CN106528001B - A kind of caching system based on nonvolatile memory and software RAID - Google Patents
A kind of caching system based on nonvolatile memory and software RAID Download PDFInfo
- Publication number
- CN106528001B CN106528001B CN201611106177.5A CN201611106177A CN106528001B CN 106528001 B CN106528001 B CN 106528001B CN 201611106177 A CN201611106177 A CN 201611106177A CN 106528001 B CN106528001 B CN 106528001B
- Authority
- CN
- China
- Prior art keywords
- raid
- cache
- metadata
- block
- software
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0614—Improving the reliability of storage systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0655—Vertical data movement, i.e. input-output transfer; data movement between one or more hosts and one or more storage devices
- G06F3/0656—Data buffering arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0628—Interfaces specially adapted for storage systems making use of a particular technique
- G06F3/0662—Virtualisation aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0668—Interfaces specially adapted for storage systems adopting a particular infrastructure
- G06F3/0671—In-line storage system
- G06F3/0683—Plurality of storage devices
- G06F3/0689—Disk arrays, e.g. RAID, JBOD
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Memory System Of A Hierarchy Structure (AREA)
Abstract
The present invention proposes a kind of caching system based on nonvolatile memory and software RAID, it is characterized in that a kind of fusion nonvolatile memory Apache Pass, the Novel storage system framework of software RAID and Device Mapper drive module Buaacache.The present invention using the Device Mapper mechanism of Linux come management software RAID, using high-speed equipment Apache Pass as the Cache of low speed RAID, thus building one two-level cache system.There is Persistent Memory media data after power down not lose the characteristic with readwrite performance brilliance.The present invention selects cache of the Persistent Memory as software RAID, under conditions of cost is far below hardware RAID, realizes that software RAID readwrite performance approaches even more than hardware RAID.
Description
Technical field
The present invention relates to technical field of computer data storage more particularly to a kind of novel caching systems, and in particular to arrives
Novel power down nonvolatile memory Persistent Memory and software RAID.
Background technique
With the rise of SSD, bedding storage side that the storage system of mainstream all gradually adopts SSD to mix with mechanical hard disk
Formula.Current storage scheme such as dm-cache, design are based on Linux Device Mapper frame with realization, essence
On be general block device cache module, using the good feature of SSD random read-write performance, using SSD as the caching of disk.With
SSD is compared, and (Persistent Memory, abbreviation PMEM are alternatively referred to as persistent storages to newborn nonvolatile memory
Device), there is High Speed I/O, large storage capacity, power down it is non-volatile and can the excellent properties such as byte addressing, therefore propose a kind of base
In the novel caching system of Persistent Memory and software RAID.
Nonvolatile memory: Persistent Memory, abbreviation PMEM, its main feature is that information is not lost after fully powered-off
It loses.Other than the characteristic that do not lost after guaranteeing media data power down, also have can byte addressing (Byte-addressable) and
The characteristics such as I/O superior performance.At the mode that writes back (write back), such equipment according to material difference, write expend the time,
The reading consuming time approaches DRAM, therefore very attractive as the cache of software RAID.Existing PMEM is ground
Study carefully scheme to focus mostly in hardware art, releases the solution party for combining PMEM and software RAID there has been no manufacturer and research institution at present
Case proposes.
AEP:Apache Pass, Intel Company be will 2017 release server platform " Purley " and research and develop
Persistent Memory of new generation, the caching in caching system as software RAID uses.
Software RAID:Software Redundant Arrays of Independent Disks, referred to as software are independent
Redundant arrays of inexpensive disks, abbreviation software array.Compared with hardware RAID, software RAID does not need the realization of extrapolation RAID card, but will
The bottom of several mechanical hard disks or solid state hard disk etc stores equipment, and polymerization becomes a big virtual RAID in the form of pure software
Equipment.By PMEM as cache, software RAID can be under the premise of cost be far below hardware RAID, and aspect of performance approaches
Even more than the latter.Software is realized in a manner of MD (Multiple Devices) Virtual Block Device at present in linux system
RAID fictionalizes a new Virtual Block Device using the block device of multiple bottoms, and utilizes striping (stripping) skill
Data block is evenly distributed on multiple disks the readwrite performance for improving virtual unit by art.At present under linux, software RAID
Support RAID-0, RAID-1, RAID-4, RAID-5, RAID-6 and RAID-10.
