CN105786656B - Redundant array of independent disks disaster tolerance storage method based on random matrix - Google Patents
Redundant array of independent disks disaster tolerance storage method based on random matrix Download PDFInfo
- Publication number
- CN105786656B CN105786656B CN201610089410.7A CN201610089410A CN105786656B CN 105786656 B CN105786656 B CN 105786656B CN 201610089410 A CN201610089410 A CN 201610089410A CN 105786656 B CN105786656 B CN 105786656B
- Authority
- CN
- China
- Prior art keywords
- block
- data block
- matrix
- redundant
- failure
- 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
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/16—Error detection or correction of the data by redundancy in hardware
- G06F11/20—Error detection or correction of the data by redundancy in hardware using active fault-masking, e.g. by switching out faulty elements or by switching in spare elements
- G06F11/2053—Error detection or correction of the data by redundancy in hardware using active fault-masking, e.g. by switching out faulty elements or by switching in spare elements where persistent mass storage functionality or persistent mass storage control functionality is redundant
- G06F11/2094—Redundant storage or storage space
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A kind of redundant array of independent disks disaster tolerance storage method based on random matrix, including constructing coding parameter according to the size of file to be stored；The file to be stored is divided into k number according to block, insufficient place zero padding；Obtain one section of storage region in storage system；The storage region is divided into n memory block；The preceding k memory block of the storage region is chosen as initial data memory block, remaining n-k memory block is as redundant data storage block；The k number is filled into the preceding k initial data memory block by column according to block；The random coded matrix G that scale is n × k is constructed on GF (2) according to the coding parametern×k；According to the random coded matrix Gn×kAnd the k number generates the redundant data block of the n-k memory block according to block；The n-k redundant data block is inserted by column in the n-k redundant data storage block.The present invention improves the redundancy ability and scalability of storage system.
Description
Technical field
The present invention relates to field of data storage, in particular to a kind of redundant array of independent disks disaster tolerance based on random matrix
Storage method.
Background technique
With the continuous development of information technology and popularizing for network, data start that explosive growth is presented, and 2010 complete
Ball data volume has reached ZB magnitude.More and more data and people's lives are closely bound up, and the storage of these data with
Protection is greatly challenged, and one of the hot issue of current social discussion is had become, and urgently obtains adequate solution.
Now, main storage system has following three classes: one kind is the storage system based on duplication, typical such as Google
File system, this storage system carries out the storage and disaster tolerance of data by the way of three backups, its advantage is that easy to operate, disadvantage
It is that storage region utilization rate is lower, causes great storage region waste.Another kind of is the array code storage based on exclusive or
System, such as RAID-6, common coding method have EVENODD code, Star code, RDP code, X-Code etc., such storage system
The advantages of be using xor operation, compiling code efficiency with higher, in addition the utilization rate of storage region also has compared with copy mode
Very big raising can not arbitrary extension the disadvantage is that coding parameter is limited by prime number, it is difficult to cope under different disaster tolerance demands
Code storage.It is the code storage system based on Reed-Solomon code, such as OceanStore there are also one kind.Such storage system
The advantages of system is that the utilization rate of storage region is higher, and coding parameter has a degree of freedom in finite field scale, is lacked
Point is coding and decoding operation in finite field, and currently generally acknowledged, the computation complexity in finite field is very high, is especially advising
In the biggish situation of mould, this has seriously affected the coding and rebuilding rate of the storage system, cannot be generalized to and actually answer well
In.
In conclusion current main three classes storage system all respectively has superiority and inferiority, current extensive number cannot be met well
According to storage demand.Therefore, a kind of existing higher compiling bit rate is constructed, and storage size is completely unrestricted, simultaneously, moreover it is possible to
The High Performance Cache and Memory System for realizing the code storage under different fault-tolerant demands, is one kind that big data era people store data
Urgent need.
Summary of the invention
In order to solve the above technical problems, strong the present invention provides a kind of redundancy ability and scalability is good, high reliablity, deposits
Store up efficiency and the high redundant array of independent disks disaster tolerance storage method based on random matrix of space utilization rate.
The redundant array of independent disks disaster tolerance storage method based on random matrix that the present invention provides a kind of, comprising: according to
The size of file to be stored constructs coding parameter；The file to be stored is divided into k number according to block, insufficient place zero padding；
Obtain one section of storage region in storage system；The storage region is divided into n memory block；Choose the storage region
Preceding k memory block is as initial data memory block, remaining n-k memory block is as redundant data storage block；By the k number
It is filled by column according to block in the preceding k initial data memory block；Constructing scale on GF (2) according to the coding parameter is n
The random coded matrix G of × kn×k；According to the random coded matrix Gn×kAnd the k number generates the n-k storage according to block
The redundant data block of block；The n-k redundant data block is inserted by column in the n-k memory block.
Preferably, the storage region crosses over multiple and different physical store nodes.
Preferably, the redundant digit of the n-k memory block is generated according to block according to the random coded matrix and the k number
It include: firstly, by the random coded matrix G according to blockn×kLast n-k non-unity row in, 1 in every row corresponding to data
Block carries out exclusive or summation, to obtain the corresponding redundant block of the row；Then, by the corresponding redundant block fill to the row one
In the redundant data block of cause.
Preferably, the random coded matrix G that scale is n × k is constructed on GF (2) according to the coding parametern×kInclude:
First, it is determined that the random coded matrix Gn×kIt whether is sequency spectrum, if not then reconfiguring the random coded matrix
Gn×k, the random coded matrix G until constructing sequency spectrumn×k；Secondly, by the random coded matrix Gn×kUnit turns toFinally, willAs final random coded matrix Gn×k。
The present invention also provides a kind of redundant array of independent disks disaster tolerance reconstructing method based on random matrix, comprising: obtain
Take RECONFIGURATION REQUEST；Corresponding storage region, pre-stored random coded matrix and coding are obtained according to the RECONFIGURATION REQUEST
Parameter；Judge whether each memory block in the storage region fails, if failure is original data block, by its serial number
It is added in failure original data block sequence, what is otherwise failed is redundant block, then failure redundant data block sequence is added in its serial number
In；If there is the original data block of failure, then do not lose according to the random coded matrix, the failure original data block sequence and
The original data block of the failure is reconstructed in the memory block of effect first；If the original data block not failed, basis
The random coded matrix, the failure redundant data block sequence and all original data blocks to the redundant data block of failure into
Row reconstruct.
Preferably, if there is the data block of failure, then according to the random coded matrix, the fail data block sequence and
It includes: all data not failed chosen in storage region that the data block of the failure, which is reconstructed, in the memory block not failed
Block；Corresponding with the data block not failed row and the new matrix G ' of composition are extracted in the random coded matrixs×k；Structure
Build encoding equtions group Gs'×k*αk×1=βs×1, wherein α indicates that original data block vector, β indicate non-fail data block vector.Benefit
The data of all fail data blocks of the storage region are acquired with Gaussian reduction.
Preferably, if the original data block not failed, according to the random coded matrix, the failure redundant block
It includes: directly to utilize the random coded matrix that the redundant data block of failure, which is reconstructed, in sequence and all original data blocks
In corresponding row and all original data blocks recompile, generate all failure redundant data blocks.
The present invention realizes a kind of coding and decoding technology of novel redundant array of independent disks, greatlys improve with this solution
The redundancy ability and scalability of storage system, improve the reliability of storage system, meanwhile, the storage efficiency of storage system and
Space utilization rate is also greatly improved.
Detailed description of the invention
Fig. 1 is the overall response flow chart of the redundant array of independent disks storage and disaster tolerance method based on random matrix；
Fig. 2 is the code storage flow chart of the redundant array of independent disks storage and disaster tolerance method based on random matrix；
Fig. 3 is the random coded matrix generation figure of the redundant array of independent disks storage and disaster tolerance method based on random matrix；
Fig. 4 is the disaster tolerance reconstruct flow chart of the redundant array of independent disks storage and disaster tolerance method based on random matrix；
Fig. 5 is the redundancy block coding demonstration of the redundant array of independent disks storage and disaster tolerance method based on random matrix；
Fig. 6 is the fail data block decoding demonstration of the redundant array of independent disks storage and disaster tolerance method based on random matrix；
Fig. 7 is the demonstration of a storage region of the redundant array of independent disks storage and disaster tolerance method based on random matrix.
Specific embodiment
Some explanations first are carried out to some symbols used in the present invention and concept:
Disaster tolerance demand r: the storage block number of maximum allowable deletion in customer requirement storage system accounts for the ratio of storage region.
Coding parameter: n presentation code informational capacity, k indicate raw information amount, and δ indicates random redundancy.
