CN110688345A - Multi-granularity structured space management mechanism of memory file system - Google Patents
Multi-granularity structured space management mechanism of memory file system Download PDFInfo
- Publication number
- CN110688345A CN110688345A CN201910918678.0A CN201910918678A CN110688345A CN 110688345 A CN110688345 A CN 110688345A CN 201910918678 A CN201910918678 A CN 201910918678A CN 110688345 A CN110688345 A CN 110688345A
- Authority
- CN
- China
- Prior art keywords
- block
- granularity
- blocks
- file
- structured
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/11—File system administration, e.g. details of archiving or snapshots
- G06F16/122—File system administration, e.g. details of archiving or snapshots using management policies
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/16—File or folder operations, e.g. details of user interfaces specifically adapted to file systems
- G06F16/162—Delete operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/17—Details of further file system functions
- G06F16/174—Redundancy elimination performed by the file system
- G06F16/1748—De-duplication implemented within the file system, e.g. based on file segments
- G06F16/1752—De-duplication implemented within the file system, e.g. based on file segments based on file chunks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Memory System (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a multi-granularity structural space management mechanism of a memory file system, and relates to the field of computer system software. The invention provides a multi-granularity structured block management method for managing the free storage space of a memory file system, and the method uses a plurality of structured block lists to manage different types of structured blocks, thereby improving the concurrency performance of the file system. In the write operation, a proper structured large block is distributed through a multi-granularity block distribution mechanism, so that the times of space distribution and the times of operation of a file mapping table are reduced, and the write file performance of the memory file system is improved. In the operation of releasing the file, the released space is recovered by the structure of the structured block, so that the times of space recovery are reduced, and the performance of deleting the file by the memory file system is improved.
Description
Technical Field
The invention relates to the field of computer system software, in particular to a multi-granularity structured space management mechanism of a memory file system.
Background
The new Non-Volatile Memory (NVM) has the properties of byte addressing, high storage density, fast read/write speed and Non-volatility. Similar to DRAM, NVM can be directly attached to a memory bus, allowing processes to access directly using CPULoad/Store instructions. Because NVM is non-volatile, NVM can be used as a permanent Storage device in a Memory Class, called Storage Class Memory (SCM). To manage such storage class memory, various memory file systems have been devised, such as PMFS, NOVA, SIMFS. The file systems use the NVM as a storage device of file data, so that software overhead of a general block layer and an operating system cache of the traditional block-oriented device is avoided, data is directly copied between the cache of an application program and the NVM, and better performance improvement is obtained.
The file system manages file data hierarchically, and usually includes 3 subsystems, namely a file data management subsystem, a file metadata management subsystem, and a free storage space management subsystem. The file data management subsystem is responsible for organization and operation of file data, such as reading and writing files. The file metadata management system is responsible for the organization and operation of various metadata. The free memory space management subsystem is responsible for data layout and management of the NVM physical space. Existing free storage management subsystems manage free space in the file system based on simple lists. For example, SIMFS uses a simple List to link all free blocks, PMFS uses a BlockNode List with ordered block numbers to manage allocated pages to achieve the purpose of free space management, NOVA divides the file system free space by the number of CPUs, and manages the free space by each CPU List.
The file corresponds to the Inode in the file system, and the Inode includes attribute information of the file, such as file creation/access/last modification time, file size, file owner, root pointer of file data block mapping table, file access mode, and the like. The data blocks allocated to a file are organized by a file mapping table. Common data structures for file mapping tables include arrays and tree structures, such as binary arrays, B-trees, and indirect block mapping tables. In the process of writing in a file, a file system firstly allocates a new idle block for the file according to the position (offset) of the written file and the size of a written buffer, and then embeds the newly allocated block into a mapping table of the file to expand the size of the file; the data is then copied cyclically from the user-mode Buffer to the space on the NVM where the file is located.
