CN105933386A - Construction method and device of storage system - Google Patents
Construction method and device of storage system Download PDFInfo
- Publication number
- CN105933386A CN105933386A CN201610211840.1A CN201610211840A CN105933386A CN 105933386 A CN105933386 A CN 105933386A CN 201610211840 A CN201610211840 A CN 201610211840A CN 105933386 A CN105933386 A CN 105933386A
- Authority
- CN
- China
- Prior art keywords
- node
- check
- storage system
- vector
- information
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/06—Digital input from, or digital output to, record carriers, e.g. RAID, emulated record carriers or networked record carriers
- G06F3/0601—Interfaces specially adapted for storage systems
- G06F3/0602—Interfaces specially adapted for storage systems specifically adapted to achieve a particular effect
- G06F3/0614—Improving the reliability of storage systems
- G06F3/0619—Improving the reliability of storage systems in relation to data integrity, e.g. data losses, bit errors
-
- 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
- G06F3/067—Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
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)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiments of the invention provide a construction method and device of a storage system, wherein the storage system has a dynamically adjustable characteristic and a high disaster recovery characteristic. According to the technical schemes of the invention, the method includes the following steps that: the node parameters of the storage system are determined; information nodes and check nodes are numbered sequentially; a {0, 1} random vector is generated for each of the check nodes; and encoding relations are established between each of the check nodes and the information nodes: elements in corresponding random vectors of the check nodes, of which the value is 1, are determined, information nodes corresponding to positions where the elements of which the value is 1 in the random vectors are located are determined, and exclusive OR operation is performed on the corresponding information nodes sequentially, and an operation result is adopted as the data of updated check nodes. With the construction method and device provided by the technical schemes of the invention adopted, operation cost is decreased when the storage system is constructed, and computational efficiency is improved, high reliability can be realized with low redundancy consumption. The construction method and device have excellent node dynamic adjustment capability.
Description
Technical field
The present invention relates to coding tissue and the technical field of memory of electronic information file, particularly relate to one
Plant storage system constituting method and device.
Background technology
In recent years, along with the fast development of the Internet, cloud computing, mobile terminal and Internet of Things, entirely
Ball data volume had been enter into the ZB epoch in 2010, and annual also in sustainable growth.The whole world is already
March toward the data-centered information age, and the most efficient, reliable, safe preservation data are
Become the foundation stone of information age, also become the matter of utmost importance of field of storage simultaneously.
Currently, distributed memory system just novel memory module with the Internet as medium, progressively take
Generation tradition single-point memory module.It with low cost architecture system, and can obtain highly reliable magnanimity number
According to storage effect, become each " big data " company's preferred option.
Current distributed memory system is still mainly to replicate disaster recovery method guarantee data reliability.It has
Store the advantages such as simple to operate, read operation is quick;But data redudancy is high, especially join when disaster tolerance
Number or memory data output are continuously increased, and storage system cost based on replication strategy will be continuously increased.
Defect for replication strategy High redundancy is day by day obvious, distributed memory system disaster recovery method
The most progressively transfer employing to based on Reed Solomon (RS) correcting and eleting codes strategy.RS code is due to self
Encoding characteristics can realize the demand of any disaster tolerance parameter, and amount of redundant data can be made to be far below duplication plan
Slightly.
But, the calculating of RS code is built upon on finite field, and computing cost is big, limits pushing away of it
Extensively.
Summary of the invention
The embodiment of the present invention provides one storage system constituting method and device, in order to solve existing skill
Store present in art and carry out the cost of computing between system interior joint greatly, the problem that computational efficiency is low.
The embodiment of the present invention provides one storage system constituting method, and described storage system includes information
Node and check-node, described information node is used for storing data, and described check-node is used for recovering
In data, and described storage system, the memory capacity of each node is identical, and described method includes:
Determine storage nodes parameter n, k, T}, wherein, n is the node total number of described storage system,
K is the information node number of described storage system, and n-k is the check-node number of described storage system, T
For individual element { in 0,1}, value is the probability of 1;Wherein n, k are positive integer, 0 < T < 1;
K information node is numbered in order as { D1, D2..., Dk, and by n-k check-node
Number in order as { P1, P2..., Pn-k, wherein, the primary data of n-k check-node is zero;
The each check-node being respectively directed in n-k check-node, generates { the 0,1} of a length of 1 × k
Random vector gj=(aj1, aj2..., ajk), 1≤j≤n-k, a ∈ { 0,1}, gjFor check-node PjCorresponding
Random vector, wherein, each element a in vectorjiWith probability T value for 1;
The each check-node being respectively directed in n-k check-node, execution following steps:
Determine check-node PjCorresponding vector gjIntermediate value is each element a of 1ji;
Determine with value be respectively 1 each element ajiAt vector gjCorresponding each in middle position
Information node Di;
By check-node PjEach information node D with described correspondenceiCarry out XOR successively,
Result is as the P after updatingj。
Further, also include:
I-th information node D in described storage systemiStorage data change Δ DiAfter, pin
To each check-node P in n-k check-nodejIf, this check-node PjCorresponding vector gj's
I-th element ajiValue be 1, then by this check-node PjVariation delta D with this information nodeiEnter
Row XOR, result is as the P after updatingj, otherwise, keep check-node PjConstant.