Redundancy ability t: the block number of maximum allowable deletion accounts for the ratio of entire storage region in storage region.
Crash rate s: the block number that fails in storage region accounts for the ratio of entire storage region,
Storage region (Region): the storage region that file raw information and coding redundancy information to be deposited occupy in total.It deposits
Storage area domain is either on one disk, be also possible on disk array, even by the different common structures of memory node
The one section of storage region built, therefore it is a concept in logic
Block (Block): storage region is divided into n parts, and storage region occupied by every part is expressed as one piece.There are two types of forms
Block, be original data block and redundant data block respectively.
Original data block: preceding k block, the block filled by the initial data of file to be deposited in storage region.
Redundant data block: (n-k) a block after in storage region, namely the block all filled by redundant data.
Core of the invention is exactly to pass through encoder matrix G and original file data block, using XOR operation, realizes memory block
The generation of redundant block in domain, and be filled into storage region by column, thus when reconstructing fail block, it can be according to entire memory block
The failure conditions in domain, first carry out the reconstruct of failure original data block, then carry out the reconstruct of failure redundant data block, make entirely to store
The reconstruction property of system is fully played, and the utilization of computing resource is optimized, and is ensured the reliability of data to the maximum extent, is mentioned
Rise the reliability of storage system.
Fig. 1 to 7 is please referred to, the present invention provides a kind of redundant array of independent disks disaster tolerance storage side based on random matrix
Method, comprising the following steps: coding parameter is constructed according to the size of file to be stored；The file to be stored is divided into k number
According to block, insufficient place zero padding；Obtain one section of storage region in storage system；The storage region is divided into n storage
Block；The preceding k memory block of the storage region is chosen as initial data memory block, remaining n-k memory block is as redundancy
DSB data store block；The k number is filled into the preceding k initial data memory block by column according to block；Joined according to the coding
Number constructs the random coded matrix G that scale is n × k on GF (2)n×k；According to the random coded matrix Gn×kAnd the k
Data block generates the redundant data block of the n-k memory block；The n-k redundant data block is inserted the n-k by column
In memory block.Preferably, the storage region crosses over multiple and different physical store nodes.
In one embodiment, it is assumed that including 20 pieces of hard disks in a disk array, currently there is the file of a 200MB to need
Want code storage, it is desirable that the fault-tolerant ability of this document is more than 20%, and the blocks of files after also even encoding has 100, then and wherein 20
It is a and with next blocks of files failure when, still be able to obtain this document.
Therefore, settable n=280, k=200, t=60, δ=20, at this moment, the fault-tolerant ability of this document reach 21%, full
The fault-tolerant demand of foot.In this case, current embodiment require that the storage region of 280MB carries out file storage.
In order to maximumlly improve the fault-tolerant ability of file, the present embodiment chooses 14 pieces of hard disks, and the storage region is drawn
It is divided into 280 pieces, the data volume namely 20 blocks of 20MB is stored on each disk, every block size 1MB, then we are respectively every
The memory space of 20 blocks is divided on a disk, then we carry out 1 to 280 numbers to all pieces respectively.So far entire storage
Region division finishes.
Firstly, the coding and storing method of the storage system constructed according to the present invention, carries out the code storage of file to be deposited.
We construct the equally distributed random coded matrix G of 0-1 of 280 × 200 scalesn×k, and it is saved in the volume of the storage region
In code information.The file of this 200MB is divided into 200 parts by us, then chooses preceding 200 blocks of storage region as data block,
And file data blocks are filled into respectively in their preceding 200 initial data memory blocks.We are then according to random coded square
Battle array Gn×kAnd this 200 data blocks generate after 80 redundant blocks data, be finally filled into corresponding storage region.When all
After redundant block encodes filling, then process terminates.
Preferably, the redundant digit of the n-k memory block is generated according to block according to the random coded matrix and the k number
It include: firstly, by the random coded matrix G according to blockn×kLast n-k non-unity row in, 1 in every row corresponding to data
Block carries out exclusive or summation, to obtain the corresponding redundant block of the row；Then, by the corresponding redundant block fill to the row one
In the redundant data block of cause.
Preferably, the random coded matrix G that scale is n × k is constructed on GF (2) according to the coding parametern×kInclude:
First, it is determined that the random coded matrix Gn×kIt whether is sequency spectrum, if not then reconfiguring the random coded matrix
Gn×k, the random coded matrix G until constructing sequency spectrumn×k；Secondly, by the random coded matrix Gn×kUnit turns toFinally, willAs final random coded matrix Gn×k。
The present invention also provides a kind of redundant array of independent disks disaster tolerance reconstructing method based on random matrix, comprising: obtain
Take RECONFIGURATION REQUEST；Corresponding storage region, pre-stored random coded matrix and coding are obtained according to the RECONFIGURATION REQUEST
Parameter；Judge whether each memory block in the storage region fails, if failure is original data block, by its serial number
It is added in failure original data block sequence, what is otherwise failed is redundant data block, then failure redundant data block is added in its serial number
In sequence；If there is the original data block of failure, then according to the random coded matrix, the failure original data block sequence and
The original data block of the failure is reconstructed in the memory block not failed first；If the original data block not failed,
According to the random coded matrix, the failure redundant data block sequence and all original data blocks to the redundant block of failure into
Row reconstruct.
Preferably, if there is the original data block of failure, then according to the random coded matrix, the failure initial data
It includes: to choose owning in storage region that the original data block of the failure, which is reconstructed, in block sequence and the memory block not failed
The data block not failed；Extract corresponding with the data block not failed row in the random coded matrix and composition newly
Matrix G 's×k；Construct encoding equtions group Gs'×k*αk×1=βs×1, wherein α indicates that original data block vector, β indicate non-failure number
According to block vector.The data of all fail data blocks of the storage region are acquired using Gaussian reduction.
Preferably, if the original data block not failed, according to the random coded matrix, the failure redundant digit
It includes: directly to utilize the random coded matrix that the redundant block of failure, which is reconstructed, according to block sequence and all original data blocks
In corresponding row and all original data blocks recompile, generate all failure redundant blocks.
The present invention realizes a kind of coding and decoding technology of novel redundant array of independent disks, greatlys improve with this solution
The redundancy ability and scalability of storage system, enhance the reliability of storage system, meanwhile, the storage efficiency of storage system and
Space utilization rate is also greatly improved.
Below with reference to Fig. 2 to Fig. 7, the committed step in the present invention is illustrated.
Fig. 2 by the present invention redundant array of independent disks storage and disaster tolerance method of the construction based on random matrix code storage
Flow chart.The process can number file block, and be saved in k original data block of storage region respectively, produce simultaneously
Its raw corresponding redundant block data, realize the disaster tolerance guarantee to file.Wherein, encoder matrix G is utilizedn×kWith k original having been filled with
The step of beginning data block, generation residue (n-k) a redundant block, carries out as shown in Figure 5.
Fig. 5 is that the coding of the redundant array of independent disks storage and disaster tolerance method based on random matrix is demonstrated.Wherein equation is left
It is original file data block vector that Bian Yilie vertical bar, which is to be encoded, is resulting coded data block vector after coding on the right of equation,
Including original data block and redundant block two parts.As shown in figure 5, utilizing the random coded matrix G and original document of upper unit
As original data block, coding redundancy block is generated.Due to the special shape of random coded matrix, so that coding has systematic code side
The preceding k block of encoding block is original file data on the right of formula namely equation, and rear (n-k) block is coding redundancy block.
Fig. 3 by the present invention redundant array of independent disks storage and disaster tolerance method of the construction based on random matrix random coded
Matrix generates figure.The process can construct the random coded matrix G that scale on GF (2) is n × kn×k。
The disaster tolerance weight for the redundant array of independent disks storage and disaster tolerance method based on random matrix that Fig. 4 is constructed for the present invention
Structure flow chart.The process can use the block not failed in storage region, and the failure feelings of redundancy ability t are not higher than to crash rate s
Condition realizes the reconstruct of fail block.Wherein, using encoder matrix G, fail data block sequence and non-fail block, to storage region
In the process that is reconstructed of fail data block it is as shown in Figure 6.
Fig. 6 is that the fail data block decoding of the redundant array of independent disks storage and disaster tolerance method based on random matrix is demonstrated.
Assuming that we carry out the reconstruct of fail block in storage region, if all fail blocks are redundant data block, we are directly sharp
With the original data block in encoder matrix G and storage region, generation failure redundant block is recompiled.If contained in fail block
There is original data block, then we extract the corresponding row of non-fail block in encoder matrix G and form new matrix G', utilize gaussian elimination
Method finds out all failure original data blocks first, and encoder matrix G and all original data blocks is then recycled to generate all failures
Redundant block.As shown in fig. 6, deleting row corresponding to the fail block of equation the right in encoder matrix, new matrix is generated, then
Using fail data block as unknown number, and equation group, last solve system of equation are constructed using non-fail block, solution of equations is to lose
The data block data of effect.
One of the redundant array of independent disks storage and disaster tolerance method based on random matrix that Fig. 7 is constructed for the present invention deposits
The demonstration in storage area domain, wherein the right vertical line disk is redundancy disk.The storage region is individual data file in entire storage system
Required storage region can also regard a Stripe band as.
Claims (7)
1. a kind of redundant array of independent disks disaster tolerance storage method based on random matrix characterized by comprising
Coding parameter is constructed according to the size of file to be stored；
The file to be stored is divided into k number according to block, insufficient place zero padding；
Obtain one section of memory space in storage system；
The memory space is divided into n memory block；
The preceding k memory block of the memory space is chosen as initial data memory block, remaining n-k memory block is as redundancy
DSB data store block；
The k number is filled into the preceding k initial data memory block by column according to block；
The random coded matrix G that scale is n × k is constructed on GF (2) according to the coding parametern×k；
According to the random coded matrix Gn×kAnd the k number generates the redundant data block of the n-k memory block according to block；
The n-k redundant data block is inserted by column in the n-k memory block.
2. the method according to claim 1, wherein the memory space crosses over multiple and different physical store sections
Point.
3. the method according to claim 1, wherein according to the random coded matrix and the k number according to block
The redundant data block for generating the n-k memory block includes:
Firstly, by the random coded matrix Gn×kLast n-k non-unity row in, 1 in every row corresponding to data block carry out
Exclusive or summation, to obtain the corresponding redundant block of the row；
Then, by the corresponding redundant block fill to in the consistent redundant data block of the row.
4. the method according to claim 1, wherein constructing scale on GF (2) according to the coding parameter is n
The random coded matrix G of × kn×kInclude:
Firstly, constructing random matrix G on GF (2)n×k, and judge the random matrix Gn×kIt whether is sequency spectrum, if not
Then reconfigure random matrix Gn×k, the random coded matrix G until constructing sequency spectrumn×k；
Secondly, by the random coded matrix Gn×kUnit turns to
Finally, willAs final required random coded matrix Gn×k。
5. a kind of redundant array of independent disks disaster tolerance reconstructing method based on random matrix, which is characterized in that
Obtain the RECONFIGURATION REQUEST；
Corresponding memory space, pre-stored random coded matrix and coding parameter are obtained according to the RECONFIGURATION REQUEST；
Judge whether each memory block in the storage region fails, if failure is original data block, by its serial number
It is added in failure original data block sequence, what is otherwise failed is redundant data block, then failure redundant data block is added in its serial number
In sequence；
If there is the original data block of failure, then according to the random coded matrix, the failure original data block sequence and not
The data block of the failure is reconstructed in the memory block of failure first；
If the original data block not failed, according to the random coded matrix, the failure redundant data block sequence and
The redundant data block of failure is reconstructed in all original data blocks.
6. according to the method described in claim 5, it is characterized in that, if there is failure original data block, then according to it is described with
Machine encoder matrix, the failure original data block sequence and the memory block not failed carry out weight to the original data block of the failure
Structure includes:
Choose all data blocks not failed in memory space；
Corresponding with the data block not failed row and the new matrix G ' of composition are extracted in the random coded matrixs×k；
Construct encoding equtions group G 's×k*αk×1=βs×1, wherein α indicates original data block vector, β indicate non-fail data block to
Amount；
The data of all failure original data blocks of the memory space are acquired using Gaussian reduction.
7. according to the method described in claim 5, it is characterized in that, if the original data block not failed, according to
Random coded matrix, the failure redundant block sequence and all original data blocks are reconstructed the redundant block of failure and include:
It is directly recompiled using corresponding row in the random coded matrix and all original data blocks, it is superfluous to generate all failures
Remaining block.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610089410.7A CN105786656B (en) | 2016-02-17 | 2016-02-17 | Redundant array of independent disks disaster tolerance storage method based on random matrix |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610089410.7A CN105786656B (en) | 2016-02-17 | 2016-02-17 | Redundant array of independent disks disaster tolerance storage method based on random matrix |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105786656A CN105786656A (en) | 2016-07-20 |
| CN105786656B true CN105786656B (en) | 2019-08-13 |
Family
ID=56402274
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610089410.7A Active CN105786656B (en) | 2016-02-17 | 2016-02-17 | Redundant array of independent disks disaster tolerance storage method based on random matrix |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105786656B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106383669B (en) * | 2016-09-19 | 2019-04-16 | 渤海大学 | Distributed storage method and system of the one kind based on (n, k, m) coding |
| CN106339975A (en) * | 2016-10-10 | 2017-01-18 | 中科院成都信息技术股份有限公司 | Digital image secret sharing method, restoration method and devices |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6148414A (en) * | 1998-09-24 | 2000-11-14 | Seek Systems, Inc. | Methods and systems for implementing shared disk array management functions |
| US7895502B2 (en) * | 2007-01-04 | 2011-02-22 | International Business Machines Corporation | Error control coding methods for memories with subline accesses |
| CN101340193B (en) * | 2008-08-12 | 2011-04-20 | 普天信息技术研究院有限公司 | Constructing method, encoder and decoder of low density check code |
| CN101387975B (en) * | 2008-10-20 | 2010-09-15 | 中科院成都信息技术有限公司 | Magnetic disc array system |
| CN101834898B (en) * | 2010-04-29 | 2013-01-30 | 中科院成都信息技术有限公司 | Method for storing network distributed codes |
| US8799705B2 (en) * | 2012-01-04 | 2014-08-05 | Emc Corporation | Data protection in a random access disk array |
| JP2013152778A (en) * | 2013-02-28 | 2013-08-08 | Qualcomm Inc | Concurrent multiple-dimension word-addressable memory architecture |
| US20140250269A1 (en) * | 2013-03-01 | 2014-09-04 | Lsi Corporation | Declustered raid pool as backup for raid volumes |
-
2016
- 2016-02-17 CN CN201610089410.7A patent/CN105786656B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105786656A (en) | 2016-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Vajha et al. | Clay codes: Moulding {MDS} codes to yield an {MSR} code | |
| CN106100801B (en) | A kind of non-homogeneous erasure code method of cloud storage system | |
| CN105260259B (en) | A kind of locality based on system minimum memory regeneration code repairs coding method | |
| CN111149093B (en) | Data encoding, decoding and repairing method of distributed storage system | |
| Li et al. | A generic transformation for optimal repair bandwidth and rebuilding access in MDS codes | |
| CN104461781B (en) | A kind of data block method for reconstructing based on correcting and eleting codes | |
| US7934120B2 (en) | Storing data redundantly | |
| CN107844272A (en) | A kind of cross-packet coding and decoding method for improving error correcting capability | |
| US20140310571A1 (en) | Local Erasure Codes for Data Storage | |
| CN106776129B (en) | A kind of restorative procedure of the multinode data file based on minimum memory regeneration code | |
| CN1801105A (en) | System and method for tolerating multiple storage device failures in a storage system using horizontal and vertical parity layouts | |
| CN110089035B (en) | Storage controller, data processing chip and data processing method | |
| CN106484559A (en) | A kind of building method of check matrix and the building method of horizontal array correcting and eleting codes | |
| CN104052576A (en) | Data recovery method based on error correcting codes in cloud storage | |
| CN110383251A (en) | Storage system, computer-readable recording medium, and method for controlling system | |
| CN110389848B (en) | Partial repetition code construction method based on block construction and fault node repair method | |
| CN103617006A (en) | Storage resource management method and device | |
| CN102843212B (en) | Coding and decoding processing method and device | |
| CN108762978B (en) | Grouping construction method of local part repeated cyclic code | |
| CN105786656B (en) | Redundant array of independent disks disaster tolerance storage method based on random matrix | |
| CN105703782A (en) | Incremental shift matrix construction method, network coding method and system | |
| CN101901115B (en) | Method for constructing redundant array of inexpensive disks (RAID) 6 level | |
| CN107153661A (en) | A kind of storage, read method and its device of the data based on HDFS systems | |
| CN104516679A (en) | RAID (Redundant Array Of Independent Disk) data processing method and device | |
| Ivanichkina et al. | Mathematical methods and models of improving data storage reliability including those based on finite field theory |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |