CN106970764B - It is a kind of based on band version watt record a RAID write ordered method and system - Google Patents
It is a kind of based on band version watt record a RAID write ordered method and system Download PDFInfo
- Publication number
- CN106970764B CN106970764B CN201710102277.9A CN201710102277A CN106970764B CN 106970764 B CN106970764 B CN 106970764B CN 201710102277 A CN201710102277 A CN 201710102277A CN 106970764 B CN106970764 B CN 106970764B
- Authority
- CN
- China
- Prior art keywords
- band
- version
- lower brush
- group
- verification
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000012795 verification Methods 0.000 claims abstract description 43
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 9
- 230000001960 triggered effect Effects 0.000 claims abstract 2
- 230000008569 process Effects 0.000 abstract description 11
- 230000003321 amplification Effects 0.000 description 3
- 238000003199 nucleic acid amplification method Methods 0.000 description 3
- 239000002699 waste material Substances 0.000 description 3
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 230000001680 brushing effect Effects 0.000 description 2
- 230000002045 lasting effect Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000004064 recycling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
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
- G06F3/0668—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
-
- 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
- G06F3/061—Improving I/O performance
-
- 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
- G06F3/0638—Organizing or formatting or addressing of data
-
- 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
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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention proposes that an a kind of watt record RAID based on band version writes ordered method and system, it is related to computer memory technical field, this method includes setting high speed persistently caching, for receiving all non-sequential write requests, lower brush is triggered when the high speed, which persistently caches, is written to water level, it wherein carries out lower brush as unit of group, generates new version for each band in group to be written when lower brush;For each version of band, primary verification is calculated, and will be in the write-in band of strip data and verification sequence.The RMW process of the RMW process and watt recording disc that generate the invention avoids traditional RAID5/6 small letter, improves the overall performance of system.
Description
Technical field
The present invention relates to computer memory technical field, in particular to an a kind of watt record RAID based on band version writes suitable
Sequence method and system.
Background technique
The surface density of traditional magnetic disk as primary storage medium quickly will be close to 1TB/In2The limit also means that
The memory space of single-deck has been difficult to have greatly improved again.In order to solve this problem, in recent years, a kind of new disk skill
Art proposed in the industry, i.e. watt recording technique.Its storage that disk is further promoted by the way of partial stack adjacent track is close
Degree, however, it is this superposition magnetic track mode but make while increasing disk storage capacity disk can not directly original place more
Newly, data on its adjacent track can be covered because updating and writing the data on a tracks.For the damage for avoiding adjacent track data
It is bad, it needs to make reading modification and writes (Read-Modify-Write, RMW).As name word description, read from disk first
Then partial data modifies a part of data therein, modified and unmodified data is finally write back disk together again.
The partial data wherein read will guarantee the valid data that other positions on disk will not be covered when writing back.Therefore, the worst
In the case of, if disk is filled with valid data, a RMW may need to operate entire disk, cause it is serious write amplification, greatly
Reduction disk non-sequential write performance.Amplification is write in order to reduce, in a certain range by RMW control, reduces write performance expense,
Disk can be divided into continuous region, i.e. band (Band), as shown in Figure 1 (assuming that writing the farthest only meeting of a tracks (Track)
The data on next tracks, i.e. coverage rate K=2 are covered, each band is made of two tracks), the addressing between is continuous
, but reserved certain space (Gap) is needed between band, it is ensured that write any position on band 1 all and will not cover and is on band 2 any
Data.In this manner it is possible to which amplification control will be write in the range of a band, therefore write the size that magnification level depends on band
(Band Size).Due to needing to waste disk storage capacity in adjacent stripe reserved space with band partitioning discs, to reduce
Disk size waste, band size cannot be too small, because smaller with size, the quantity of band is more on disk, disk space waste
It is more serious.Even if band is dimensioned to tens MB or so, each write request of user may cause RMW, the worst
In the case where need the data that will entirely take all to read, modified and be written back, performance can also decline to a great extent.
The RAID system of tape verifying for RAID5/6 or more higher-dimension based on band, under non-sequential write field scape, due to
The presence of small letter, each in disk takes and can introduce frequent non-sequential read-write operation, since watt recording disc is non-suitable
Sequence write performance is very poor, and the non-sequential write performance of RAID system entirety will receive serious influence.
Summary of the invention
In order to solve the above technical problems, the structure that the present invention uses circular logging to the band of watt recording disc carries out pipe
Reason, such as Fig. 2, write request write head pointer position, and head pointer is moved forward, black region is valid data block, and shadow region is
The invalid data block generated due to updating operation, since watt recording technique does not support original place to update, invalid data block can only pass through
Reclaimer operation could be utilized again.Recycling is to read valid data to head pointer direction sequence from tail pointer, and it is invalid to skip over
Data block, and by valid data block write head pointer position, so that invalid data block the space occupied is released, to realize in band
The release in space.It is ordered for ensuring to write with interior circular logging structure, avoids the RMW operation of watt recording disc.This method is kept away
The small letter process for having exempted from traditional RAID5/6, with the interim method for increasing band redundancy, is avoided small by doing version to band
Random read operation during writing, and by it is non-sequential write be converted to it is non-to improve watt record RAID system with interior sequential write
Sequence write performance, the present invention propose that an a kind of watt record RAID based on band version writes ordered method and system.
The present invention proposes that an a kind of watt record RAID based on band version writes ordered method,
Step 1, setting high speed persistently caches, for receiving all non-sequential write requests, when the high speed persistently caches quilt
Triggering lower brush when writing water level generates new edition for each band in group to be written when lower brush wherein carry out lower brush as unit of group
This;
Step 2, for each version of band, primary verification is calculated, and by the write-in of strip data and verification sequence
In band.
Band one group of composition in the step 1 in different disk with identical reel number.
It further include that data and verification invalid in old version are removed by automatic de-redundant and versions merging.
Automatic de-redundant is including being first band build version bitmap, when each column of remaining all version on version bitmap are patrolled
Volume or the resulting value of operation be all larger than or equal to same column on a certain legacy version value, then by the data in a certain legacy version
It is discharged with verification, otherwise carries out versions merging.
Versions merging include before data block carries out lower brush in organizing judge in group each band whether have sufficient space into
The lower brush of row, if had, directly generates new version and corresponding band is written, the tail for the band otherwise never brushed under sufficient space
Pointer carries out data block scan to head pointer, and reading scan to the affiliated band of data block it is newest in all versions in organizing
Data block recalculates verification, generates the latest edition of the band, writes back in array, and by old version data and verifies
It is invalid that the space occupied is labeled as, thus the space that release takes.
The present invention also proposes that an a kind of watt record RAID based on band version writes ordered system,
New version module is generated, for high speed persistently caching to be arranged, for receiving all non-sequential write requests, when the height
Triggering lower brush when fast persistently caching is written to water level, wherein carry out lower brush as unit of group, when lower brush is each in group to be written
Band generates new version;
Writing module calculates primary verification for each version for band, and by strip data and verification sequence
Write-in band in.
Band one group of composition in the generation new version module in different disk with identical reel number.
It further include that data and verification invalid in old version are removed by automatic de-redundant and versions merging.
Automatic de-redundant is including being first band build version bitmap, when each column of remaining all version on version bitmap are patrolled
Volume or the resulting value of operation be all larger than or equal to same column on a certain legacy version value, then by the data in a certain legacy version
It is discharged with verification, otherwise carries out versions merging.
Versions merging include before data block carries out lower brush in organizing judge in group each band whether have sufficient space into
The lower brush of row, if had, directly generates new version and corresponding band is written, the tail for the band otherwise never brushed under sufficient space
Pointer carries out data block scan to head pointer, and reading scan to the affiliated band of data block it is newest in all versions in organizing
Data block recalculates verification, generates the latest edition of the band, writes back in array, and by old version data and verifies
It is invalid that the space occupied is labeled as, thus the space that release takes.
As it can be seen from the above scheme the present invention has the advantages that
By high speed, persistently caching assembles non-sequential write request, is that the data of brush and verification to be played to generate band before lower brush
Version, and the automatic de-redundant of version and versions merging are done, it establishes version for band in lower brush and realizes sequential write disk.The present invention is kept away
The RMW process for having exempted from RMW process and watt recording disc that traditional RAID5/6 small letter generates, improves the overall performance of system.
Detailed description of the invention
Fig. 1 is disc space by band division figure;
Fig. 2 is with interior circular logging structure chart;
Fig. 3 is system architecture diagram;
Fig. 4 is band version figure;
Fig. 5 is band version bitmap;
Fig. 6 is the automatic de-redundant figure of version;
Fig. 7 is band versions merging figure.
Specific embodiment
The present invention joined one layer of high speed on traditional RAID and persistently cache (such as: SSD), all non-sequential for receiving
Write request triggers lower brush when caching is written to water level, and such as Fig. 3, lower brush is carried out (in Fig. 3 in dotted line frame as unit of group
One group of band composition in different disk with identical reel number), it will not mutually be generated between group when lower brush and write covering, group is interior non-whole
Band writes the small letter process that will not trigger, but generates new version for the band to be written, and organizes 0 as shown in Figure 4: 4 of band 0
Version.Wherein No. 0 version includes data D00 (the 0 of left side indicates version number, and the 0 of right side indicates reel number) and corresponding verification P0
(0 indicate band version number), No. 1 version include data D11 and corresponding verification P1, No. 2 versions include data D22 and D23 with
And corresponding verification P2, No. 3 versions include data D30, D31, D33 and verification P3.For each version of band, all need to count
Primary verification is calculated, and by the corresponding band of the write-in of the strip data of new version and verification sequence, so as to avoid RAID5/6 system
The small letter process of system also avoids the RWM process of watt recording disc.Due to the different editions of band can introduce data redundancy and
The redundancy of verification, therefore, the method that the present invention proposes two kinds of version de-redundant, i.e., automatic de-redundant and versions merging are gone through for removing
Invalid data and verification in history version.It is first band build version bitmap, such as Fig. 5, band version before automatic de-redundant
In have the position marks 1 of data, the not no position mark 0 of data.It, can for each of band version bitmap old version
Automatically the principle discharged is that each column of remaining all version on version bitmap do logic or the resulting value of operation is all larger than or waits
In the value of same column in version to be released, such as Fig. 6, when band 0:V3 will be brushed down, V0 and V1 can discharge automatically, at this time will
V0 and V1 has been written into data in band and verification labeled as hole, i.e., in vain, waits to be recycled, i.e. versions merging.Due to going automatically
It is superfluous to accomplish to completely remove all redundant datas and verification, it is therefore desirable to the complete of version is realized in conjunction with versions merging method
Full de-redundant needs to judge whether the interior each band of group has and brushes under sufficient space, that is, take pointer end to end that is, before the lower brush of group
Between whether have and brushed under enough spaces, if had, directly generate new version write-in, otherwise need to start version conjunction
And the tail pointer for the band never brushed under sufficient space carries out data block scan, and the data that reading scan is arrived to head pointer
Newest data block of the affiliated band of block in organizing in all versions, recalculates verification, generates the latest edition of the band, and
Old version data and verification the space occupied are labeled as hole, thus the space that release takes, such as Fig. 7.It is this to be based on band
The RAID of version writes the multiple random read-write disk that ordered method had both avoided traditional RAID5/6 small letter, in turn avoids a watt note
It is ordered to realize RAID system writing on single-deck for the RMW expense for recording disk.
A specific embodiment of the invention is given below, detailed description is made that the present invention in conjunction with attached drawing.Packet of the present invention
It includes non-sequential write request aggregation and ordered two parts is write based on band version.
(1) non-sequential write request aggregation
As shown in figure 3, D indicates to play the data block of brush, the corresponding check block of band where P indicates data block.To own
In the non-sequential lasting caching of write-in high speed (such as SSD).After lasting caching is written into certain water level (water level), touching
Give brush operation.Selection will play the data block of brush and calculate its verification as unit of group when lower brush, it is ensured that each of in each group
Band is interior to be written as sequential write when brush under a wheel.Since SSD has accumulated the non-sequential write request in a large amount of bands, does in band and write
When ordered can by it is more it is non-sequential write it is ordered, so that it is wide to maximize lower brush.
(2) ordered based on writing for band version
Such as Fig. 4, by taking band 0 as an example, Dxy indicates that the y number block in the band of version number x, Px indicate version number x's
The check block of band.Assuming that band 0 has had 3 versions to be present in disk, it is V0, V1 and V2, the data of V3 version respectively
D30, D31, D33 are that the epicycle assembled in SSD needs the lower data block brushed, corresponding to be verified as P3.Before being brushed at V3 first
The automatic de-redundant for doing band is first 0 build version bitmap of band such as Fig. 5 and Fig. 6, and does version de-redundant, de-redundant principle: version
Each column of remaining all version do logic on bitmap or the resulting value of operation is all larger than or is equal to same column in version to be released
Value.V0 and V1 can be by automatic de-redundant before brushing at V3, and its space occupied can be labeled as hole, waits versions merging
Operation discharges its space occupied.
After the completion of the automatic de-redundant of version, before brushing at V3, need to judge each take whether have sufficient space in group
Write-in, if space is inadequate, needs to start versions merging, the tail pointer for the band never brushed under sufficient space starts to refer to head
Needle traversal, it is assumed that the position that current tail pointer is directed toward is the data block in band 0, then will organize in interior all bands includes band 0
Newest data in old version are read, and recalculate verification, and by newest data and verification in company with the data that play brush
It is sequentially written in lower brush in band with verification.Such as Fig. 7, v0 and v1 is eliminated by automatic de-redundant, v2 is eliminated by versions merging, so
New version v0 is constructed with the latest data for each taking band 0 afterwards, corresponding disk in write-in group when waiting lower brush.One wheel version
This merging moves forward the tail pointer of current band after finishing, to release disk space, repeats this process, takes until currently
There are enough spaces to complete lower brush.By automatic de-redundant and versions merging, the sky that the old version in group occupies can be discharged
Between, so that it is guaranteed that lower brush can be normally written, RMW process and watt recording disc for avoiding RAID5/6 system small letter are non-sequential
RMW process when writing.
The present invention also proposes that an a kind of watt record RAID based on band version writes ordered system, comprising:
New version module is generated, for high speed persistently caching to be arranged, for receiving all non-sequential write requests, when the height
Triggering lower brush when fast persistently caching is written to water level, wherein carry out lower brush as unit of group, when lower brush is each in group to be written
Band generates new version;
Writing module calculates primary verification for each version for band, and by strip data and verification sequence
Write-in band in.
Band one group of composition in the generation new version module in different disk with identical reel number.
It further include that data and verification invalid in old version are removed by automatic de-redundant and versions merging.
Automatic de-redundant is including being first band build version bitmap, when each column of remaining all version on version bitmap are patrolled
Volume or the resulting value of operation be all larger than or equal to same column on a certain legacy version value, then by the data in a certain legacy version
It is discharged with verification, otherwise carries out versions merging.
Versions merging is included in front of group carries out lower brush, judges whether each band has sufficient space to carry out lower brush in group,
If had, directly generate new version and corresponding band be written, the tail pointer of the band otherwise never brushed under sufficient space to
Head pointer carries out data block scan, and reading scan to latest data of the affiliated band of data block in organizing in all versions
Block recalculates verification, generates the latest edition of the band, and old version data and verification the space occupied are labeled as nothing
Effect, thus the space that release takes.
Claims (8)
1. an a kind of watt record RAID based on band version writes ordered method characterized by comprising
Step 1, setting high speed persistently caches, and for receiving all non-sequential write requests, is written to when the high speed persistently caches
Lower brush is triggered when water level, wherein carry out lower brush as unit of group, generates new version for each band in group to be written when lower brush;
Step 2, for each version of band, primary verification is calculated, and will be in the write-in band of strip data and verification sequence;
It wherein further include that the tail pointer of current band is moved forward later by automatic de-redundant and versions merging, the history in release group
The occupied space of version.
2. watt record RAID based on band version writes ordered method as described in claim 1, which is characterized in that the step
Band one group of composition in rapid 1 in different disk with identical reel number.
3. watt record RAID based on band version writes ordered method as described in claim 1, which is characterized in that go automatically
It is superfluous include first be band build version bitmap, when each column of remaining all version on version bitmap do logic or operation is resulting
Value be all larger than or equal to same column on a certain legacy version value, then by a certain legacy version data and verification carry out version
It is discharged when this merging.
4. watt record RAID based on band version writes ordered method as claimed in claim 3, which is characterized in that version is closed
And including before data block carries out lower brush in organizing, judging whether each band has sufficient space to carry out lower brush in group, if all
Have, then directly generate new version and corresponding band is written, the tail pointer of the band otherwise never brushed under sufficient space is to head pointer
Carry out data block scan, and reading scan to latest data block of the affiliated band of data block in organizing in all versions, again
Verification is calculated, the latest edition of the band is generated, writes back in array, and by old version data and verifies the space occupied mark
It is invalid to be denoted as, thus the space that release takes.
5. an a kind of watt record RAID based on band version writes ordered system characterized by comprising
New version module is generated, for high speed persistently caching to be arranged, for receiving all non-sequential write requests, when the high speed is held
Triggering lower brush when caching is written to water level long is each band in group to be written when lower brush wherein carry out lower brush as unit of group
Generate new version;
Writing module calculates primary verification for each version for band, and writing strip data and verification sequence
Enter in band;
It wherein further include that the tail pointer of current band is moved forward later by automatic de-redundant and versions merging, the history in release group
The occupied space of version.
6. watt record RAID based on band version writes ordered system as claimed in claim 5, which is characterized in that the life
At band one group of composition in different disk in new version module with identical reel number.
7. watt record RAID based on band version writes ordered system as claimed in claim 5, which is characterized in that go automatically
It is superfluous include first be band build version bitmap, when each column of remaining all version on version bitmap do logic or operation is resulting
Value be all larger than or equal to same column on a certain legacy version value, then by a certain legacy version data and verification carry out version
It is discharged when this merging.
8. watt record RAID based on band version writes ordered system as claimed in claim 7, which is characterized in that version is closed
And including before data block carries out lower brush in organizing, judging whether each band has sufficient space to carry out lower brush in group, if all
Have, then directly generate new version and corresponding band is written, the tail pointer of the band otherwise never brushed under sufficient space is to head pointer
Carry out data block scan, and reading scan to latest data block of the affiliated band of data block in organizing in all versions, again
Verification is calculated, the latest edition of the band is generated, writes back in array, and by old version data and verifies the space occupied mark
It is invalid to be denoted as, thus the space that release takes.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710102277.9A CN106970764B (en) | 2017-02-24 | 2017-02-24 | It is a kind of based on band version watt record a RAID write ordered method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710102277.9A CN106970764B (en) | 2017-02-24 | 2017-02-24 | It is a kind of based on band version watt record a RAID write ordered method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106970764A CN106970764A (en) | 2017-07-21 |
CN106970764B true CN106970764B (en) | 2019-08-16 |
Family
ID=59328685
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710102277.9A Expired - Fee Related CN106970764B (en) | 2017-02-24 | 2017-02-24 | It is a kind of based on band version watt record a RAID write ordered method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106970764B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108496161A (en) * | 2017-09-29 | 2018-09-04 | 深圳市大疆创新科技有限公司 | Data buffer storage device and control method, data processing chip, data processing system |
CN107908499B (en) * | 2017-11-02 | 2021-03-05 | 华为技术有限公司 | Data verification method and device |
CN108376053B (en) * | 2018-01-23 | 2020-02-28 | 中国科学院计算技术研究所 | Tile record writing sequencing method and device supporting in-place updating |
CN109582222B (en) * | 2018-10-31 | 2020-11-24 | 华中科技大学 | Method for cleaning persistent cache in host sensing tile recording disk |
CN109710184B (en) * | 2018-12-19 | 2022-02-18 | 中国人民解放军国防科技大学 | Hierarchical hybrid storage method and system for tile record disk perception |
CN111522501B (en) * | 2019-02-01 | 2022-01-21 | 杭州海康威视数字技术股份有限公司 | Disk array space dividing method and device, electronic equipment and storage medium |
CN111949434B (en) * | 2019-05-17 | 2022-06-14 | 华为技术有限公司 | RAID management method, RAID controller and system |
CN110413229A (en) * | 2019-07-10 | 2019-11-05 | 华中科技大学 | A kind of method for writing data |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105824579A (en) * | 2016-03-17 | 2016-08-03 | 中国科学院计算技术研究所 | Write amplification overhead reducing method including read and write for shingled magnetic recording |
CN105955664A (en) * | 2016-04-29 | 2016-09-21 | 华中科技大学 | Method for reading and writing segment-based shingle translation layer (SSTL) |
CN106293543A (en) * | 2016-08-25 | 2017-01-04 | 中国科学院计算技术研究所 | A kind of based on incidentally reclaim watt record a RAID write ordered method and device |
-
2017
- 2017-02-24 CN CN201710102277.9A patent/CN106970764B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105824579A (en) * | 2016-03-17 | 2016-08-03 | 中国科学院计算技术研究所 | Write amplification overhead reducing method including read and write for shingled magnetic recording |
CN105955664A (en) * | 2016-04-29 | 2016-09-21 | 华中科技大学 | Method for reading and writing segment-based shingle translation layer (SSTL) |
CN106293543A (en) * | 2016-08-25 | 2017-01-04 | 中国科学院计算技术研究所 | A kind of based on incidentally reclaim watt record a RAID write ordered method and device |
Non-Patent Citations (4)
Title |
---|
"Design and Implementation of Hybrid Shingled Recording RAID System";ZW Lu等;《2016 IEEE 14th Intl Conf on Dependable,Autonomic & Secure Computing,Intl Conf on Pervasive Intelligence & Computing & Cyber Science & Technology Congress》;20160812;I137-20 * |
"RAID5小写更新的冗余管理机制";刘秀菊;《微电子学与计算机》;20120731;第29卷(第07期);第113-115页 * |
"提高磁盘阵列小写性能的方法研究";李洁琼等;《计算机工程》;20031031;第29卷(第18期);第57-59页 * |
"瓦记录磁盘写特征研究与优化设计";周志华;《中国优秀硕士学位论文全文数据库 信息科技辑》;20130815(第08期);第937-942页 * |
Also Published As
Publication number | Publication date |
---|---|
CN106970764A (en) | 2017-07-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106970764B (en) | It is a kind of based on band version watt record a RAID write ordered method and system | |
CN106293543B (en) | It is a kind of that ordered method and device is write based on watt record RAID incidentally recycled | |
US9514055B2 (en) | Distributed media cache for data storage systems | |
CN103135940B (en) | Implementing enhanced fragmented stream handling in a shingled disk drive | |
US10394493B2 (en) | Managing shingled magnetic recording (SMR) zones in a hybrid storage device | |
CN107037976A (en) | Storage device and its method of work | |
CN108958656B (en) | Dynamic stripe system design method based on RAID5 solid state disk array | |
US5875455A (en) | Information recording and reproducing apparatus merging sequential recording requests into a single recording request, and method of data caching for such apparatus | |
JP2010211910A (en) | Data storage device | |
JPH03167644A (en) | Device and method for handling defect in semiconductor memory | |
JP2011134307A (en) | Semiconductor recording device and control method of semiconductor recording device | |
US11334277B2 (en) | Issuing efficient writes to erasure coded objects in a distributed storage system with two tiers of storage | |
US11334276B2 (en) | Using segment pre-allocation to support large segments | |
US20140208005A1 (en) | System, Method and Computer-Readable Medium for Providing Selective Protection and Endurance Improvements in Flash-Based Cache | |
US20130246886A1 (en) | Storage control apparatus, storage system, and storage control method | |
US20090307563A1 (en) | Replacing bad hard drive sectors using mram | |
JP2017010602A (en) | Disk storage device and data processing method | |
CN105845161A (en) | Method for reducing adjacent track interference effects in a data storage system | |
CN105824579B (en) | A method of incidentally read-write reduction watt record writes amplification expense | |
US7073029B2 (en) | Storage system using fast storage and log-structured storage | |
CN100468548C (en) | Defect management for HDD with variable index architecture | |
CN107402890B (en) | Data processing method and system based on solid-state disk array and cache | |
JP2007087062A (en) | Array type storage device | |
US11625370B2 (en) | Techniques for reducing data log recovery time and metadata write amplification | |
CN108376053A (en) | A kind of newer watt of record in support original place writes ordered method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190816 |
|
CF01 | Termination of patent right due to non-payment of annual fee |