Further, also include:
When needing to add l the information node that memory capacity is identical within the storage system
{Dk+1, Dk+2..., Dk+lTime, each check-node being respectively directed in n-k check-node, generate
{ 0, the 1} random vector h of a length of 1 × lj=(bj1, bj2..., bjl), 1≤j≤n-k, b ∈ { 0,1}, hjFor school
Test node PjCorresponding random vector, wherein, each element b in vectorjiWith probability T value for 1;
The each check-node being respectively directed in n-k check-node, execution following steps:
Determine check-node PjCorresponding vector hjIntermediate value is each element b of 1ji;
Determine with value be respectively 1 each element bjiAt vector hjCorresponding each in middle position
Information node Dk+i;
By check-node PjEach information node D with described correspondencek+iCarry out XOR successively,
Result is as the P after updatingj;
By l information node { Dk+1, Dk+2..., Dk+lAdd in described storage system, and respectively
By random vector hjCorrespondence adds random vector g tojAfterbody, the random vector g after being updatedj。
Further, at the described each check-node being respectively directed in n-k check-node, hold
Before row following steps, also include:
Determine l the information node { D that will addk+1, Dk+2..., Dk+lThe most all do not store data,
If it is not, then each check-node being respectively directed to described in Zhi Hanging in n-k check-node, hold
The step for of row following steps;And
If it is, directly by l information node { Dk+1, Dk+2..., Dk+lAdd described storage system to
In system, and respectively by random vector hjCorrespondence adds random vector g tojAfterbody, updated
After random vector gj。
Further, also include:
When l information node { C deleted within the storage system by needs1, C2..., ClTime, determine and treat
L the information node { C deleted1, C2..., ClNumbering in information node set;
The each check-node being respectively directed in n-k check-node, generates { the 0,1} of a length of 1 × l
Vector fj=(cj1, cj2..., cjl), 1≤j≤n-k, fjFor check-node PjCorresponding vector, wherein,
cjiFor information node CiNumbering in information node set is at gjThe element of middle correspondence position;
The each check-node being respectively directed in n-k check-node, execution following steps:
Determine check-node PjCorresponding vector fjIntermediate value is each element c of 1ji;
Determine with value be respectively 1 each element cjiIn vector fjCorresponding each in middle position
Information node Ci;
By check-node PjEach information node C with described correspondenceiCarry out XOR successively,
Result is as the P after updatingj;
By l information node { C1, C2..., ClFrom described storage system delete, and respectively from
Machine vector gjMiddle deletion each information node { C1, C2..., ClThe element of numbering correspondence position, obtain more
Random vector g after Xinj。
Further, each check-node in being respectively directed to n-k check-node, generate long
Degree is { 0, the 1} vector f of 1 × lj=(cj1, cj2..., cjlBefore), also include:
Determine l the information node { C that will delete1, C2..., ClThe most all do not store data, if
, then do not perform each check-node being respectively directed in n-k check-node, generate a length of
{ 0, the 1} vector f of 1 × lj=(cj1, cj2..., cjl) the step for;And
If it is, directly by l information node { C1, C2..., ClDelete from described storage system,
And respectively from random vector gjMiddle deletion each information node { C1, C2..., ClNumbering correspondence position
Element, the random vector g after being updatedj。
Further, also include:
When needing to add l the check-node that memory capacity is identical within the storage system
{Pn-k+1, Pn-k+2..., Pn-k+lTime, each check-node being respectively directed in l check-node, generate
{ 0, the 1} random vector g of a length of 1 × kn-k+j=(aN-k+j, 1, aN-k+j, 2..., aN-k+j, k), 1≤j≤l,
A ∈ { 0,1}, gn-k+jFor check-node Pn-k+jCorresponding random vector, wherein, each element a in vectorN-+j, i
With probability T value for 1;
The each check-node being respectively directed in l check-node, execution following steps:
Determine check-node Pn-k+jCorresponding vector gn-k+jIntermediate value is each element a of 1N-k+j, i;
Determine with value be respectively 1 each element aN-k+j, iAt vector gn-k+jMiddle position pair
The each information node D answeredi;
By check-node Pn-k+jEach information node D with described correspondenceiCarry out XOR successively, knot
Fruit is as the P after updatingn-k+j。
Further, also include:
Determine check-node to be deleted, from described storage system, delete described verification to be deleted
Node.
Further, described storage system also has disaster tolerance parameter { t, δ }, and wherein, t+ δ is described
The check-node number of storage system, described storage system allows that at most t node is complete and makes mistakes, and δ is institute
State storage system and ensure additional nodes number needed for high redundancy ability, and t+ δ < k, t+ δ+k=n, institute
State method, also include:
When the storage loss of data of l information node of described storage system, from described storage system
Remaining n-l node is arbitrarily chosen δ node of k+, and including the individual check-node of δ 'The individual information node of k+ δ-δ ', l≤t;
By each for individual for δ ' check-node self-corresponding 1 × k 0,1} random vector take out, be arranged as δ ' × k's
{ 0,1} matrix Rδ′×k, wherein matrix jth V&V of behavior node1 corresponding × k 0,1} random vector,
1≤j≤δ′;
Structural matrix Hδ′×(k+δ′), matrix Hδ′×(k+δ′)Left part is divided into matrix Rδ′×k, right part is divided into list
Position battle array Iδ′×δ′;
Based on the known k+ individual information node of δ-δ ' and the individual check-node of δ ', structure vector β(k+δ′)×1,
Wherein β(k+δ′)×1K position is information node { D from left to right1, D2..., Dk, wherein unknown information node becomes
Amount X replaces, and right-hand member is
For relational expression Hδ′×(k+δ′)·β(k+δ′)×1=0 solves, and obtains l the information node lost
Storage data;
By in the data recovery storage of l information node that obtains to described storage system.
The embodiment of the present invention also provides for a kind of storage system constructing device, it is characterised in that described in deposit
Storage system includes that information node and check-node, described information node are used for storing data, described school
Test node for recovering data, and in described storage system, the memory capacity of each node is identical, described
Device includes:
Parameter determination unit, is used for determining that { n, k, T}, wherein, n is described to storage nodes parameter
The node total number of storage system, k is the information node number of described storage system, and T is that individual element exists
{ in 0,1}, value is the probability of 1;Wherein n, k are positive integer, 0 < T < 1;
Numbered cell, for numbering in order k information node as { D1, D2..., Dk, and will
N-k check-node numbers in order as { P1, P2..., Pn-k, wherein, at the beginning of n-k check-node
Beginning data are zero;
Vector signal generating unit, for each check-node being respectively directed in n-k check-node,
Generate { 0, the 1} random vector g of a length of 1 × kj=(aj1, aj2..., ajk), 1≤j≤n-k, a ∈ 0,1},
gjFor check-node PjCorresponding random vector, wherein, each element a in vectorjiWith probability T value
It is 1;
Encoding relation sets up unit, for each verification joint being respectively directed in n-k check-node
Point, execution following steps:
Determine check-node PjCorresponding vector gjIntermediate value is each element a of 1ji;
Determine with value be respectively 1 each element ajiAt vector gjCorresponding each in middle position
Information node Di;
By check-node PjEach information node D with described correspondenceiCarry out XOR, result successively
As the P after updatingj。
Beneficial effect of the present invention includes:
Use the above-mentioned storage system constituting method that the embodiment of the present invention provides, build storage system
During, it is only necessary to system carries out between each node simple XOR, thus drops
The low cost of computing, improves computational efficiency.
Other features and advantage will illustrate in the following description, and, partly
Become apparent from description, or understand by implementing the application.The purpose of the application
Can be by being referred in particular in the description write, claims and accompanying drawing with other advantages
The structure gone out realizes and obtains.