Traditional software RAID scheme bottleneck: the cpu performance and current loading of server are depended critically upon, mode is being write back
Under, disconnection fault will lead to the loss of data in memory and can not restore, and advantage is unobvious compared with hardware RAID.PMEM conduct
The novel storage equipment of hot topic risen in recent years, releases there has been no manufacturer and research institution combine Persistent at present
The solution of Memory and software RAID, therefore present invention selection is not lost and I/O aspect of performance in power down media data
Cache of the brilliant Persistent Memory of energy as software RAID, proposes a kind of new and effective caching system.
Summary of the invention
The purpose of the present invention is being directed to deficiency of the traditional software RAID solution in terms of power loss recovery and readwrite performance,
And for there has been no nonvolatile memories (Persistent Memory, abbreviation PMEM) in conjunction with software RAID at present
Scheme is released, and proposes a kind of caching system based on PMEM and software RAID.
The present invention is based on the Device Mapper Disk Mapping mechanism of Linux, developing kernel drive module Buaacache,
Using high-speed equipment PMEM as the cache of low speed RAID, thus one two-level cache system of building.
Low speed RAID and high speed PEME are created as by kernel-driven module Buaacache of the present invention as a whole
One virtual mixing block device, user can use mixing block device as reading and writing common block device.It is slow in Virtual Block Device
Equipment PMEM and cache object RAID is deposited to invisible per family, with can not directly access per family.User reads and writes this new creation
Virtual Block Device when, actual storing process is cached by Buaacache, if reading and writing hit in direct return cache
Data (or being written directly to cache), otherwise will read and write disk array.
The data space of buffer memory device PMEM of the present invention is divided into metadata space and data space, is connected and is reflected using group
Mode is penetrated, data space is divided into the cache blocks of pre-defined size, cache blocks are further combined into one by fixed number group
A group (Set).The mapping relations one by one of metadata essential record hard disk block and some cache blocks in a determining group, pass through
Distinctive byte addressing (Byte-addressable) mode of PMEM realizes read-write.
The characteristic that the present invention does not lose by media data after PMEM device looses power compensates for tradition and writes back mechanism power down number
According to the defect of loss, the multiple access to disk array is avoided, access time is greatly shortened.Using timely write-back memory member number
According to mode, realize and back up in realtime metadata to buffer memory device during caching system operation, it is live to save system operation.
After the loss of power accident occurs, host can run scene by recovery system from buffer memory device after re-powering, and avoid dirty in caching
Data can not write back to disk because of the corresponding relationship for the disk block for losing and being cached, and realize power loss recovery function.
The present invention selects open source scheme d m-cache to carry out secondary development, since dm-cache is as Linux maturation kernel
A part, it is that storage system plays the part of the role of caching at a slow speed for rear end that one or more fast equipments, which can be used,.However dm-
Cache latest edition only can in 3.10 version of kernel compilation run, kernel version span is excessive, and cause cannot be steady in new version
Determine to operate normally on kernel.In view of there has been no the distribution of new Kernel Porting version at present, therefore the present invention has carried out specific aim transplanting
Work.
Detailed description of the invention
Fig. 1 caching system Linux I/O hierarchy chart；
Fig. 2 Buaacache drive module figure；
Fig. 3 caching system integrated stand composition；
Fig. 4 buffer memory device Apache Pass structure chart；
Fig. 5 caching system mapping mode schematic diagram；
Fig. 6 caching system read request flow diagram；
Fig. 7 caching system write request flow diagram；
Fig. 8 Kernel Porting operation schematic diagram；
The part Fig. 9 bio Kernel Porting schematic diagram；
The part Figure 10 Device Mapper Kernel Porting schematic diagram；
Specific embodiment
To make the object, technical solutions and advantages of the present invention express to be more clearly understood, with reference to the accompanying drawing and specifically
The present invention is further described in more detail for embodiment.
The purpose of the present invention is directed to easy in storage performance and power failure data currently based on the disc array system of software RAID
The deficiency of lose etc. proposes a kind of based on nonvolatile memory (Persistent Memory, abbreviation PMEM) and software
The caching system of RAID.The present invention is based on the Device Mapper mechanism of Linux, developing kernel drive module Buaacache,
The PMEM equipment Apache Pass researched and developed using Intel, the cache as low speed RAID.The software RAID exists
Creation is completed by "-C " or "-create " parameter of mdadm software under Linux4.4 kernel, and is managed by mdadm.
Level of the present invention in Linux I/O system is as shown in Figure 1.As shown in Fig. 2, Buaacache operates in Linux
Under 64 4.4 versions of stable kernel, work the Block Device Driver layer in Linux block device I/O architecture, is registered as
Target drives (Target Driver) under Linux kernel Device Mapper frame are moved in a manner of kernel-driven
State load uses.The drive module calls the dmsetup program in Device Mapper by software diskette array and Apache
Pass is created as a mixing block device as a whole, and user can operate the equipment as reading and writing common block device.
As shown in figure 3, when user executes read-write requests, which is caught by file system system structure of the invention first
It catches, request is issued to general block device layer by file system, is specifically presented, is then proceeded in the form of kernel data structure bio
It is handed down to Mapper layers of Device, after Mapper layers of Device the operation such as is split and recombinates to bio, is finally handed down to
Buaacache drives layer.
Buffer structure of the invention as shown in figure 4, each nonvolatile memory Apache Pass first cache blocks
It can be used as superblock (Super Block), entirely cache relevant key message for storing, can also retain to stay and use it for anything else.Tightly
Then metadata space saves several metadata (Meta Data), and meta-data region is stored for managing data cache block (Data
Block metadata).Data field is divided into the data cache block (Data Block) of several specified sizes, defaults every 8 sectors
A cache blocks are formed, every 512 cache blocks form a group (Set), while the hard disk in software RAID also presses this division.Often
A cache blocks record the initial sector number of corresponding hard disk block, are recorded by metadata structure body, each metadata structure body maintenance
Corresponding state, necessary lock and tag information.Metadata and data block are linearly stored according to sequencing.
Caching mapping mode of the invention is as shown in Figure 5.The present invention is connected mapping mechanism using 512 tunnel groups, is avoided with member
Data are the high fever degree of representative according to generation clustering phenomena.Group be connected mapping can be divided into directly mapping and linear Hash detect two steps
It is rapid to carry out: step 1: directly mapping, group disk block being mapped in buffer memory device；Second: linear Hash detection,
Once finding target group, just the corresponding data block of linear search in organizing；Being connected mapping by group can be by model local on disk
The block of sequential data enclosed is cached in a target group.
Caching read-write flow chart difference of the invention is as shown in Figure 6 and Figure 7, and cache garbage collection mechanism of the present invention is only supported to write back
Mode.When user's read-write requests miss caching when, and all dirty piece of cache blocks without free buffer block when, cache garbage collection machine
System uses Two-Handed clock algorithm, writes back dirty piece had not visited in nearest 20 minutes to disk array, maintains data
Consistency.
Power loss recovery mechanism of the invention as shown in figure 8, the present invention in user's read-write requests miss, once have first number
According to update, buffer memory device is written from memory with byte addressing system in time.After accident power-off accident occurs, the metadata in memory
Backup is in buffer memory device in time, and after host re-powers, Buaacache drive module can will back up in buffer memory device
Metadata be re-loaded in memory, then by the corresponding dirty data of metadata in buffer memory device (because power-off from caching not and
When write back the data of disk array) write back to RAID, realize power loss recovery.
The present invention is based on open source scheme d m-cache to carry out secondary development, as described in summary of the invention, by dm-cache from interior
3.10 version of core is transplanted to kernel 4.4 and supports version for a long time.Work is transplanted to block device read-write requests field bio and Device
Part field in Mapper carries out kernel change tracking, is accordingly adapted to.Bio is one of generic block layer in linux kernel
A kernel data structure is responsible for connection core buffer and block device.
As shown in figure 9, the bio field in 4.4 version of kernel increases bvec_iter structure newly compared with 3.10 version of kernel
Body encapsulates and belongs to tetra- critical fielies of bi_sector in bio, bi_size, bi_idx and bi_bvec_done originally.
As shown in Figure 10, Device Mapper frame provides synchronous or asynchronous I/O service, in kernel 4.4
There are four types of dm_io_mem_type types altogether by dm-io.Compared with 3.10 version of kernel, in 4.4 version of kernel, dm_io_
Memory struct bio*bio substitutes struct bio_vec, and dm_io_memtype substitutes DM_IO_ with DM_IO_BIO
BVEC。
Claims (2)
1. a kind of caching system based on nonvolatile memory and software RAID, it is characterised in that: a kind of fusion is non-volatile
Memory Apache Pass, the novel caching system framework of software RAID and Device Mapper drive module Buaacache；
The nonvolatile memory refers to not to be lost with fully powered-off rear information, byte addressing and I/O superior performance characteristic
Novel storage equipment, the caching in caching system as software RAID uses；
The software RAID refers in (SuSE) Linux OS, is completed using "-C " or "-create " parameter of mdadm software
Creation, is managed by mdadm；The RAID is redundancy RAID or nonredundancy RAID；
The Device Mapper drive module Buaacache is driven based on the kernel under Linux Device Mapper frame
Dynamic model block Buaacache, work the Block Device Driver layer in Linux block device I/O architecture, in logic will be non-volatile
Caching (Cache) of the memory Apache Pass as low speed software RAID, physically by software RAID and non-volatile memories
Device is created as a mixing block device as a whole；Buaacache drive module includes data cached molded tissue block, caching
Mapping block, cache garbage collection module, metadata management module and power loss recovery module；
The data cached molded tissue block, which refers to, is divided into reserved area, meta-data region and data cache block area for Apache Pass,
Reserved area is responsible for storage file system superblock or reserves for other use, and meta-data region is responsible for storing essential record cache blocks and hard disk block
The metadata of corresponding relationship, data cache block area is with cache blocks organizational form storing data；
The disk in Apache Pass and software RAID is divided into according to fixed size if the caching mapping block refers to
Dry group (Set) is connected mapping mechanism using 512 tunnel groups, avoids using metadata being the high fever degree of representative according to clustering phenomena occurs；
Group be connected mapping be divided into directly mapping and linear Hash detect two steps carry out；
The cache garbage collection module refers to using Two-Handed clock algorithm, writes back in Apache Pass nearest 20 minutes not
Dirty piece accessed arrives disk array；
The metadata management module refers to from memory write-in metadata to Apache Pass or loads from Apache Pass
For metadata to memory, addressing system is byte addressing；When from RAID be written caching or from Apache Pass write back dirty piece to
When RAID, progress metadata updates operation；
The power loss recovery module refers to once there is metadata updates, writes back to Apache Pass in time with byte addressing system；
After accident power-off accident occurs, the metadata in memory backs up in buffer memory device in time, after host re-powers,
The metadata backed up in buffer memory device is re-loaded in memory by Buaacache drive module, then will according to metadata
Corresponding dirty data writes back to disk array in buffer memory device, realizes power loss recovery.
2. the caching system according to claim 1 based on nonvolatile memory and software RAID, it is characterised in that: institute
State Device Mapper drive module Buaacache be based on open source scheme d m-cache carry out secondary development, by dm-cache from
3.10 version of kernel is transplanted to kernel 4.4 and supports version for a long time, and specific step of transplanting includes the 1.1st step and the 1.2nd step:
1.1st step: increasing bvec_iter structural body newly in block device read-write requests field bio field, and encapsulation belongs to bio originally
Middle bi_sector, bi_size, tetra- critical fielies of bi_idx and bi_bvec_done；
1.2nd step: in Device Mapper field dm_io_memory and dm_io_memtype, struct is used respectively
Bio*bio substitutes struct bio_vec, and DM_IO_BIO substitutes DM_IO_BVEC.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611106177.5A CN106528001B (en) | 2016-12-05 | 2016-12-05 | A kind of caching system based on nonvolatile memory and software RAID |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611106177.5A CN106528001B (en) | 2016-12-05 | 2016-12-05 | A kind of caching system based on nonvolatile memory and software RAID |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106528001A CN106528001A (en) | 2017-03-22 |
| CN106528001B true CN106528001B (en) | 2019-08-23 |
Family
ID=58341307
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611106177.5A Active CN106528001B (en) | 2016-12-05 | 2016-12-05 | A kind of caching system based on nonvolatile memory and software RAID |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106528001B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20180130140A (en) * | 2017-05-29 | 2018-12-07 | 에스케이하이닉스 주식회사 | Data processing system and data processing method |
| CN107577492A (en) * | 2017-08-10 | 2018-01-12 | 上海交通大学 | The NVM block device drives method and system of accelerating file system read-write |
| CN109933279B (en) * | 2017-12-19 | 2021-01-22 | 中国科学院声学研究所 | Content filtering method supporting hybrid storage system |
| CN108108130A (en) * | 2017-12-22 | 2018-06-01 | 深圳忆联信息系统有限公司 | A kind of method and solid state disk for promoting solid state disk read-write performance |
| CN109086425B (en) * | 2018-08-09 | 2022-05-06 | 北京百度网讯科技有限公司 | Data processing method and device for database |
| CN109240609B (en) * | 2018-08-23 | 2022-02-18 | 郑州云海信息技术有限公司 | Data space read-write method and system based on AEP memory |
| CN110032526B (en) * | 2019-04-16 | 2021-10-15 | 苏州浪潮智能科技有限公司 | Page caching method, system and equipment based on nonvolatile medium |
| CN110532136B (en) * | 2019-08-30 | 2023-05-02 | 四川效率源信息安全技术股份有限公司 | Raid data recovery method based on metadata |
| CN111158599B (en) * | 2019-12-29 | 2022-03-22 | 北京浪潮数据技术有限公司 | Method, device and equipment for writing data and storage medium |
| CN112069091B (en) * | 2020-08-17 | 2023-09-01 | 北京科技大学 | Memory access optimization method and device applied to molecular dynamics simulation software |
| CN113326072B (en) * | 2021-05-24 | 2023-11-24 | 北京计算机技术及应用研究所 | Real-time monitoring method based on nonvolatile memory under Feiteng server platform |
| CN114201115A (en) * | 2021-12-14 | 2022-03-18 | 北京达佳互联信息技术有限公司 | Data storage system, method, computer device and storage medium |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1527973A (en) * | 2000-06-23 | 2004-09-08 | 英特尔公司 | Non-volatile cache |
| CN101236526A (en) * | 2007-01-30 | 2008-08-06 | 扬智科技股份有限公司 | Computer system having cache system directly connected to nonvolatile storage device |
| CN102043731A (en) * | 2010-12-17 | 2011-05-04 | 天津曙光计算机产业有限公司 | Cache system of storage system |
| CN102053929A (en) * | 2010-12-17 | 2011-05-11 | 天津曙光计算机产业有限公司 | IO cache operation method and system based on DM layer of Linux system |
| CN102147802A (en) * | 2010-12-17 | 2011-08-10 | 曙光信息产业(北京)有限公司 | Pseudo-random type NFS application acceleration system |
| CN102203751A (en) * | 2008-10-30 | 2011-09-28 | 惠普开发有限公司 | Redundant array of independent disks (raid) write cache sub-assembly |
| CN102521147A (en) * | 2011-11-17 | 2012-06-27 | 曙光信息产业(北京)有限公司 | Management method by using rapid non-volatile medium as cache |
| CN104035887A (en) * | 2014-05-22 | 2014-09-10 | 中国科学院计算技术研究所 | Block device caching device and method based on simplification configuration system |
| CN104679442A (en) * | 2013-12-02 | 2015-06-03 | 中兴通讯股份有限公司 | Method and device for improving performance of disk array |
| CN105339885A (en) * | 2013-05-23 | 2016-02-17 | 网络装置公司 | Efficient storage of small random changes to data on disk |
| CN105955841A (en) * | 2016-04-18 | 2016-09-21 | 华中科技大学 | Method for RAID dual-controller to write cache mirror with disk |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7930499B2 (en) * | 2007-08-15 | 2011-04-19 | Digi-Data Corporation | Method to accelerate block level snapshots in archiving storage systems |
-
2016
- 2016-12-05 CN CN201611106177.5A patent/CN106528001B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1527973A (en) * | 2000-06-23 | 2004-09-08 | 英特尔公司 | Non-volatile cache |
| CN101236526A (en) * | 2007-01-30 | 2008-08-06 | 扬智科技股份有限公司 | Computer system having cache system directly connected to nonvolatile storage device |
| CN102203751A (en) * | 2008-10-30 | 2011-09-28 | 惠普开发有限公司 | Redundant array of independent disks (raid) write cache sub-assembly |
| CN102043731A (en) * | 2010-12-17 | 2011-05-04 | 天津曙光计算机产业有限公司 | Cache system of storage system |
| CN102053929A (en) * | 2010-12-17 | 2011-05-11 | 天津曙光计算机产业有限公司 | IO cache operation method and system based on DM layer of Linux system |
| CN102147802A (en) * | 2010-12-17 | 2011-08-10 | 曙光信息产业(北京)有限公司 | Pseudo-random type NFS application acceleration system |
| CN102521147A (en) * | 2011-11-17 | 2012-06-27 | 曙光信息产业(北京)有限公司 | Management method by using rapid non-volatile medium as cache |
| CN105339885A (en) * | 2013-05-23 | 2016-02-17 | 网络装置公司 | Efficient storage of small random changes to data on disk |
| CN104679442A (en) * | 2013-12-02 | 2015-06-03 | 中兴通讯股份有限公司 | Method and device for improving performance of disk array |
| CN104035887A (en) * | 2014-05-22 | 2014-09-10 | 中国科学院计算技术研究所 | Block device caching device and method based on simplification configuration system |
| CN105955841A (en) * | 2016-04-18 | 2016-09-21 | 华中科技大学 | Method for RAID dual-controller to write cache mirror with disk |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106528001A (en) | 2017-03-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106528001B (en) | A kind of caching system based on nonvolatile memory and software RAID | |
| Yang et al. | I-CASH: Intelligently coupled array of SSD and HDD | |
| US10289544B2 (en) | Mapping tables for storage devices | |
| US10521131B2 (en) | Storage apparatus and storage control apparatus | |
| CN102521147B (en) | Management method by using rapid non-volatile medium as cache | |
| CN103530237B (en) | A kind of rubbish recovering method of Solid-state disc array | |
| CN111679795B (en) | Lock-free concurrent IO processing method and device | |
| US9996557B2 (en) | Database storage system based on optical disk and method using the system | |
| CN106445405B (en) | Data access method and device for flash memory storage | |
| WO2018019119A1 (en) | Method and device for dynamic partial-parallel data layout for continuous data storage | |
| JP2013514561A (en) | Storage system | |
| Mao et al. | SAR: SSD assisted restore optimization for deduplication-based storage systems in the cloud | |
| US8862819B2 (en) | Log structure array | |
| CN103207883A (en) | Method For Metadata Persistence | |
| TW201028843A (en) | Method and system for dynamic storage tiering using allocate-on-write snapshots | |
| CN105446665A (en) | Computer storage acceleration system and optimization method thereof | |
| Myers | On the use of NAND flash memory in high-performance relational databases | |
| CN109739696B (en) | Double-control storage array solid state disk caching acceleration method | |
| CN109144899A (en) | Method for managing table recovery | |
| US11379326B2 (en) | Data access method, apparatus and computer program product | |
| CN114356246B (en) | Storage management method and device for SSD internal data, storage medium and SSD device | |
| CN107451069A (en) | The rubbish recovering method of Solid-state disc array based on headspace | |
| CN104461930A (en) | Cache writing method and device | |
| CN115480697A (en) | Data processing method and device, computer equipment and storage medium | |
| US20220221988A1 (en) | Utilizing a hybrid tier which mixes solid state device storage and hard disk drive storage |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |