科技有限公司 一种存储设备元数据管理处理方法及系统

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080082736A1 (en) * 2004-03-11 2008-04-03 Chow David Q Managing bad blocks in various flash memory cells for electronic data flash card
US8510370B2 (en) * 2008-02-26 2013-08-13 Avid Technology, Inc. Array-based distributed storage system with parity
US8219776B2 (en) * 2009-09-23 2012-07-10 Lsi Corporation Logical-to-physical address translation for solid state disks
US8407449B1 (en) * 2010-02-26 2013-03-26 Western Digital Technologies, Inc. Non-volatile semiconductor memory storing an inverse map for rebuilding a translation table
EP2764516A4 (en) * 2011-10-05 2015-07-08 Lsi Corp SELF-HISTORIZATION AND HIERARCHICAL CONSISTENCY FOR NON-VOLATILE STORES
CN102819496B (zh) * 2012-08-16 2015-02-18 常州新超电子科技有限公司 闪存ftl的地址转换方法
US9671962B2 (en) * 2012-11-30 2017-06-06 Sandisk Technologies Llc Storage control system with data management mechanism of parity and method of operation thereof
CN103176916B (zh) * 2013-03-07 2016-03-09 中国科学院苏州纳米技术与纳米仿生研究所 闪存及闪存的地址转换方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1936867A (zh) * 2005-09-25 2007-03-28 深圳市朗科科技有限公司 闪存介质数据管理方法
CN102866956A (zh) * 2012-09-14 2013-01-09 上海宝存信息科技有限公司 基于固态存储介质的数据实时跟踪存储系统及方法
US20140089630A1 (en) * 2012-09-24 2014-03-27 Violin Memory Inc Virtual addressing
CN103744795A (zh) * 2013-12-19 2014-04-23 记忆科技(深圳)有限公司 固态硬盘快速启动的方法及其固态硬盘
CN104461391A (zh) * 2014-12-05 2015-03-25 上海宝存信息科技有限公司 一种存储设备元数据管理处理方法及系统

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019089838A1 (en) * 2017-10-31 2019-05-09 Pure Storage, Inc. Overlapping raid groups
EP3652649B1 (en) * 2017-10-31 2023-12-06 Pure Storage, Inc. Overlapping raid groups
CN114691418A (zh) * 2022-05-30 2022-07-01 阿里巴巴（中国）有限公司 存储设备的信息恢复方法、装置、电子设备及存储介质
CN116880781A (zh) * 2023-09-08 2023-10-13 合肥康芯威存储技术有限公司 一种存储设备及其控制方法
CN116880781B (zh) * 2023-09-08 2023-12-26 合肥康芯威存储技术有限公司 一种存储设备及其控制方法

Also Published As

Publication number Publication date
US10445231B2 (en) 2019-10-15
CN104461391B (zh) 2019-08-02
US20170344474A1 (en) 2017-11-30
CN104461391A (zh) 2015-03-25

Similar Documents

Publication Publication Date Title
WO2016086899A1 (zh) 一种存储设备元数据管理处理方法及系统
US8909986B2 (en) Data storing method for solid state drive to preserve data integrity after power failure
TWI661307B (zh) 資料儲存裝置、主機裝置、以及資料寫入方法
US8850114B2 (en) Storage array controller for flash-based storage devices
EP2920698B1 (en) Methods and devices for avoiding lower page corruption in data storage devices
TWI716416B (zh) 儲存系統、用於管理儲存媒體的方法以及系統、操作儲存系統的方法、以及非暫存性電腦可讀媒體
US8055873B2 (en) Data writing method for flash memory, and controller and system using the same
US10061710B2 (en) Storage device
US20190102291A1 (en) Data storage device and method for operating non-volatile memory
TW201737094A (zh) 包含動態超區塊之記憶體裝置及相關方法與電子系統
US20130019057A1 (en) Flash disk array and controller
US9213495B2 (en) Controller management of memory array of storage device using magnetic random access memory (MRAM) in a mobile device
JP2016506585A (ja) データストレージのための方法及びシステム
US20120166709A1 (en) File system of flash memory
WO2014074449A2 (en) Wear leveling in flash memory devices with trim commands
KR102585883B1 (ko) 메모리 시스템의 동작 방법 및 메모리 시스템
JPWO2015015611A1 (ja) ストレージシステム及びデータライト方法
US11334272B2 (en) Memory system and operating method thereof
TWI696916B (zh) 資料儲存裝置以及非揮發式記憶體控制方法
CN111737162A (zh) 一种快速恢复数据的固态硬盘使用方法及计算机可读存储介质
CN106844229A (zh) 固态硬盘固件映射表的组织方法、系统及装置
WO2018142622A1 (ja) 計算機
US11113205B2 (en) Die addressing using a reduced size translation table entry
CN109002265B (zh) 一种数据处理的方法以及相关装置
CN111338846A (zh) 一种基于多核恢复l2p表的方法和装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15864484

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 15528839

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15864484

Country of ref document: EP

Kind code of ref document: A1