Accompanying drawing explanation
Accompanying drawing is for providing a further understanding of the present invention, and constitutes a part for description,
It is used for explaining the present invention together with the embodiment of the present invention, is not intended that limitation of the present invention.Attached
In figure:
The Organization Chart of the storage system that Fig. 1 provides for the embodiment of the present invention;
The flow chart of the storage system constituting method that Fig. 2 provides for the embodiment of the present invention;
Fig. 3 is the schematic diagram building storage system in the embodiment of the present invention;
The flow chart adding information node within the storage system that Fig. 4 provides for the embodiment of the present invention;
Fig. 5 is the schematic diagram adding information node in the embodiment of the present invention within the storage system;
The flow chart deleting information node from storage system that Fig. 6 provides for the embodiment of the present invention;
Fig. 7 is the schematic diagram deleting information node in the embodiment of the present invention from storage system;
The flow chart adding check-node within the storage system that Fig. 8 provides for the embodiment of the present invention;
Fig. 9 is the schematic diagram adding check-node in the embodiment of the present invention from storage system;
Figure 10 is to recover the flow chart of the storage data of loss in storage system in the embodiment of the present invention;
The structural representation of the storage system constructing device that Figure 11 provides for the embodiment of the present invention.
Detailed description of the invention
In order to provide the computing cost reducing in storage system between each node, and improve and calculate effect
The implementation of rate, embodiments provides a kind of storage system constituting method and device, with
The preferred embodiments of the present invention are illustrated by lower combination Figure of description, it will be appreciated that this place
The preferred embodiment described is merely to illustrate and explains the present invention, is not intended to limit the present invention.And
And in the case of not conflicting, the embodiment in the application and the feature in embodiment can mutual groups
Close.
The embodiment of the present invention provides one storage system constituting method, as it is shown in figure 1, constructed
This storage system includes information node and check-node, and information node is used for storing data, verification joint
It is identical that point is used for recovering the memory capacity of each node in data, and storage system.
Further, this storage system is the most unrestricted to storage medium or storage tool in application process,
The information node of storage data can be the most independent frame of sector, USB flash disk, hard disk etc..
Further, in this storage system, each node can be connected by large-scale the Internet, also may be used
Being connected by the LAN of little scope, the network type connecting each node does not limit.Further, deposit
In storage system, each node can be positioned over different geographical, also can be the diverse location in same frame, with
Sample does not limit.
As in figure 2 it is shown, the embodiment of the present invention provide storage system constituting method, can include as
Lower step:
Step 21, determine that { n, k, T}, wherein, n is the joint of storage system to storage nodes parameter
Point sum, k is the information node number of storage system, and n-k is the check-node number of storage system, T
For individual element { in 0,1}, value is the probability of 1;Wherein n, k are positive integer, 0 < T < 1.
Step 22, k information node is numbered in order as { D1, D2..., Dk, and by n-k
Check-node numbers in order as { P1, P2..., Pn-k, wherein, the primary data of n-k check-node
It is zero.
Step 23, each check-node being respectively directed in n-k check-node, generate a length of
{ 0, the 1} random vector g of 1 × kj=(aj1, aj2..., ajk), 1≤j≤n-k, a ∈ { 0,1}, gjFor verification joint
Point PjCorresponding random vector, wherein, each element a in vectorjiWith probability T value for 1.
Step 24, each check-node being respectively directed in n-k check-node, with information node
Set up encoding relation, specifically can perform following steps:
Step A1, determine check-node PjCorresponding vector gjIntermediate value is each element a of 1ji;
Step B1, determine with value be respectively 1 each element ajiAt vector gjMiddle position is corresponding
Each information node Di;
Step C1, by check-node PjEach information node D with this correspondenceiCarry out XOR fortune successively
Calculating, result is as the P after updatingj。
Concrete, as a example by the framework storing system shown in Fig. 3, to the storage shown in above-mentioned Fig. 2
System constituting method is described in detail.
Where it is assumed that storage systematic parameter index is that { n=10, k=6} wherein have 6 information
Node and 4 check-nodes.
According to above-mentioned steps 22,6 information nodes are numbered in order as { D1, D2, D3, D4, D5, D6,
And 4 check-nodes are numbered in order as { P1, P2, P3, P4, wherein, at the beginning of 4 check-nodes
Beginning data are zero.
According to above-mentioned steps 23, it is respectively directed to each check-node in 4 check-nodes, generates
{ 0, the 1} random vector g of a length of 1 × 6j=(aj1, aj2..., aj6), 1≤j≤4, a ∈ { 0,1}, gjFor school
Test node PjCorresponding random vector, wherein, each element a in vectorjiWith probability T value for 1.
Such as, concrete, as shown in Figure 3, g1=(1,0,1,1,1,0), g2=(0,1,1,0,0,0),
g3=(0,1,0,0,1,0), g4=(1,1,0,0,1,1).
Then according to above-mentioned steps 24, it is respectively directed to each check-node in 4 check-nodes,
Encoding relation is set up with information node, concrete, below with check-node P1As a example by be described:
According to step A1, check-node P1Corresponding random vector g1=(1,0,1,1,1,0), its intermediate value is 1
Each element include the 1st, 3,4,5 elements;
According to step B1, with value be respectively 1 each element at vector g1Corresponding each in middle position
Information node includes: D1、D3、D4And D5;
According to step C1, by check-node P1Each information node D with this correspondence1、D3、D4、D5
Carry out XOR successively, i.e.Initial due to each check-node
Data are zero, so, check-node P1Final operation result can also be expressed asResult is as the P after updating1, thus complete for P1Coding close
The foundation of system.
In the embodiment of the present invention, the XOR between node can be identical storage position in two nodes
The nodulo-2 addition of the binary bits symbol put.
Identical method is used to complete for P2、P3And P4The foundation of encoding relation, thus complete
The structure of this storage system.
After completing the structure of this storage system, as the i-th information node D of this storage systemi's
Storage data change Δ DiAfter, for each check-node P in n-k check-nodejIf,
This check-node PjCorresponding vector gjI-th element ajiValue be 1, then can be by this check-node
PjWith this information node variation delta DiCarry out XOR, i.e.Result is as more
P after Xinj, thus complete the renewal of nodes encoding relation, otherwise, keep check-node PjConstant.
Such as, for the storage system shown in Fig. 3, if data to be stored in information node D2, then pin
To each check-node { P1, P2, P3, P4Corresponding random vector, second element value be 1 random to
Amount includes g2、g3And g4, the most accordingly, for check-node P2、P3And P4It is required to carry out more
Newly, concrete renewal process refers to foregoing, is no longer described in detail at this.
Use the above-mentioned storage system constituting method that the embodiment of the present invention provides, build storage system
During, it is only necessary to system carries out between each node simple XOR, thus drops
The low cost of computing, improves computational efficiency.
In the embodiment of the present invention, for the above-mentioned storage system built, it is also possible to according to actual needs
Information node quantity therein is adjusted, including adding information node and deleting information node,
It is described in detail below in conjunction with the accompanying drawings.
When needs add information node within the storage system, as shown in Figure 4, can include as follows
Step:
Step 41, determine l information joint that the memory capacity needing to add within the storage system is identical
Point { Dk+1, Dk+2..., Dk+l}。
Step 42, each check-node being respectively directed in n-k check-node, generate a length of
{ 0, the 1} random vector h of 1 × lj=(bj1, bj2..., bjl), 1≤j≤n-k, b ∈ { 0,1}, hjFor check-node
PjCorresponding random vector, wherein, each element b in vectorjiWith probability T value for 1.
Step 43, each check-node being respectively directed in n-k check-node, with new information
Node sets up encoding relation, specifically can perform following steps:
Step A2, determine check-node PjCorresponding vector hjIntermediate value is each element b of 1ji;
Step B2, determine with value be respectively 1 each element bjiAt vector hjMiddle position is corresponding
Each information node Dk+i;
Step C2, by check-node PjEach information node D with this correspondencek+iCarry out XOR fortune successively
Calculating, result is as the P after updatingj。
Step 44, by l information node { Dk+1, Dk+2..., Dk+lAdd in this storage system, with
And respectively by random vector hjCorrespondence adds random vector g tojAfterbody, random after being updated
Vector gj。
Concrete, still as a example by the framework storing system shown in Fig. 3, to shown in above-mentioned Fig. 4
The method adding information node within the storage system is described in detail.
On the basis of storage system shown in Fig. 3, need to add 3 information node { D7, D8, D9}。
According to above-mentioned steps 42, it is respectively directed to each check-node in 4 check-nodes, generates
{ 0, the 1} random vector h of a length of 1 × 3j=(bj1, bj2, bj3), 1≤j≤4, b ∈ { 0,1}, hjFor verification joint
Point PjCorresponding random vector, wherein, each element b in vectorjiWith probability T value for 1.
Such as, concrete, as shown in Figure 5, h1=(1,0,0), h2=(0,1,1), h3=(0,1,0),
h4=(1,1,1).
According to above-mentioned steps 43, it is respectively directed to each check-node in 4 check-nodes, with newly
Information node set up encoding relation, concrete, below with check-node P1As a example by be described:
According to step A2, check-node P1Corresponding random vector h1=(1,0,0), its intermediate value is 1
Each element includes the 1st element;
According to step B2, with value be respectively 1 each element at vector h1Corresponding each in middle position
Information node includes: D7;
According to step C2, by check-node P1Information node D with this correspondence7Carry out XOR,
I.e.Based in flow process shown in above-mentioned Fig. 2In this step,
Check-node P1Final operation result can also be expressed asResult
As the P after updating1, thus complete for P1The renewal of encoding relation.
Then according to above-mentioned steps 44, as it is shown in figure 5, by 3 information node { D7, D8, D9Add
In this storage system, and respectively by random vector hjCorrespondence adds random vector g tojAfterbody,
Random vector g after being updatedj, after i.e. updating, it is respectively as follows: g1=(1,0,1,1,1,0,1,0,0),
g2=(0,1,1,0,0,0,0,1,1), g3=(0,1,0,0,1,0,0,1,0), g4=(1,1,0,0,1,1,1,1,1).
In the embodiment of the present invention, when adding information node in storage system, if will add
L information node { Dk+1, Dk+2..., Dk+lAll do not store data, due to information node to be added not
Affect the original internodal encoding relation of storage system, it is possible to do not perform above-mentioned steps 43,
Directly by l information node { Dk+1, Dk+2..., Dk+lAdd in storage system, and respectively will be with
Machine vector hjCorrespondence adds random vector g tojAfterbody, the random vector g after being updatedj。
When needs delete information node from storage system, as shown in Figure 6, can include as follows
Step:
Step 61, from storage system, delete l information node { C when needs1, C2..., ClTime, really
Fixed l information node { C to be deleted1, C2..., ClNumbering in information node set.
Step 62, each check-node being respectively directed in n-k check-node, generate a length of
{ 0, the 1} vector f of 1 × lj=(cf1, cj2..., cjl), 1≤j≤n-k, fjFor check-node PjCorresponding to
Amount, wherein, cjiFor information node CiNumbering in information node set is at gjMiddle correspondence position
Element.
Step 63, each check-node being respectively directed in n-k check-node, update and information
The encoding relation of node, specifically can perform following steps:
Step A3, determine check-node PjCorresponding vector fjIntermediate value is each element c of 1ji;
Step B3, determine with value be respectively 1 each element cjiIn vector fjMiddle position is corresponding
Each information node Ci;
Step C3, by check-node PjEach information node C with this correspondenceiCarry out XOR fortune successively
Calculating, result is as the P after updatingj。
Step 64, by l information node { C1, C2..., ClDelete from storage system, and respectively
From random vector gjMiddle deletion each information node { C1, C2..., ClThe element of numbering correspondence position,
Random vector g after renewalj。
Concrete, still as a example by the framework storing system shown in Fig. 3, to shown in above-mentioned Fig. 6
The method deleting information node from storage system is described in detail.
On the basis of storage system shown in Fig. 3, need to delete 2 information node { D1, D4, right
Should be { C1, C2}。
According to above-mentioned steps 62, it is respectively directed to each check-node in 4 check-nodes, generates
{ 0, the 1} vector f of a length of 1 × 2j=(cj1, cj2), 1≤j≤4, fjFor check-node PjCorresponding vector,
Wherein, cjiFor information node CiNumbering in information node set is at gjThe element of middle correspondence position.
Such as, concrete, as shown in Figure 7, { D1, D4Numbering in information node set divides
It is not 1 and 4, the most respectively from check-node { P1, P2, P3, P41 × 6 { 0,1} random vector g1、g2、
g3And g4In extract the 1st, 4 bits element, obtain f1=(1,1), f2=(0,0), f3=(0,0), f4=(1,0).
According to above-mentioned steps 63, it is respectively directed to each check-node in 4 check-nodes, updates
With the encoding relation of information node, concrete, below with check-node P1As a example by be described:
According to step A3, check-node P1Corresponding vector f1=(1,1), its intermediate value is each element of 1
Including the 1st, 2 elements;
According to step B3, it is each element c of 1 respectively with valuejiIn vector fjMiddle position is corresponding
Each information node includes: { C1, C2, correspond to { D1, D4};
According to step C3, by check-node P1Each information node { C with this correspondence1, C2Carry out successively
XOR, i.e. with information node { D1, D4Carry out XOR successively, i.e.
Based in flow process shown in above-mentioned Fig. 2In this step, check-node P1's
Final operation result can also be expressed asResult
As the P after updating1, thus complete for P1The renewal of encoding relation.
Then according to above-mentioned steps 64, as it is shown in fig. 7, by 2 information node { C1, C2, i.e.
{D1, D4Delete from storage system, and respectively from random vector gjThe each information node of middle deletion
{C1, C2The element of numbering correspondence position in information node set, after being updated random to
Amount gj, after i.e. updating, it is respectively as follows: g1=(0,1,1,0), g2=(1,1,0,0), g3=(1,0,1,0), g4=(1,0,1,1).
In the embodiment of the present invention, when deleting information node from storage system, if will delete
L information node { C1, C2..., ClAll do not store data, owing to information node to be deleted does not affects
The original internodal encoding relation of storage system, it is possible to do not perform above-mentioned steps 62 and step
63, directly by l information node { C1, C2..., ClFrom storage system delete, and respectively from
Machine vector gjMiddle deletion each information node { C1, C2..., ClNumbering in information node set is corresponding
The element of position, the random vector g after being updatedj。
At present, in prior art, there is part distributed memory system application RAID (Redundant
Arrays of Independent Disks, disk array) coding techniques, it can realize each storage data
Direct xor operation so that amount of redundant data be far below replication strategy.But, due to RAID
Coding techniques limits, and systematic parameter is the most limited, is not appropriate for what data storage size constantly expanded
Trend.
On the other hand, it is big that big data age not only stores data volume radix, and it also presents quick-fried simultaneously
Fried formula growing trend.But, existing distributed storage method based on coded system, systematic parameter
Adjustability is poor, once it is determined that, it is difficult to change;Data storage method is complicated, by first piecemeal re-encoding
Rear storage mode so that storage system scale autgmentability is poor, causes adapting to data to be stored
The most quickly increase.
And based on the embodiment of the present invention provide storage system, use shown in above-mentioned Fig. 4 to storage
System is added the flow process of information node, and deleting from storage system shown in above-mentioned Fig. 6 is believed
The flow process of breath node, can quickly be extended storage system according to the actual requirements flexibly,
Thus preferably adapt to the most quickly increasing of data to be stored.
In the embodiment of the present invention, for the above-mentioned storage system built, it is also possible to according to actual needs
Check-node quantity therein is adjusted, including adding check-node and deleting check-node,
Thus realize the adjustment of the redundancy ability to storage system, it is described in detail below in conjunction with the accompanying drawings.
When needs add check-node within the storage system, as shown in Figure 8, can include as follows
Step:
Step 81, determine and need to add l the check-node that memory capacity is identical within the storage system
{Pn-k+1, Pn-k+2..., Pn-k+l}。
Step 82, each check-node being respectively directed in l check-node, generate a length of 1 × k
{ 0,1} random vector gn-k+j=(aN-k+j, 1, aN-k+j, 2..., aN-k+j, k), 1≤j≤l, a ∈ { 0,1}, gn-k+jFor
Check-node Pn-k+jCorresponding random vector, wherein, each element a in vectorN-+j, iWith probability T value
It is 1.
Step 83, each check-node being respectively directed in l check-node, build with information node
Vertical encoding relation, specifically can perform following steps:
Step A4, determine check-node Pn-k+jCorresponding vector gn-k+jIntermediate value is each element of 1
aNk+j, i;
Step B4, determine with value be respectively 1 each element aN-k+j, iAt vector gn-k+jMiddle in place
Put each information node D of correspondencei;
Step C4, by check-node Pn-k+jEach information node D with this correspondenceiCarry out XOR fortune successively
Calculating, result is as the P after updatingn-k+j。
Concrete, on the basis of storage system shown in Fig. 3, storage system will newly-increased 2 verifications
Node { P5, P6}。
According to above-mentioned steps 82, it is respectively directed to each check-node in 2 check-nodes, generates
{ 0, the 1} random vector g of a length of 1 × 6j=(aj1, aj2..., aj6), 5≤j≤6, a ∈ { 0,1}, gjFor school
Test node PjCorresponding random vector, wherein, each element a in vectorjiWith probability T value for 1.
Such as, concrete, as shown in Figure 9, g5=(0,1,0,1,0,1), g6=(1,1,0,0,0,1).
Then according to above-mentioned steps 83, it is respectively directed to each check-node in 2 check-nodes,
Encoding relation is set up with information node, concrete, below with check-node P5As a example by be described:
According to step A4, check-node P5Corresponding random vector g5=(0,1,0,1,0,1), its intermediate value is
Each element of 1 includes the 2nd, 4,6 elements;
According to step B4, with value be respectively 1 each element at vector g5Corresponding each in middle position
Information node includes: D2、D4And D6;
According to step C4, by check-node P5Each information node D with this correspondence2、D4、D6Depend on
The secondary XOR that carries out, i.e.At the beginning of 2 newly-increased check-nodes
Beginning data are zero, so, check-node P5Final operation result can also be expressed asResult is as the P after updating5, thus complete for P5Encoding relation
Set up.
Adopt in the same way for P6Setting up encoding relation, result is:
In the embodiment of the present invention, when check-node deleted by needs, determine check-node to be deleted,
Then from storage system, delete this check-node to be deleted, no longer illustrate at this and retouch
State.
In the embodiment of the present invention, above-mentioned storage system can also have a disaster tolerance parameter { t, δ }, wherein,
T+ δ is the check-node number of storage system, and storage system is allowed that at most t node is complete and made mistakes, and δ is
Storage system ensures additional nodes number needed for high redundancy ability, and t+ δ < k, t+ δ+k=n.
Based on above-mentioned disaster tolerance parameter { t, δ }, when the information node storing loss of data in storage system
When quantity is not more than t, the storage data of loss can be recovered, below in conjunction with the accompanying drawings, to recovery
The handling process losing data is described in detail, and as shown in Figure 10, specifically can include walking as follows
Rapid:
Step 101, when the storage loss of data of l information node of the system of storage, determine loss
L information node numbering in information node set, l≤t.
Step 102, from storage system spare n-l node arbitrarily choose δ node of k+, and
Including the individual check-node of δ 'The individual information node of k+ δ-δ '.
Step 103, by each for individual for δ ' check-node self-corresponding 1 × k 0,1} random vector take out, row
It is classified as { 0, the 1} matrix R of δ ' × kδ′×k, wherein matrix jth V&V of behavior node1 corresponding × k's
{ 0,1} random vector, 1≤j≤δ '.
Step 104, structural matrix Hδ′×(k+δ′), matrix Hδ′×(k+δ′)Left part is divided into matrix Rδ′×k, right
Part is unit battle array Iδ′×δ′。
Step 105, based on the known k+ individual information node of δ-δ ' and the individual check-node of δ ', construct to
Amount β(k+δ′)×1, wherein β(k+δ′)×1K position is information node { D from left to right1, D2..., Dk, wherein unknown
Information node variable X replaces, and right-hand member is
Step 106, for relational expression Hδ′×(k+δ′)·β(k+δ′)×1=0 solves, and obtains the l lost
The storage data of individual information node.
Step 107, by the data recovery storage of l information node that obtains to described storage system.
Specifically, on the basis of storage system shown in Fig. 3, it is known that each ginseng of current storage system
Number is for { n=10, k=6, δ=1, t=3}, wherein, have k=6 information node, and δ+t=4 verification saves
Point, it is ensured that arbitrarily t=3 node damages, δ=1 is to ensure the necessary redundancy joint that high probability recovers
Count.
Assume information node { D1, D5The loss of data that stores or damage, according to step 102, from surplus
8 remaining nodes take out 7 nodes, such as, is respectively { D2, D3, D4, P1, P2, P3, P4}。
According to step 103, by 4 check-node { P1, P2, P3, P40,1} random vector take out, structure
Make matrixWherein matrix jth V&V of behavior nodeCorresponding
{ 0,1} random vector, 1≤j≤4 of 1 × k.
According to step 104, structural matrix
According to step 105, based on known 3 information nodes and 4 check-nodes
{D2, D3, D4, P1, P2, P3, P4, structure vector
β(k+δ′)×1=β10×1={ X1, D2, D3, D4, X2, X3, P1, P2, P3, P4, wherein, { X1, X2Corresponding loss
{D1, D5, X3Corresponding D6;Although D6Storage data are not lost, but D6The most selected carry out extensive
Multiple operation, its value is the most unknown.
According to step 106, H4×10With β10×1Meet relation H4×10)·β10×1=0, carry out for this relation
Solve, both can obtain 2 the information node { D lost1, D5Storage data, and then will obtain
2 information node { D1, D5Data recovery storage in storage system.
In the embodiment of the present invention, for the relational expression in above-mentioned steps 106
Hδ′×(k+δ′)·β(k+δ′)×1=0, as T=0.5, there is the probability of solution more than 1-2δ, so, work as T=0.5,
During δ=20, can guarantee that the probability higher than 0.999999 has solution, along with δ constantly increases, have the general of solution
Rate the most constantly increases, so, it is also preferred that the left as T=0.5, δ=20.
Further, in actual applications, when k, t value is much larger than δ, δ redundancy check node
Memory space, for the most whole storage system, be negligible.
Based on same inventive concept, the storage system constructing side provided according to the above embodiment of the present invention
Method, correspondingly, another embodiment of the present invention additionally provides a kind of storage system constructing device, its knot
As shown in figure 11, described storage system includes information node and check-node, described letter to structure schematic diagram
Breath node is used for storing data, and described check-node is used for recovering in data, and described storage system
The memory capacity of each node is identical, and described device includes:
Parameter determination unit 111, is used for determining that { n, k, T}, wherein, n is storage nodes parameter
The node total number of described storage system, k is the information node number of described storage system, and T is single unit
Element is { in 0,1}, value is the probability of 1;Wherein n, k are positive integer, 0 < T < 1;
Numbered cell 112, for numbering in order k information node as { D1, D2..., Dk, with
And n-k check-node is numbered in order as { P1, P2..., Pn-k, wherein, n-k check-node
Primary data be zero;
Vector signal generating unit 113, for each verification joint being respectively directed in n-k check-node
Point, generates { 0, the 1} random vector g of a length of 1 × kj=(aj1, aj2..., ajk), 1≤j≤n-k,
A ∈ { 0,1}, gjFor check-node PjCorresponding random vector, wherein, each element a in vectorjiWith generally
Rate T value is 1;
Encoding relation sets up unit 114, for each verification being respectively directed in n-k check-node
Node, execution following steps:
Determine check-node PjCorresponding vector gjIntermediate value is each element a of 1ji;
Determine with value be respectively 1 each element ajiAt vector gjCorresponding each in middle position
Information node Di;
By check-node PjEach information node D with described correspondenceiCarry out XOR, result successively
As the P after updatingj。
Further, the function of above-mentioned each unit may correspond to the phase in flow process shown in Fig. 1 to Figure 10
Step should be processed, do not repeat them here.
The storage system constructing device that embodiments herein is provided can be real by computer program
Existing.Those skilled in the art are it should be appreciated that above-mentioned Module Division mode is only numerous module
One in dividing mode, if being divided into other modules or not dividing module, if storage system
Construction device has above-mentioned functions, all should be within the protection domain of the application.
The application is with reference to method, equipment (system) and the computer journey according to the embodiment of the present application
The flow chart of sequence product and/or block diagram describe.It should be understood that can be real by computer program instructions
Show each flow process in flow chart and/or block diagram and/or square frame and flow chart and/or side
Flow process in block diagram and/or the combination of square frame.These computer program instructions can be provided to general meter
The processor of calculation machine, special-purpose computer, Embedded Processor or other programmable data processing device
To produce a machine so that by computer or the processor of other programmable data processing device
The instruction performed produces for realizing at one flow process of flow chart or multiple flow process and/or block diagram one
The device of the function specified in individual square frame or multiple square frame.
These computer program instructions may be alternatively stored in and can guide at computer or other programmable datas
In the computer-readable memory that reason equipment works in a specific way so that being stored in this computer can
The instruction read in memorizer produces the manufacture including command device, and this command device realizes in flow process
The merit specified in one flow process of figure or multiple flow process and/or one square frame of block diagram or multiple square frame
Energy.
These computer program instructions also can be loaded into computer or other programmable data processing device
On so that on computer or other programmable devices, perform sequence of operations step to produce calculating
The process that machine realizes, thus the instruction performed on computer or other programmable devices provides and is used for
Realize in one flow process of flow chart or multiple flow process and/or one square frame of block diagram or multiple square frame
The step of the function specified.
Obviously, those skilled in the art can carry out various change and modification to the present invention and not take off
From the spirit and scope of the present invention.So, if these amendments of the present invention and modification belong to this
Within the scope of bright claim and equivalent technologies thereof, then the present invention be also intended to comprise these change and
Including modification.
Claims (10)
1. a storage system constituting method, it is characterised in that described storage system includes that information saves
Point and check-node, described information node is used for storing data, and described check-node is used for recovering number
According to, and in described storage system, the memory capacity of each node is identical, described method includes:
Determine storage nodes parameter n, k, T}, wherein, n is the node total number of described storage system,
K is the information node number of described storage system, and n-k is the check-node number of described storage system, T
For individual element { in 0,1}, value is the probability of 1;Wherein n, k are positive integer, 0 < T < 1;
K information node is numbered in order as { D1,D2,…,Dk, and by n-k check-node
Number in order as { P1,P2,…,Pn-k, wherein, the primary data of n-k check-node is zero;
The each check-node being respectively directed in n-k check-node, generates { the 0,1} of a length of 1 × k
Random vector gj=(aj1,aj2,…,ajk), 1≤j≤n-k, a ∈ { 0,1}, gjFor check-node PjCorresponding
Random vector, wherein, each element a in vectorjiWith probability T value for 1;
The each check-node being respectively directed in n-k check-node, execution following steps:
Determine check-node PjCorresponding vector gjIntermediate value is each element a of 1ji;
Determine with value be respectively 1 each element ajiAt vector gjCorresponding each in middle position
Information node Di;
By check-node PjEach information node D with described correspondenceiCarry out XOR successively,
Result is as the P after updatingj。
2. the method for claim 1, it is characterised in that also include:
I-th information node D in described storage systemiStorage data change Δ DiAfter, pin
To each check-node P in n-k check-nodejIf, this check-node PjCorresponding vector gj's
I-th element ajiValue be 1, then by this check-node PjVariation delta D with this information nodeiEnter
Row XOR, result is as the P after updatingj, otherwise, keep check-node PjConstant.
3. the method for claim 1, it is characterised in that also include:
When needing to add l the information node that memory capacity is identical within the storage system
{Dk+1,Dk+2,…,Dk+lTime, each check-node being respectively directed in n-k check-node, generate
{ the 0,1} random vector h of a length of 1 × lj=(bj1,bj2,…,bjl), 1≤j≤n-k, b ∈ { 0,1}, hjFor school
Test node PjCorresponding random vector, wherein, each element b in vectorjiWith probability T value for 1;
The each check-node being respectively directed in n-k check-node, execution following steps:
Determine check-node PjCorresponding vector hjIntermediate value is each element b of 1ji;
Determine with value be respectively 1 each element bjiAt vector hjCorresponding each in middle position
Information node Dk+i;
By check-node PjEach information node D with described correspondencek+iCarry out XOR successively,
Result is as the P after updatingj;
By l information node { Dk+1,Dk+2,…,Dk+lAdd in described storage system, and respectively
By random vector hjCorrespondence adds random vector g tojAfterbody, the random vector g after being updatedj。
4. method as claimed in claim 3, it is characterised in that be respectively directed to n-k described
Each check-node in check-node, before performing following steps, also includes:
Determine l the information node { D that will addk+1,Dk+2,…,Dk+lThe most all do not store data,
If it is not, then each check-node being respectively directed to described in Zhi Hanging in n-k check-node, hold
The step for of row following steps;And
If it is, directly by l information node { Dk+1,Dk+2,…,Dk+lAdd described storage system to
In system, and respectively by random vector hjCorrespondence adds random vector g tojAfterbody, updated
After random vector gj。
5. the method for claim 1, it is characterised in that also include:
When l information node { C deleted within the storage system by needs1,C2,…,ClTime, determine and treat
L the information node { C deleted1,C2,…,ClNumbering in information node set;
The each check-node being respectively directed in n-k check-node, generates { the 0,1} of a length of 1 × l
Vector fj=(cj1,cj2,…,cjl), 1≤j≤n-k, fjFor check-node PjCorresponding vector, wherein,
cjiFor information node CiNumbering in information node set is at gjThe element of middle correspondence position;
The each check-node being respectively directed in n-k check-node, execution following steps:
Determine check-node PjCorresponding vector fjIntermediate value is each element c of 1ji;
Determine with value be respectively 1 each element cjiIn vector fjCorresponding each in middle position
Information node Ci;
By check-node PjEach information node C with described correspondenceiCarry out XOR successively,
Result is as the P after updatingj;
By l information node { C1,C2,…,ClFrom described storage system delete, and respectively from
Machine vector gjMiddle deletion each information node { C1,C2,…,ClThe element of numbering correspondence position, obtain more
Random vector g after Xinj。
6. method as claimed in claim 5, it is characterised in that be respectively directed to n-k verification
Each check-node in node, generates { 0, the 1} vector f of a length of 1 × lj=(cj1,cj2,…,cjl) it
Before, also include:
Determine l the information node { C that will delete1,C2,…,ClThe most all do not store data, if
, then do not perform each check-node being respectively directed in n-k check-node, generate a length of
{ the 0,1} vector f of 1 × lj=(cj1,cj2,…,cjl) the step for;And
If it is, directly by l information node { C1,C2,…,ClDelete from described storage system,
And respectively from random vector gjMiddle deletion each information node { C1,C2,…,ClNumbering correspondence position
Element, the random vector g after being updatedj。
7. the method for claim 1, it is characterised in that also include:
When needing to add l the check-node that memory capacity is identical within the storage system
{Pn-k+1,Pn-k+2,…,Pn-k+lTime, each check-node being respectively directed in l check-node, generate
{ the 0,1} random vector g of a length of 1 × kn-k+j=(an-k+j,1,an-k+j,2,…,an-k+j,k), 1≤j≤l,
A ∈ { 0,1}, gn-k+jFor check-node Pn-k+jCorresponding random vector, wherein, each element a in vectorn-+j,i
With probability T value for 1;
The each check-node being respectively directed in l check-node, execution following steps:
Determine check-node Pn-k+jCorresponding vector gn-k+jIntermediate value is each element a of 1n-k+j,i;
Determine with value be respectively 1 each element an-k+j,iAt vector gn-k+jMiddle position pair
The each information node D answeredi;
By check-node Pn-k+jEach information node D with described correspondenceiCarry out XOR successively, knot
Fruit is as the P after updatingn-k+j。
8. the method for claim 1, it is characterised in that also include:
Determine check-node to be deleted, from described storage system, delete described verification to be deleted
Node.
9. the method for claim 1, it is characterised in that described storage system also has appearance
Calamity parameter { t, δ }, wherein, t+ δ is the check-node number of described storage system, described storage system
Allowing that at most t node is complete to make mistakes, δ is additionally to save needed for described storage system ensures high redundancy ability
Count, and t+ δ < k, t+ δ+k=n, described method, also include:
When the storage loss of data of l information node of described storage system, from described storage system
Remaining n-l node is arbitrarily chosen δ node of k+, and including the individual check-node of δ 'The individual information node of k+ δ-δ ', l≤t;
By each for individual for δ ' check-node self-corresponding 1 × k 0,1} random vector take out, be arranged as δ ' × k's
{ 0,1} matrix Rδ'×k, wherein matrix jth V&V of behavior node1 corresponding × k 0,1} random vector,
1≤j≤δ';
Structural matrix Hδ'×(k+δ'), matrix Hδ'×(k+δ')Left part is divided into matrix Rδ'×k, right part is divided into list
Position battle array Iδ'×δ';
Based on the known k+ individual information node of δ-δ ' and the individual check-node of δ ', structure vector β(k+δ')×1,
Wherein β(k+δ')×1K position is information node { D from left to right1,D2,…,Dk, wherein unknown information node becomes
Amount X replaces, and right-hand member is
For relational expression Hδ'×(k+δ')·β(k+δ')×1=0 solves, and obtains l the information node lost
Storage data;
By in the data recovery storage of l information node that obtains to described storage system.
10. a storage system constructing device, it is characterised in that described storage system includes information
Node and check-node, described information node is used for storing data, and described check-node is used for recovering
In data, and described storage system, the memory capacity of each node is identical, and described device includes:
Parameter determination unit, is used for determining that { n, k, T}, wherein, n is described to storage nodes parameter
The node total number of storage system, k is the information node number of described storage system, and T is that individual element exists
{ in 0,1}, value is the probability of 1;Wherein n, k are positive integer, 0 < T < 1;
Numbered cell, for numbering in order k information node as { D1,D2,…,Dk, and will
N-k check-node numbers in order as { P1,P2,…,Pn-k, wherein, at the beginning of n-k check-node
Beginning data are zero;
Vector signal generating unit, for each check-node being respectively directed in n-k check-node,
Generate { the 0,1} random vector g of a length of 1 × kj=(aj1,aj2,…,ajk), 1≤j≤n-k, a ∈ 0,1},
gjFor check-node PjCorresponding random vector, wherein, each element a in vectorjiWith probability T value
It is 1;
Encoding relation sets up unit, for each verification joint being respectively directed in n-k check-node
Point, execution following steps:
Determine check-node PjCorresponding vector gjIntermediate value is each element a of 1ji;
Determine with value be respectively 1 each element ajiAt vector gjCorresponding each in middle position
Information node Di;
By check-node PjEach information node D with described correspondenceiCarry out XOR, result successively
As the P after updatingj。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610211840.1A CN105933386B (en) | 2016-04-06 | 2016-04-06 | A kind of storage system construction method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610211840.1A CN105933386B (en) | 2016-04-06 | 2016-04-06 | A kind of storage system construction method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105933386A true CN105933386A (en) | 2016-09-07 |
CN105933386B CN105933386B (en) | 2018-11-27 |
Family
ID=56840200
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610211840.1A Active CN105933386B (en) | 2016-04-06 | 2016-04-06 | A kind of storage system construction method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105933386B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109460185A (en) * | 2018-10-29 | 2019-03-12 | 锐捷网络股份有限公司 | Set up method, apparatus, storage pool and the storage medium of piecemeal group CKG |
CN110278222A (en) * | 2018-03-15 | 2019-09-24 | 华为技术有限公司 | The method, system and relevant device of data management in distributed file storage system |
CN111435323A (en) * | 2019-01-15 | 2020-07-21 | 阿里巴巴集团控股有限公司 | Information transmission method, device, terminal, server and storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102420616B (en) * | 2011-11-16 | 2013-08-14 | 西安电子科技大学 | Error correction method by using quasi-cyclic LDPC code based on Latin square |
CN102624866B (en) * | 2012-01-13 | 2014-08-20 | 北京大学深圳研究生院 | Data storage method, data storage device and distributed network storage system |
-
2016
- 2016-04-06 CN CN201610211840.1A patent/CN105933386B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102420616B (en) * | 2011-11-16 | 2013-08-14 | 西安电子科技大学 | Error correction method by using quasi-cyclic LDPC code based on Latin square |
CN102624866B (en) * | 2012-01-13 | 2014-08-20 | 北京大学深圳研究生院 | Data storage method, data storage device and distributed network storage system |
Non-Patent Citations (2)
Title |
---|
JAMES S. PLANK; CATHERINE D. SCHUMAN; B. DEVIN ROBISON: ""Heuristics for optimizing matrix-based erasure codes for fault-tolerant storage systems"", 《IEEE/IFIP INTERNATIONAL CONFERENCE ON DEPENDABLE SYSTEMS AND NETWORKS (DSN 2012)》 * |
JIANQIANG LUO; KEVIN D. BOWERS; ALINA OPREA; LIHAO XU;: ""Efficient software implementations of large finite fields GF (2 n </sup>) for secure storage applications"", 《ACM TRANSACTIONS ON STORAGE (TOS)》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110278222A (en) * | 2018-03-15 | 2019-09-24 | 华为技术有限公司 | The method, system and relevant device of data management in distributed file storage system |
CN110278222B (en) * | 2018-03-15 | 2021-09-14 | 华为技术有限公司 | Method, system and related device for data management in distributed file storage system |
CN109460185A (en) * | 2018-10-29 | 2019-03-12 | 锐捷网络股份有限公司 | Set up method, apparatus, storage pool and the storage medium of piecemeal group CKG |
CN109460185B (en) * | 2018-10-29 | 2022-03-18 | 锐捷网络股份有限公司 | Method and device for building block group CKG, virtual storage pool and storage medium |
CN111435323A (en) * | 2019-01-15 | 2020-07-21 | 阿里巴巴集团控股有限公司 | Information transmission method, device, terminal, server and storage medium |
CN111435323B (en) * | 2019-01-15 | 2023-06-20 | 阿里巴巴集团控股有限公司 | Information transmission method, device, terminal, server and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN105933386B (en) | 2018-11-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104052576B (en) | Data recovery method based on error correcting codes in cloud storage | |
CN103336785B (en) | A kind of distributed storage method based on network code and device thereof | |
CN105024704B (en) | A kind of row layering ldpc decoder implementation method of low complex degree | |
CN106484559B (en) | A kind of building method of check matrix and the building method of horizontal array correcting and eleting codes | |
CN103746774B (en) | The fault-tolerant coding method that a kind of efficient data is read | |
CN106100801A (en) | A kind of non-homogeneous erasure code method of cloud storage system | |
CN107844272A (en) | A kind of cross-packet coding and decoding method for improving error correcting capability | |
CN105721611A (en) | General method for generating minimal storage regenerating code with maximum distance separable storage code | |
CN105846830A (en) | Data processing device | |
CN106790408B (en) | A kind of coding method for distributed memory system node reparation | |
CN106776129A (en) | A kind of restorative procedure of the multinode data file based on minimum memory regeneration code | |
CN103838649B (en) | Method for reducing calculation amount in binary coding storage system | |
CN105933386A (en) | Construction method and device of storage system | |
CN111078460A (en) | Fast erasure code calculation method | |
CN105808170B (en) | A kind of RAID6 coding methods that can repair single disk error | |
CN106788891A (en) | A kind of optimal partial suitable for distributed storage repairs code constructing method | |
CN108132854A (en) | A kind of correcting and eleting codes coding/decoding method that can restore data element and redundant elements simultaneously | |
CN113901069B (en) | Data storage method and device of distributed database | |
CN102064835B (en) | Decoder suitable for quasi-cyclic LDPC decoding | |
CN107153661A (en) | A kind of storage, read method and its device of the data based on HDFS systems | |
Li et al. | A numerical study of bravyi-bacon-shor and subsystem hypergraph product codes | |
CN106788455B (en) | A kind of building method of the optimal partial repairable system code based on packet | |
CN108449090A (en) | A kind of configurable multi-code is long, multi code Rate of Chinese character ldpc decoder | |
CN103929188A (en) | Coding method, device and system | |
WO2018209541A1 (en) | Coding structure based on t-design fractional repetition codes, and coding method |
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 |