CN104765660A - Single-disc quick recovery method machine based on RAID6 of SSD - Google Patents
Single-disc quick recovery method machine based on RAID6 of SSD Download PDFInfo
- Publication number
- CN104765660A CN104765660A CN201510199457.4A CN201510199457A CN104765660A CN 104765660 A CN104765660 A CN 104765660A CN 201510199457 A CN201510199457 A CN 201510199457A CN 104765660 A CN104765660 A CN 104765660A
- Authority
- CN
- China
- Prior art keywords
- matrix
- dish
- newly
- data
- redundancy
- 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.)
- Granted
Links
Abstract
The invention relates to a single-disc quick recovery method machine based on RAID6 of SSD. According to the method, based on the RAID6 system of the SSD, wherein the RAID6 system uses the RDP to code, by adding an additional piece of redundancy equipment, on the premise that the reliability is high when errors are allowed to happen to double-disc, the segmentation sum required by the single-disc recovery is further reduced, so that the recovery efficiency of the system single-disc is improved, the reconsitution time is shortened, the load balancing of the system is achieved, and finally the service quality of the system can be promoted. Aiming at the error of any single data disc in an original system, the fragmentation sum required by the single-disc recovery of the system can be reduced; aiming at the error of any single data disc in the original system, the bottleneck might borne by the original system in the single-disc recovery process can be fractured, and the load balancing of all equipment is achieved; by using the binary xor sum, the updating cost of a newly added redundancy disc is reduced, and therefore the service life of the newly added redundancy disc is prolonged.
Description
Technical field
The present invention is applicable to flash memory solid state storage technologies field, provide a kind of 6 grades of redundancy array in disc RAID 6(Redundant Arrays of Independent Disks 6 based on solid-state disk SSD (Solid State Drive)) the single-deck quick recovery method of system, the single-deck improving system recovers efficiency and reliability.
Background technology
Relative to traditional magnetic storage medium, flash memory (Flash Memory) technology relies on its significant advantage on volume, weight, power consumption and readwrite performance to start an emerging change of field of storage, is a kind of perfect medium building novel non-volatile main memory.
In order to technology and the Related product of compatible and inheriting tradition disk better, the solid-state storage SSD (Solid State Drive) that multiple chip portfolio based on nand flash memory is packaged in order to similar traditional magnetic disk by industry manufacturer applies, and utilize flash translation layer (FTL) FLT(Flash Translation Layer) and shield the internal diversity of SSD and traditional magnetic disk, make external user can obviously to obtain while performance promotion feeling is subject to perception less than the change of storage itself.
Nowadays, SSD based on flash memory obtains application widely in the data center of movable storage device, personal memory device and different levels, and along with the lifting of technique and the reduction of unit bit price, the application based on the SSD of flash memory can be more extensive.
In application based on the SSD of flash memory, being more the raid-array RAID(Redundant Arrays of Independent Disks utilizing polylith SSD to build high-performance and high reliability) system realizes.The core of raid-array RAID utilizes multiple device build can the carrying out redundancy encoding technology across stripe and extra disk and improve performance and system reliability respectively of concurrent reading and concurrent writing.The RAID technique of present comparative maturity comprises RAID0, RAID1, RAID5, RAID6, RAID10, RAID01 etc., and user can be configured according to different application demands.Must, it is relatively extensive that RAID5, RAID6 apply.But with regard to reliability, because RAID5 only has a check disk, and RAID6 has two check disks, so in reliability, RAID6 is more reliable.
We are for the RAID6(6 block dish of encoding based on RDP), its band by the identical four parts of data blocks of size and two parts of check blocks totally six piecemeals form, each piecemeal is stored in a different equipment respectively.First block check block be all original four parts of data blocks on this band XOR and, the second block check block be then a certain correspondence in piecemeal array positive oblique line on piecemeal XOR and, as Fig. 1.Owing to introducing redundancy, in order to ensure the consistance of data, so when upgrading original data block, the two block check blocks relevant to it are needed to upgrade.
With regard to storage system, because the instability of equipment itself can cause the unavailable of one or more memory device.Statistics shows, sets up standby probability of makeing mistakes up to more than 90%, and the probability that double plate is made mistakes is then that about 8%, three dishes are made mistakes simultaneously, is less than 1%.So typically, reliable storage system generally needs to hold double plate and makes mistakes, and can make mistakes for the single-deck of high probability and recover fast.So although RAID5 realizes simple, its reliability is far from being enough.In order to ensure better reliability, study for the RAID6 system based on SSD that double plate makes mistakes can be held herein.
In RAID6 system, to make mistakes due to individual equipment or its serviceable life reaches capacity, cause its data unavailable.Now, we need bad SSD equipment to reject, and use new SSD equipment to replace, and are reconstructed the data in faulty equipment.This restructuring procedure needs to read data on other healthy equipment and redundancy carries out the recovery calculating and realize data of decoding, and this can increase the request of system I/O, and takies corresponding resource.User needs to wait for that data reconstruction could be read and write the data on faulty equipment after completing, so how to shorten in RAID6 system single-deck release time, thus promotes the major issue that Consumer's Experience becomes storage system.
At present based on the correlative study of the fast quick-recovery of single-deck in RAID6 system.This research is the RAID6 system based on using RDP coding, and it indicates the inefficiencies be just reconstructed with a class check block in the storage system that RDP encodes, and proposes and demonstrates a kind of single-deck recovery policy of optimum.This strategy is reconstructed by utilizing two class check blocks of relative number, obtains the overlap partition of maximum number, thus makes the piecemeal minimum number needed for reconstructing single, and carried out equilibrium to loading condition.The method realizes simple and versatility is comparatively strong, but its performance boost is limited, is difficult to meet the demand that single-deck capacity explodes.
In addition, single-deck fast recovery problem in the storage system also having research to be encoded to EVENODD by similar method is studied, the single-deck that theory deduction has gone out EVENODD coding recovers lower limit, and the strategy of single-deck quick reconfiguration under giving isomerous environment, make restructuring procedure reach near-optimization.Meanwhile, other similar research is then the single-deck fast recovery problem towards the RAID system under some other coded system, and this is similar with research before.
These researchs are all carry out when single-deck capacity is less, and this capacity that have ignored single memory device is in recent years along with the feature of technique and technical development sharp increase, and unit stores price cheap system service quality is become become more and more important.And these above researchs, be all based on the changeless situation of redundancy condition.
Summary of the invention
Technical matters to be solved by this invention is carried out in the RAID6 system of encoding at the use RDP based on SSD, by increasing by one piece of extra redundance unit, ensureing to hold under the high reliability prerequisite that double plate makes mistakes, piecemeal sum needed for further minimizing single-deck recovers, thus improve system single-deck recovery efficiency, shorten reconstitution time, and realize the load balancing of system, finally realize elevator system service quality.
Technical scheme of the present invention is: because the single-deck recovery policy of traditional RAID6 system based on fixing redundancy reaches optimum, cannot continue to improve again, so the development trend being difficult to adapt to the increase of single-deck capacity and causing reconstitution time to increase.So the present invention proposes the thought increasing limited additional redundancy and carry out improving performance instead of increase reliability, newly increase a solid storage device and be used as additional redundancy dish, and for newly-increased redundancy dish is specifically encoded on the basis that original RDP encodes, with reduce any one data disks in original system break down be reconstructed time required piecemeal sum.The present invention had both been applicable to the solid-state disk system based on nand flash memory, went for again the storage system of traditional magnetic disk composition.Due to the similarity that EVENODD coding and RDP encode, the present invention can be used in the system of EVENODD coding.Core of the present invention is how to find the redundancy coding method that can obtain good compromise effect in many aspects such as performance, reliability, renewal expense, load balancing.By analysis and prove, the present invention utilizes some particular data in original system to carry out computing, the binary XOR obtaining meeting specified conditions and the coding realized newly-increased redundancy.
So, in the present invention for the key of the coding of newly-increased additional redundancy block be how to find suitable binary XOR and.
Concrete technical scheme is:
The first step, because original system uses RDP coding, we define three data structures, are the data that matrix matrix deposits in disk respectively, matrix diagonal houses the coordinate of diagonal line data in disk, and temp is the coordinate of the element in matrix principal diagonal.
Second step, initialization matrix matrix and matrix diagonal;
3rd step: assignment is carried out to matrix matrix and matrix diagonal;
4th step: by perform findTwoTuples (matrix [] [], diagonal [] [], int p) find qualified binary XOR and:
Each element in the current matrix of 4.1 traversal,
4.2 judge the position of the element in current matrix is in which clinodiagonal,
4.3 find out corresponding diagonally opposing corner, judge element in clinodiagonal whether in temp (in principal diagonal),
If 4.4 in principal diagonal, then can not form two tuples, if do not exist, then can form two tuples, and record;
5th step: find qualified binary XOR and afterwards, utilize specific binary XOR and newly-increased redundancy dish is encoded;
6th step: the single-deck recovery operation after newly-increased redundancy dish is encoded.
The process that the present invention realizes increasing additional redundancy Code And Decode refers to accompanying drawing explanation.
Use the present invention can reach following beneficial effect:
1, make mistakes for any individual data dish in original system, the burst sum needed for System recover single-deck can be reduced;
2, make mistakes for any individual data dish in original system, original system possibility load bottleneck in single-deck rejuvenation can be abolished, realize the load balancing of each equipment;
3, by use binary XOR and, as much as possible when obtaining performance near-optimization, reducing the renewal expense of newly-increased redundancy dish, thus extending its life-span.
Realize expense of the present invention very little, comprise following 2 points:
1, space expense: for whole RAID6 system adds an extra redundancy dish, the size of this dish is identical with the solid-state disk size in original system.Because individual equipment is relatively inexpensive, and have downward trend year by year, for the price of whole storage system, the space expense only increasing single redundancy dish is acceptable.
2, time overhead: owing to only relating to arithmetic speed xor operation quickly to the coding of additional redundancy dish, and look for special binary XOR with algorithm also relative simple and quick, and to carry out in the process that can encode at original RDP of this cataloged procedure simultaneously.In addition because newly-increased redundancy encoding can not reduce the reliability of original system, its computing after can waiting system stability at one's leisure section carry out.So time overhead of the present invention is also negligible.
accompanying drawing illustrates (all for the RAID6 system of six blocks of data dishes)
Fig. 1 is the coding schematic diagram of the RDP of the present invention before newly-increased redundancy dish;
Fig. 2 is that the single-deck of the RDP of the present invention before newly-increased redundancy dish recovers schematic diagram;
Fig. 3 be the present invention search for suitable binary XOR and algorithm flow chart;
Fig. 4 is the schematic diagram that the present invention utilizes specific binary XOR and encodes to newly-increased redundancy dish;
Fig. 5 is that the present invention utilizes specific binary XOR and the single-deck recovery operation schematic diagram after encoding to newly-increased redundancy dish;
Fig. 6 is the implementation effect figure that the present invention and original system carry out contrasting.
Embodiment
Fig. 1 is the coding schematic diagram of the RDP of the present invention before newly-increased redundancy dish.Totally eight pieces of memory devices in figure, wherein two pieces are used as redundancy, are respectively P and Q equipment.Redundant block in P equipment is all walked crosswise upper all data by corresponding level and is done XOR and obtained, as shown in the black arrow in figure.Redundant block in Q equipment is then by being positioned at data block on same positive diagonal line and P redundant block carries out XOR and obtains, namely in figure same shape symbol shown in.
Fig. 2 is that the single-deck of the RDP of the present invention before newly-increased redundancy dish recovers schematic diagram.
Fig. 3 be the present invention search for suitable binary XOR and algorithm flow chart.Concrete implementation is:
The first step: initialization prime P, matrix matrix and diagonal, and assignment.
Second step: judge in current matrix, in the current matrix of horizontal ordinate, whether horizontal ordinate i is less than p-1.If not be, then terminate, otherwise perform the 3rd step;
3rd step: put i=i+1;
4th step: judge in current matrix, whether ordinate j is less than p.If not, then perform second step;
5th step: can conclude that this operates the logical page (LPAGE) related to is general data, puts page_type=1, performs the 8th step;
6th step: put j=j+1;
7th step: judge position in matrix [i] [j] whether in diagonal [i] [j].If not, then perform the 6th step, otherwise perform the 8th step;
8th step: find corresponding clinodiagonal;
9th step: judge the element whether had in clinodiagonal in principal diagonal.If so, then perform the 7th step, otherwise perform the tenth step;
Tenth step: matrix [i] [j] and current diagonal [i] [j] form binary XOR and, and record.