Existing memory file systems use NVM that is close to DRAM access speed as storage device, repeatedly invoke the free storage space management subsystem to allocate free blocks for files in the write process, and embed these blocks into the file mapping table. Further, in the process of deleting a file, it is necessary to repeatedly reclaim blocks one by one. Therefore, larger expenditure is generated, and the performance of writing files and deleting files of the memory file system is reduced.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a multi-granularity structural space management mechanism of a memory file system so as to improve the performance of writing files and deleting files of the memory file system. The invention comprises three parts: a multi-granularity storage space management mechanism, a multi-granularity block allocation mechanism and a multi-granularity block recovery mechanism.
1. The multi-granularity structured space management mechanism manages the free storage space of the memory file system, allocates the storage space for the file and recovers and deletes the space released by the file, and comprises a free storage space initialization module, a structured large block construction module, a structured block allocation module, a structured block recovery module and a dynamic adjustment module for the number of various blocks.
And the free storage space initialization module is responsible for initializing the spatial layout of the file system when the file system is mounted. The initialization module organizes all free storage space in the file system into a basic block list according to the size of basic blocks, then calls a structured large block building module to create structured large blocks with various granularities, and each type of structured large block is managed through the list.
The structured large block building module is responsible for building different types of structured large blocks, and the structured large blocks with different sizes are formed by basic blocks through a file mapping table.
The structured block distribution module provides various block distribution interfaces and is responsible for distributing different types of structured large blocks. The required structured blocks are allocated from the head of the corresponding block list according to the number of current blocks in the super-block.
The structured block recovery module provides a plurality of block recovery interfaces and is responsible for recovering different types of structured blocks. The freed structured chunk is linked to the head of the corresponding chunk list.
The dynamic adjusting module for the number of the various types of blocks is responsible for dynamically adjusting the number of the various types of structural blocks and is realized by a kernel thread. The number of each type of block corresponds to a threshold value in a file system super block, and when the number of the types of the currently allocated blocks is smaller than the threshold value, the kernel thread is started. And the kernel thread allocates a certain number of blocks from the block class with the maximum idle number currently to construct/decompose the required target blocks according to the number of each block type in the super-speed.
2. And the multi-granularity block allocation mechanism is responsible for calculating the number of different types of structural blocks needing to be allocated for the current write operation so as to decide which space allocation interface is called.
The main idea of the multi-granularity block allocation mechanism is to allocate large structural blocks as much as possible so as to reduce the times of allocating storage space and inserting a file mapping table and improve the performance of writing files. And the block allocation mechanism calculates and allocates proper structuralization and the quantity of structuralization according to the position of the written file, the data volume of the written file at this time and the current size of the file.
3. And the multi-granularity block recovery mechanism is responsible for judging the block type of the released space when the file is released, and determining whether the storage space to be released needs to be split or not so as to call the corresponding block recovery interface to recover the storage space.
The multi-granularity block recovery mechanism recovers the mapping table structure of the release space to the multi-granularity storage space management module as a whole as possible, reduces the calling of recovery interfaces, and improves the performance of file release. The multi-granularity block reclamation mechanism first determines whether the size of the file space being freed is aligned with the maximum structured block size. If so, the data is directly recycled according to the maximum structured block. Otherwise, the data is split into a plurality of small structured block recycling.
The invention provides a multi-granularity structural space management mechanism of a memory file system, which flexibly selects proper blocks for space allocation according to the offset of a written file and the size of a written data cache, avoids frequently calling a space allocation module and improves the performance of the written file. In addition, when the file recovery space is deleted, the deleted space is directly added into the corresponding structured block size list according to the size of the deleted file space, and the file deletion performance is improved.
Drawings
FIG. 1 is a schematic diagram of a multi-granular structured space management mechanism
FIG. 2 is a 32KB structural Block schematic
FIG. 3 is a schematic diagram of a multi-granularity block allocation mechanism
FIG. 4 is a flow diagram of a multi-granularity block allocation mechanism
FIG. 5 is a schematic diagram of a multi-granularity block reclamation mechanism
FIG. 6 is a flow diagram of a multi-granularity chunk reclamation mechanism
FIG. 7 is a graph comparing write performance for multi-granular space management when writing different file sizes
Detailed Description
FIG. 1 is a diagram of a multi-granular structured space management mechanism. As shown in FIG. 1, a multi-granular structured space management mechanism manages the free storage space of an organized memory file system. To efficiently allocate and reclaim space, structured blocks of multiple granularity sizes are provided. The structured block is formed by a plurality of basic blocks (such as 4KB in size) through a tree-structured file mapping table and is created by a structured block construction module. The multi-granularity structural space management mechanism manages the structural blocks of each category through a simple list, provides corresponding distribution and recovery interfaces for the structural blocks of different categories, and effectively improves the concurrency of space distribution and recovery. The initialization process of the multi-granularity structured space management mechanism is as follows:
s1, in the process of mounting the internal storage file system, the number of basic 4KB blocks is calculated according to the mounting size of the file system, and the number of the basic blocks is recorded into a super block of the file system. Free storage space of the file system is managed in a list of basic blocks of size 4KB, the head pointers of the basic block list being recorded into the super block.
S2 calls the structured chunk construction module, constructs the structured chunk using the basic chunk, and manages the structured chunk of the category with the list data structure. The number of large blocks and the head pointer of the list are recorded into the super block.
S3 initializes other classes of structured chunks.
The background kernel thread in fig. 1 is responsible for dynamically balancing the number of granularity blocks. Each category structured block list corresponds to a block quantity threshold in the super block. After the system runs for a period of time, the number of different kinds of free structured blocks in the file system is different due to the randomness of the size of the written file. In order to obtain the required structured blocks during space allocation, the background kernel thread is triggered to execute during space allocation and release each time, and whether the number of the current idle structured blocks is smaller than a threshold value is judged. And if the number of the current free blocks is less than the threshold value, scanning the current free number of other structured blocks, acquiring the structured blocks from the block list with the maximum number to construct or decompose the structured blocks into the current structured blocks, and balancing the number of the structured blocks with various granularities. If no list is found that is greater than the block number threshold, a lack of space cue is returned.
The structured block allocation module provides a variety of block allocation interfaces to allocate different classes of structured chunks. When applying for space allocation, the structured block allocation module obtains the head pointer of the structured block list from the super block, assigns the address of the next block to the head pointer, and returns the block pointed by the original head pointer to complete block allocation.
The structured chunk reclamation module provides a variety of chunk reclamation interfaces to reclaim different classes of structured chunks. When the space is recovered, the structured block recovery module acquires a head pointer of the structured block list from the super block, inserts the recovered block into the head of the list, modifies the head pointer to point to the recovered block, and then modifies the pointer of the recovered block to point to the original first free block.
FIG. 2 is a diagram of a 32KB structured block, with a 32KB structured block being composed of 8 basic blocks, with 8 basic blocks being managed by 8 consecutive pointers. One basic block can manage 64 32KB structured blocks. Since these blocks are unused free space, the address of the next block is stored with the previous block to manage all blocks through the list.
Fig. 3 is a schematic diagram of a multi-granularity block allocation mechanism. The multi-granularity block allocation mechanism determines which structured block is allocated to perform space expansion of the file according to the current size of the written file, the file offset and the size of the written Buffer. The multi-granularity block allocation mechanism is a pre-allocation mechanism that pre-allocates a segment of space for a file through structured chunks according to a file write pattern.
FIG. 4 is a flow diagram of a multi-granularity chunk allocation mechanism, the detailed steps are as follows:
the S1 application calls the write file system to call write data to the file, with the write file request being transmitted to the VFS write function.
S2 VFS calls the write operation of the bottom-layer memory file system, and the memory file system write operation searches the maximum structural block meeting allocation according to the current size of the file, the size of the write buffer and the circulation.
S3, to avoid the occurrence of holes in the file allocated space, it is determined whether the current file size is aligned with the structured chunk. If not, the small-granularity blocks are allocated to align the size of the file space with the size of the large structured block, and then the required large structured block is calculated according to the Buffer size.
S4 if the file size is aligned with the selected structured size, then the number of structured chunks that need to be allocated is calculated directly from the Buffer size.
S5 calls the corresponding structure block distribution interface, distributes a certain quantity of structure blocks and inserts the structure blocks into the mapping table of the file.
Fig. 5 is a schematic diagram of a multi-granularity block recycling mechanism, where the multi-granularity block recycling mechanism determines which type of structured large block the storage space belongs to according to the size of the storage space of the released file, and determines whether the storage space needs to be split. As shown in FIG. 5, the file system frees up a storage space of size 1023MB1020 KB. The multi-granularity block reclamation mechanism first divides the freed space into 511 structured large blocks of 2MB and invokes the corresponding reclamation interfaces to reclaim the blocks. The last 1M1020KB is then divided into 1MB, 1 512KB, 1 256KB, … …, 1 32KB structured block, and 7 4KB basic blocks, respectively. The multi-granularity recovery mechanism effectively reduces the times of calling the space recovery module and improves the performance of deleting files.
FIG. 6 is a flow diagram of a multi-granularity chunk reclamation mechanism, the detailed steps of which are as follows:
the S1 application calls a release file system call and the release file request is passed to the VFS release function.
S2 VFS calls file releasing operation of the bottom memory file system, and the memory file system releasing operation searches the largest block smaller than the size of the released file circularly according to the size of the file.
S3 if the size of the released file is aligned with the searched maximum block size, the number of blocks corresponding to the released space is obtained by dividing the file size by the maximum block size. Otherwise, the released space is divided into a smaller structured block, and then the remaining space is compared with the smaller structured block to determine whether the remaining space is aligned with the smaller structured block until the base block is searched. Num [ i ] holds the number of blocks corresponding to each class.
S4 calls the corresponding block recycle interface according to Num [ i ].
FIG. 7 is a graph comparing write performance for multi-granular structured space management when writing different file sizes. As shown in the figure, the Multi-granularity Space management mechanism is represented by MSMS (Multi-granularity Space management Scheme), and promotes the write performance of the memory file system by 23.25% on average. In addition, the influence of different written file data volume sizes on the writing performance is compared in experiments, and the performance is gradually increased when the written file is larger.
The protection of the present invention is not limited to the above embodiments. Those skilled in the art will recognize that changes and advantages may be made therein without departing from the spirit and scope of the inventive concept, which is defined by the appended claims.
Claims (8)
1. A multi-granularity structural space management mechanism of a memory file system at least comprises four functional modules, an allocation mechanism and a recovery mechanism. A multi-granularity structural space management mechanism, which creates structural blocks with various granularities through a structural construction module; the structured block is formed by a plurality of basic blocks in a file mapping table structure; the blocks of each category are managed by a multi-granularity structured block list; providing a plurality of categories of structured block allocation and reclamation modules to allocate reclamation blocks; and dynamically adjusting the storage space of each category of block according to the number of each category of block. In the file writing process, a multi-granularity block allocation mechanism is realized, and the required target block number is calculated. In the file release process, a multi-granularity block recovery mechanism is realized, and a corresponding block recovery interface is called.
2. The mechanism as claimed in claim 1, wherein the large structured block is composed of basic blocks and a file mapping table.
3. The mechanism as claimed in claim 1, wherein the multi-granularity structural space management mechanism of the memory file system is characterized by a multi-granularity structural block list, each granularity block list manages a corresponding structural block, and the block list manages a plurality of blocks by using an address of a previous free structural block to store a next free structural block.
4. The mechanism as claimed in claim 1, wherein the multi-granularity structural space management module allocates the structural blocks from the corresponding block list.
5. The mechanism as claimed in claim 1, wherein the multi-granularity structure block recycling module is configured to recycle the structure blocks to the corresponding block list.
6. The mechanism as claimed in claim 1, wherein the module for dynamically adjusting the number of each type of block balances the storage space of each type of block, compares the current block number with a threshold, and constructs or decomposes the block into the required structural blocks by other blocks.
7. The mechanism of claim 1, wherein the multi-granularity block allocation mechanism calculates the number of blocks to be allocated and the number of blocks to be allocated according to the size of the written data, the size of the file, and the writing location, and calls the corresponding block allocation module.
8. The mechanism according to claim 1, wherein the multi-granularity block recycling mechanism calculates which type of structured block the file-released space corresponds to, and calls the corresponding structured block recycling module.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910918678.0A CN110688345A (en) | 2019-09-26 | 2019-09-26 | Multi-granularity structured space management mechanism of memory file system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910918678.0A CN110688345A (en) | 2019-09-26 | 2019-09-26 | Multi-granularity structured space management mechanism of memory file system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110688345A true CN110688345A (en) | 2020-01-14 |
Family
ID=69110446
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910918678.0A Pending CN110688345A (en) | 2019-09-26 | 2019-09-26 | Multi-granularity structured space management mechanism of memory file system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110688345A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111240599A (en) * | 2020-01-17 | 2020-06-05 | 北京马赫谷科技有限公司 | Data stream storage method and device |
CN111309258A (en) * | 2020-02-14 | 2020-06-19 | 苏州浪潮智能科技有限公司 | B + tree access method and device and computer readable storage medium |
CN111339053A (en) * | 2020-03-17 | 2020-06-26 | 卡斯柯信号有限公司 | RAM file system and access method of RAM file system |
CN111597124A (en) * | 2020-04-21 | 2020-08-28 | 重庆大学 | Persistent memory file system data organization method, system and storage medium |
CN111694765A (en) * | 2020-06-05 | 2020-09-22 | 重庆大学 | Mobile application feature-oriented multi-granularity space management method for nonvolatile memory file system |
CN112804618A (en) * | 2021-01-06 | 2021-05-14 | 昆腾微电子股份有限公司 | Audio output method and device |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102024034A (en) * | 2010-11-26 | 2011-04-20 | 中国科学院声学研究所 | Fragment processing method for high-definition media-oriented embedded file system |
CN103914265A (en) * | 2014-04-09 | 2014-07-09 | 江苏物联网研究发展中心 | Cluster fine-grained memory management method |
CN104461914A (en) * | 2014-11-10 | 2015-03-25 | 浪潮电子信息产业股份有限公司 | Automatic simplified-configured self-adaptation optimization method |
CN105930356A (en) * | 2016-04-08 | 2016-09-07 | 上海交通大学 | Method for implementing log type heterogeneous hybrid memory file system |
CN106951376A (en) * | 2017-02-27 | 2017-07-14 | 清华大学 | Towards the fine granularity nonvolatile cache system and buffer memory management method of flash memory |
CN107016100A (en) * | 2017-04-10 | 2017-08-04 | 重庆大学 | A kind of metadata management method based on Nonvolatile memory file system |
CN107533435A (en) * | 2015-12-21 | 2018-01-02 | 华为技术有限公司 | The distribution method and storage device of memory space |
CN109669640A (en) * | 2018-12-24 | 2019-04-23 | 浙江大华技术股份有限公司 | A kind of date storage method, device, electronic equipment and medium |
-
2019
- 2019-09-26 CN CN201910918678.0A patent/CN110688345A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102024034A (en) * | 2010-11-26 | 2011-04-20 | 中国科学院声学研究所 | Fragment processing method for high-definition media-oriented embedded file system |
CN103914265A (en) * | 2014-04-09 | 2014-07-09 | 江苏物联网研究发展中心 | Cluster fine-grained memory management method |
CN104461914A (en) * | 2014-11-10 | 2015-03-25 | 浪潮电子信息产业股份有限公司 | Automatic simplified-configured self-adaptation optimization method |
CN107533435A (en) * | 2015-12-21 | 2018-01-02 | 华为技术有限公司 | The distribution method and storage device of memory space |
CN105930356A (en) * | 2016-04-08 | 2016-09-07 | 上海交通大学 | Method for implementing log type heterogeneous hybrid memory file system |
CN106951376A (en) * | 2017-02-27 | 2017-07-14 | 清华大学 | Towards the fine granularity nonvolatile cache system and buffer memory management method of flash memory |
CN107016100A (en) * | 2017-04-10 | 2017-08-04 | 重庆大学 | A kind of metadata management method based on Nonvolatile memory file system |
CN109669640A (en) * | 2018-12-24 | 2019-04-23 | 浙江大华技术股份有限公司 | A kind of date storage method, device, electronic equipment and medium |
Non-Patent Citations (3)
Title |
---|
孙强等: "带磨损均衡的小粒度非易失性内存管理机制", 《计算机科学》 * |
李志国等: "基于知识本体的文本分类技术及其应用研究", 《计算机科学》 * |
陈俊熹等: "混合内存页面管理策略的性能和能耗研究", 《现代计算机(专业版)》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111240599A (en) * | 2020-01-17 | 2020-06-05 | 北京马赫谷科技有限公司 | Data stream storage method and device |
CN111309258A (en) * | 2020-02-14 | 2020-06-19 | 苏州浪潮智能科技有限公司 | B + tree access method and device and computer readable storage medium |
US11762827B2 (en) | 2020-02-14 | 2023-09-19 | Inspur Suzhou Intelligent Technology Co., Ltd. | B-plus tree access method and apparatus, and computer-readable storage medium |
CN111339053A (en) * | 2020-03-17 | 2020-06-26 | 卡斯柯信号有限公司 | RAM file system and access method of RAM file system |
CN111339053B (en) * | 2020-03-17 | 2022-09-30 | 卡斯柯信号有限公司 | RAM file system and access method of RAM file system |
CN111597124A (en) * | 2020-04-21 | 2020-08-28 | 重庆大学 | Persistent memory file system data organization method, system and storage medium |
CN111694765A (en) * | 2020-06-05 | 2020-09-22 | 重庆大学 | Mobile application feature-oriented multi-granularity space management method for nonvolatile memory file system |
CN112804618A (en) * | 2021-01-06 | 2021-05-14 | 昆腾微电子股份有限公司 | Audio output method and device |
CN112804618B (en) * | 2021-01-06 | 2022-07-29 | 昆腾微电子股份有限公司 | Audio output method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110688345A (en) | Multi-granularity structured space management mechanism of memory file system | |
CN109542333B (en) | Memory system and control method for controlling nonvolatile memory | |
CN110678836B (en) | Persistent memory for key value storage | |
CN106354745B (en) | Method for providing an interface of a computer device and computer device | |
CN114115747B (en) | Memory system and control method | |
US9436597B1 (en) | Using non-volatile memory resources to enable a virtual buffer pool for a database application | |
JP6205650B2 (en) | Method and apparatus utilizing non-uniform hash function to place records in non-uniform access memory | |
US9396103B2 (en) | Method and system for storage address re-mapping for a memory device | |
US11150815B2 (en) | Information processing apparatus, information processing method, and computer program product | |
US20060218347A1 (en) | Memory card | |
US20110161621A1 (en) | Micro-update architecture for address tables | |
CN106662985B (en) | Host managed non-volatile memory | |
CN116909941A (en) | Memory system and method for controlling nonvolatile memory | |
CN107066498B (en) | Key value KV storage method and device | |
CN108733306B (en) | File merging method and device | |
US20150324281A1 (en) | System and method of implementing an object storage device on a computer main memory system | |
CN111522507B (en) | Low-delay file system address space management method, system and medium | |
CN110968269A (en) | SCM and SSD-based key value storage system and read-write request processing method | |
Lee et al. | iLSM-SSD: An intelligent LSM-tree based key-value SSD for data analytics | |
US20220035546A1 (en) | Base and compressed difference data deduplication | |
US10528284B2 (en) | Method and apparatus for enabling larger memory capacity than physical memory size | |
CN110659248B (en) | User mode file system design method and system based on nonvolatile memory | |
CN111694765A (en) | Mobile application feature-oriented multi-granularity space management method for nonvolatile memory file system | |
CN113050886B (en) | Nonvolatile memory storage method and system for embedded memory database | |
KR20210012085A (en) | Storage device for improving journal replay, operating method thereof, and electronic device comprising storage device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200114 |