To be the present invention find qualified binary XOR and afterwards, the schematic diagram (RAID6 for 6 blocks of data dishes) utilizing specific binary XOR and encode to newly-increased redundancy dish at the above-mentioned algorithm of use to Fig. 4.The RAID6 system of totally eight solid state disk devices of reason six blocks of data dish and two pieces of redundancy dish compositions, its checking data accounts for 2/8, after increasing a new redundancy dish X, becomes the system of nine equipment compositions, its checking data accounts for 3/9, and reliability is still for can tolerate that any two dishes are made mistakes.In newly-increased redundancy dish, its coding is not unique, as long as meet binary XOR and the condition of setting.As shown in the figure, having increased six redundancies newly, is a1+f1, a2+b2, b3+d3, c4+e4, d5+e5, c6+f6 respectively.After any one former data disks is made mistakes, first determine in X dish containing make mistakes dish data all binary XORs and, then select the corresponding redundant block in other two check disks P, Q to recover according to Maximum overlap principle.
Fig. 5 is that the present invention utilizes specific binary XOR and the single-deck recovery operation schematic diagram after encoding to newly-increased redundancy dish.As first data disks lost efficacy, namely need to recover a1, a2, a3, a4, a5, a6.First determine binary XOR containing element in failure disc in newly-increased redundancy dish and, as figure a2 and a4 can be undertaken recovering (a2+e2, a4+c4) by the newly-increased redundancy of correspondence in X dish respectively.Due to the binary XOR in X dish and be specific, according to Maximum overlap principle, must select respectively containing the check block q3 in the Q dish of e2 and c4, q6, a1 and a5 uses the check block p1 in corresponding P dish simultaneously, and p5 recovers, and so just can produce a repetitive sequence determined.
Fig. 6 is the implementation effect figure that the present invention and original system carry out contrasting.The data block can saved after then representing increase redundancy containing hypographous circle in figure, the performance boost namely before and after system.
For the flash array system that traditional RDP encodes, after single-deck lost efficacy, or use the redundancy of single type to recover, or use two class redundancies to recover.The former desired data amount is large, and the number of devices that the latter relates to is more, but required data volume can reach optimum.For figure six, the data volume needed for above-mentioned two kinds of reconstructing single algorithms is respectively 36 and 27.
For situation of the present invention, after a newly-increased specific redundancy dish, although add the number of devices involved by reconstruct, be the reduction of the data total amount needed for reconstruct, saved a part of data block, reconstitution time is shortened.For figure six, the data volume needed for restructing algorithm that the present invention proposes is respectively 22, and the optimized database restore before comparing improves 18.5%, and load has been distributed on more equipment.
This shows, the present invention according to utilizing newly-increased redundancy dish, reduce after any individual data dish is made mistakes and reconstructing required data total amount, reaches the effect accelerated single-deck and recover.
Claims (4)
1. the single-deck quick recovery method based on the RAID6 system of SSD, use RDP based on SSD carries out the RAID6 system of encoding, increase by one piece of additional redundancy dish, ensureing to hold under the prerequisite that double plate makes mistakes, piecemeal sum needed for further minimizing single-deck recovers, it is characterized in that, increase a solid storage device and be used as additional redundancy dish, and for newly-increased redundancy dish is encoded on the basis that original RDP encodes, with reduce any one data disks in original system break down be reconstructed time required piecemeal sum, concrete steps are:
The first step, define three data structures, be the data that matrix matrix deposits in disk respectively, matrix diagonal houses the coordinate of diagonal line data in disk, and temp is the coordinate of the element in matrix principal diagonal;
Second step, initialization matrix matrix and matrix diagonal;
3rd step: assignment is carried out to matrix matrix and matrix diagonal;
4th step: by perform findTwoTuples find qualified binary XOR and:
5th step: find qualified binary XOR and afterwards, utilize specific binary XOR and newly-increased redundancy dish is encoded;
6th step: the single-deck recovery operation after newly-increased redundancy dish is encoded.
2. the single-deck quick recovery method of a kind of RAID6 system based on SSD according to claim 1, is characterized in that, find qualified binary XOR and algorithm flow be:
2.1st step: initialization prime P, matrix matrix and diagonal, and assignment;
2.2nd step: judge that in current matrix, in the current matrix of horizontal ordinate, whether horizontal ordinate i is less than p-1, if not be, then terminates, otherwise perform the 2.3rd step;
2.3rd step: put i=i+1;
2.4th step: judge in current matrix, whether ordinate j is less than p, if not, then perform the 2.2nd step;
2.5th step: concluding that this operates the logical page (LPAGE) related to is general data, puts page_type=1, performs the 2.8th step;
2.6th step: put j=j+1;
2.7th step: judge position in matrix [i] [j] whether in diagonal [i] [j], if not, then perform the 2.6th step, otherwise perform the 2.8th step;
2.8th step: find corresponding clinodiagonal;
2.9th step: judge the element whether had in clinodiagonal in principal diagonal, if so, then performs the 2.7th step, otherwise performs the 2.10th step;
2.10th step: matrix [i] [j] and current diagonal [i] [j] form binary XOR and, and record.
3. the single-deck quick recovery method of a kind of RAID6 system based on SSD according to claim 1, is characterized in that, described 5th step detailed process is:
The RAID6 system of totally eight solid state disk devices be made up of six blocks of data dishes and two pieces of redundancy dishes, its checking data accounts for 2/8, after increasing a new redundancy dish, become the system of nine equipment compositions, its checking data accounts for 3/9, reliability is for tolerating that any two dishes are made mistakes, in newly-increased redundancy dish, its coding is not unique, as long as meet binary XOR and the condition of setting, newly-increased six redundancies, a1+f1 respectively, a2+b2, b3+d3, c4+e4, d5+e5, c6+f6, after any one former data disks is made mistakes, first determine in new redundancy dish containing make mistakes dish data all binary XORs and, then the corresponding redundant block in other two check disks is selected to recover according to Maximum overlap principle.
4. the single-deck quick recovery method of a kind of RAID6 system based on SSD according to claim 1, is characterized in that, the single-deck recovery operation after encoding to newly-increased redundancy dish, and detailed process is:
As first data disks lost efficacy, namely need to recover a1, a2, a3, a4, a5, a6, first determine binary XOR containing element in failure disc in newly-increased redundancy dish and, a2 and a4 can carry out recoverys a2+e2, a4+c4 by newly-increased redundancy corresponding in newly-increased redundancy dish respectively, due to the binary XOR in newly-increased redundancy dish be specific, according to Maximum overlap principle, must select respectively containing the check block q3 in the Q dish of e2 and c4, q6, a1 and a5 uses the check block p1 in corresponding P dish simultaneously, p5 recovers, and produces a repetitive sequence determined like this.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510199457.4A CN104765660B (en) | 2015-04-24 | 2015-04-24 | A kind of single-deck quick recovery method device of the RAID6 systems based on SSD |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510199457.4A CN104765660B (en) | 2015-04-24 | 2015-04-24 | A kind of single-deck quick recovery method device of the RAID6 systems based on SSD |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104765660A true CN104765660A (en) | 2015-07-08 |
| CN104765660B CN104765660B (en) | 2017-04-05 |
Family
ID=53647505
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510199457.4A Active CN104765660B (en) | 2015-04-24 | 2015-04-24 | A kind of single-deck quick recovery method device of the RAID6 systems based on SSD |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104765660B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105824725A (en) * | 2016-03-10 | 2016-08-03 | 中国人民解放军国防科学技术大学 | X-coding extension method capable of quickly recovering and realizing load balance |
| CN108228382A (en) * | 2018-01-11 | 2018-06-29 | 成都信息工程大学 | A kind of data reconstruction method for EVENODD code single-deck failures |
| WO2019090756A1 (en) * | 2017-11-13 | 2019-05-16 | 清华大学 | Raid mechanism-based data storage system for sharing resources globally |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102012792A (en) * | 2010-11-02 | 2011-04-13 | 华中科技大学 | Quick reconfigurable RAID-6 coding and reconfiguration method |
| CN102103533A (en) * | 2011-02-25 | 2011-06-22 | 华中科技大学 | Method for reconstructing single disk in double-disk fault-tolerance disk array |
| US8595606B1 (en) * | 2010-07-16 | 2013-11-26 | The Research Foundation Of State University Of New York | Extended row diagonal parity with optimal decoding procedure |
| CN103593253A (en) * | 2013-11-22 | 2014-02-19 | 华中科技大学 | Vertical RAID-6 coding method based on exclusive or |
| CN103809919A (en) * | 2013-12-16 | 2014-05-21 | 华中科技大学 | Efficient and multi-fault-tolerant code quick recovery method and validation matrix generating method thereof |
-
2015
- 2015-04-24 CN CN201510199457.4A patent/CN104765660B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8595606B1 (en) * | 2010-07-16 | 2013-11-26 | The Research Foundation Of State University Of New York | Extended row diagonal parity with optimal decoding procedure |
| CN102012792A (en) * | 2010-11-02 | 2011-04-13 | 华中科技大学 | Quick reconfigurable RAID-6 coding and reconfiguration method |
| CN102103533A (en) * | 2011-02-25 | 2011-06-22 | 华中科技大学 | Method for reconstructing single disk in double-disk fault-tolerance disk array |
| CN103593253A (en) * | 2013-11-22 | 2014-02-19 | 华中科技大学 | Vertical RAID-6 coding method based on exclusive or |
| CN103809919A (en) * | 2013-12-16 | 2014-05-21 | 华中科技大学 | Efficient and multi-fault-tolerant code quick recovery method and validation matrix generating method thereof |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105824725A (en) * | 2016-03-10 | 2016-08-03 | 中国人民解放军国防科学技术大学 | X-coding extension method capable of quickly recovering and realizing load balance |
| WO2019090756A1 (en) * | 2017-11-13 | 2019-05-16 | 清华大学 | Raid mechanism-based data storage system for sharing resources globally |
| US10997025B2 (en) | 2017-11-13 | 2021-05-04 | Tsinghua University | RAID-based globally resource-shared data storage system |
| CN108228382A (en) * | 2018-01-11 | 2018-06-29 | 成都信息工程大学 | A kind of data reconstruction method for EVENODD code single-deck failures |
| CN108228382B (en) * | 2018-01-11 | 2021-08-10 | 成都信息工程大学 | Data recovery method for single-disk fault of EVENODD code |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104765660B (en) | 2017-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101387975B (en) | Magnetic disc array system | |
| CN111480148B (en) | Storage system with peer-to-peer data recovery | |
| Huang et al. | Erasure coding in windows azure storage | |
| Vajha et al. | Clay codes: Moulding {MDS} codes to yield an {MSR} code | |
| CN103793182B (en) | System and method for expansible storage protection | |
| US9063910B1 (en) | Data recovery after triple disk failure | |
| CN102521068B (en) | Reconstructing method of solid-state disk array | |
| US7529970B2 (en) | System and method for improving the performance of operations requiring parity reads in a storage array system | |
| CN102880428B (en) | The creation method of distributed Redundant Array of Independent Disks (RAID) and device | |
| Park et al. | Reliability and performance enhancement technique for SSD array storage system using RAID mechanism | |
| CN111465922B (en) | Storage system with peer-to-peer data scrubbing | |
| CN101576833B (en) | Data reconstruction method for Redundant Array of Independent Disks (RAID) and appliance thereof | |
| Greenan et al. | Building flexible, fault-tolerant flash-based storage systems | |
| KR102572357B1 (en) | Raid-6 data storage device and data processing system having the same | |
| CN103488432B (en) | Hybrid disk array, deferred write verification method for hybrid disk array, and data recovery method for hybrid disk array | |
| KR102624605B1 (en) | System and method for supporting data protection across field programmable gate array solid state drives | |
| CN110187830A (en) | A kind of method and system accelerating disk array reconstruction | |
| CN105353974B (en) | A kind of two fault-tolerant coding methods for being applied to disk array and distributed memory system | |
| CN106874140B (en) | Data storage method and device | |
| CN104765660A (en) | Single-disc quick recovery method machine based on RAID6 of SSD | |
| Miyamae et al. | Erasure code with shingled local parity groups for efficient recovery from multiple disk failures | |
| CN102103533A (en) | Method for reconstructing single disk in double-disk fault-tolerance disk array | |
| CN1898650A (en) | Data storage array | |
| JP2012518231A (en) | Method and system for performing I / O operations of multiple disk arrays | |
| Dong | Coop-u: a cooperative update scheme for erasure-coded storage systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |