KR101873879B1 - Data distribution storage apparatus and method using relative difference set generated from the group having the two-dimensional element - Google Patents

Abstract

Disclosed is a data distribution storage apparatus and method using a relative difference set generated from a group having two-dimensional elements. A method for distributing a plurality of messages to a plurality of data stores, the method comprising: distributing the plurality of messages using only a minimum of data stores, selecting only a minimum data store among the data stores, A plurality of messages can be loaded so as to prevent data loss for the plurality of messages and to provide a technique for enabling fast data loading from data stores.

Description

Technical Field [0001] The present invention relates to a data distribution storage apparatus and a data distribution method using a relative difference set generated from a group having a two-dimensional element,

The present invention relates to a technique for distributing and storing a plurality of messages in a plurality of data stores so as to facilitate recovery of data when data loss occurs.

Recently, as the spread of the Internet becomes active, a variety of Internet service providers are emerging. These Internet service providers provide e-mail services, or provide various services such as blogs, cafes, and cloud data storage services.

Especially, as the popularity of social network services has increased recently, a lot of Internet service providers providing social network services are also increasing in number.

These Internet service providers need to store a large amount of data uploaded from a large number of client terminals in their own data storage to provide an e-mail service, a blog, a cafe, a cloud service, a social network service, If it is lost in the store, it is trying to prevent loss of data because it does not provide normal service to customers.

One such effort is introducing a way to back up user messages that need to be stored in a data store to multiple data stores. Such a data backup method is one of the most ideal methods for coping with data loss in that the original message can be replaced by using a backup message even if the original message is lost.

In this way, when one original message is stored in a plurality of data stores at a plurality of times, when a data load command for the original message is applied, Since the message corresponding to the original message is immediately loaded from the data storage, it is possible to provide a faster speed of data loading than a method of storing the original message in one data storage.

However, in the existing data backup method, only a method of distributing a copy message for an original message to a plurality of data stores is adopted, and therefore, from the viewpoint of Internet service providers who need to store a large number of messages in a data store, A problem may arise in that a large amount of data storage capacity is required because copy messages for each of a plurality of messages are generated and distributedly stored in the respective data stores.

Accordingly, in a method of distributing a plurality of messages to a plurality of data stores, a method of distributing a plurality of messages using only a minimum number of data stores and selecting and accessing only a minimum data store among the data stores There is a need for research on a technique for preventing data loss for the plurality of messages and enabling quick data loading from the data stores by loading all of the plurality of messages.

A method for distributing a plurality of messages to a plurality of data stores, the method comprising: distributing the plurality of messages using only a minimum of data stores, selecting only a minimum data store among the data stores, So that the plurality of messages can be loaded so as to prevent data loss for the plurality of messages and to enable fast data loading from the data stores.

According to an embodiment of the present invention, a data distribution storage apparatus using a relative difference set generated from a group having two-dimensional elements is a natural number satisfying the condition of N = q ² and q is a prime number of odd number - a distributed group generation unit for generating N message distribution groups corresponding to the N data stores when a distributed storage command to N data stores is applied to N messages, Wherein the messages distributed to the N message distribution groups may overlap each other, and a message distribution unit for allocating q messages allocated to the N message distribution groups, Separates the N messages into N data distribution groups and a data store corresponding to each message distribution group among the N data stores And a data storage unit for storing the data.

Also, in the data distribution storage method using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention, N - N is a natural number satisfying the condition of N = q ² , and q is an odd number Generating N message distribution groups corresponding to the N data stores when a distributed store command to N data stores is granted for a few number of messages, And distributing the N messages to the N distributed message groups in a distributed manner, the messages distributed to the N message distribution groups may overlap each other, and transmitting q messages allocated to the N message distribution groups Separating the N message distributing groups into a data store corresponding to each message distribution group among the N data stores, It should.

A method for distributing a plurality of messages to a plurality of data stores, the method comprising: distributing the plurality of messages using only a minimum of data stores, selecting only a minimum data store among the data stores, A plurality of messages can be loaded so as to prevent data loss for the plurality of messages and to provide a technique for enabling fast data loading from data stores.

1 is a diagram illustrating a structure of a data distribution storage apparatus using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention.
FIG. 2 is a diagram for explaining operations of a data distribution storage apparatus using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention.
3 is a flowchart illustrating a data distribution storage method using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings. It is to be understood that the description is not intended to limit the invention to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Like reference numerals in the drawings are used for similar elements and, unless otherwise defined, all terms used in the specification, including technical and scientific terms, are to be construed in a manner that is familiar to those skilled in the art. It has the same meaning as commonly understood by those who have it.

1 is a diagram illustrating a structure of a data distribution storage apparatus using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention.

Referring to FIG. 1, a data distribution storage apparatus 110 using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention includes a distributed group generation unit 111, a message distribution assignment unit 112 and a data storage unit 113.

The distributed group generation unit 111 generates N (N) data corresponding to N data stores 101, 102, and 103 when distributed storage commands to N data stores 101, 102, Message distribution groups.

Here, N is a natural number satisfying the condition of N = q ² , and q is a prime number of an odd number.

A message is a kind of data symbol to be stored in N data stores 101, 102 and 103, and the N messages are different data symbols.

The message distribution assigning unit 112 distributes the N messages to each of the N message distribution groups q-by-one.

 At this time, messages distributedly allocated to the N message distribution groups may overlap each other.

The data storage unit 113 stores q messages allocated to each of the N message distribution groups to each message distribution group among N data stores 101, And stores them separately in the corresponding data repository.

According to an embodiment of the present invention, the message distribution assignment unit 112 includes a first sequence generation unit 114, a second sequence generation unit 115, a sequence matching unit 116, and an assignment unit 117 .

으로 정의됨)에 상대적인 G(G는

로 정의됨)의 상대 차집합 D(상기 상대 차집합 D는

으로 정의되고, a=1이며, (q, q, q, 1)의 파라미터를 가짐)를 연산하고, 상기 상대 차집합 D에 포함되는 q개의 2차원 원소들에 기초하여 상기 G에 포함되어 있는 (i₁, i₂)(i₁, i₂는 모두 Z_q에 포함되는 원소임)로 표현되는 2차원 원소들에 대해 "1"과 "0"의 코드 값을 하나씩 할당함으로써, N개의 "1"과 "0"의 코드 값으로 구성된 제1 시퀀스를 생성한다.The first sequence generator 114 generates K (K) for the group Z _q for addition with the number of elements _q ,

Relative to G (G)

(The relative difference set D is defined as < RTI ID = 0.0 >

(Q, q, q, 1), a = 1 and has parameters of (q, q, q, 1) quot; 1 "and" 0 "are assigned to the two-dimensional elements represented by (i ₁ , i ₂ ) (i ₁ and i ₂ are elements included in Z _q ) Quot; 1 "and" 0 "

At this time, the first sequence generator 114 multiplies q two-dimensional elements included in the relative difference set D among the two-dimensional elements represented by (i ₁ , i ₂ ) Quot; and " 0 "are assigned to the remaining two-dimensional elements not included in the relative difference set D by allocating a code value of "0" 1 sequence can be generated.

The second sequence generator 115 calculates a right coset of the relative difference set D for all the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G, and outputs N Dimensional elements represented by (i ₁ , i ₂ ) included in the G based on q two-dimensional elements included in each of the N-1 uxcosets after generating -1 right co- 1 " and " 0 "are generated for each of the N-1 uxcodes by allocating one code value of" 1 "≪ / RTI > second sequences.

At this time, the second sequence generation unit 115 is the N-1 of the right koset s for each, corresponding to the at least one G (i _1, i ₂₎ of said N-1 of the two-dimensional element, represented by Wu Quot; 1 "is assigned to q two-dimensional elements included in each of the co-sets, and a code value of" 0 "is assigned to remaining two-dimensional elements not included in each of the N- 1 "and "0" code values by assigning N < th >

Let G be a group with order uv and K be a normal subgroup with degree u. Let G be the relative difference set of G relative to K, D can be defined as shown in Table 1 below.

D is a subset of ω elements of G, taking any two elements d ₁ , d ₂ of D and computing the inverse of d ₁ and d _{2 so} that all nonzero elements of G except K are exactly λ It has the property to occur once. At this time, the parameter of D is represented by (v, u, ω, λ).

,

라 정의하면, 이때,

에 대해 K에 상대적인 G의 상대 차집합 D는 하기의 수학식 1과 같이 정의될 수 있고, D의 파라미터는 (q, q, q, 1)이 된다.At this time, when q is an odd number, Z _q is a group for addition,

,

In this case,

The relative difference set D of G relative to K can be defined as Equation (1) below, and the parameter of D becomes (q, q, q, 1).

Here, the relative difference set D includes _q elements among the two-dimensional elements represented by (i ₁ , i ₂ ) (i ₁ , i ₂ are elements included in Z _q ) included in G .

가 있을 때,

에 포함되어 있는 2차원 원소인 (i₁, i₂)에 대한 상대 차집합 D의 우코셋은 상대 차집합 D에 포함되어 있는 q개의 2차원 원소들인

각각에 대해 아래의 수학식 2와 같이

를 연산함으로써 생성된 q개의 2차원 원소들로 구성될 수 있다.In the case of the right co-set, a relative difference set having parameters (q, q, q, 1)

Lt; / RTI >

The right co-set of the relative difference set D with respect to the two-dimensional elements (i ₁ , i ₂ ) included in the relative difference set D is q two-dimensional elements included in the relative difference set D

As shown in the following Equation 2,

Dimensional elements that are generated by computing q.

각각(

은 제외)에 대해서 상기 수학식 2와 같은 연산을 수행함으로써, 총 q²-1개가 생성될 수 있다.In this way, the right co-set of the relative difference set D is included in G

each(

, The total of q ² -1 can be generated by performing an operation such as the above-described equation (2).

이라는 N-1개의 우코셋들이 생성될 수 있다.In this situation, if N is a natural number satisfying the condition of N = q ² , and a = 1, the relative difference set D is calculated according to Equation (1) , The relative difference set D and

RTI ID = 0.0 > N-1 < / RTI >

이라는 N-1개의 우코셋들이 생성되었을 때, 상대 차집합 D와

이라는 N-1개의 우코셋들 각각에 대해서, 상기 G에 포함되어 있는 (i₁, i₂)로 표현되는 2차원 원소들 중 상대 차집합 D와 N-1개의 우코셋들 각각에 포함되는 q개의 2차원 원소들에 대해 "1"의 코드 값을 할당하고, 상대 차집합 D와 N-1개의 우코셋들 각각에 포함되지 않는 나머지 2차원 원소들에 대해 "0"의 코드 값을 할당하는 방식으로 상기 G에 포함되어 있는 (i₁, i₂)로 표현되는 2차원 원소들에 대해 "1"과 "0"의 코드 값을 할당하게 되면, 총 N개의 "1"과 "0"의 코드 값으로 구성된 시퀀스들을 획득할 수 있다.Thus, the relative difference set D and

When the N-1 right cosets are generated, the relative difference set D

(I ₁ , i ₂ ) included in the G, for each of the N-1 uxcosets, which are included in the relative difference D and the N-1 right co- Quot; 1 "is assigned to the two-dimensional elements, and a code value of" 0 "is assigned to the remaining two-dimensional elements not included in each of the relative difference set D and the N-1 right co- 1 "and" 0 "are assigned to the two-dimensional elements represented by (i ₁ , i ₂ ) It is possible to obtain sequences composed of code values.

개 이상이 되고, "0"의 코드 값의 개수는 항상

개 이하가 되는 특징을 가지고 있다.At this time, the number of code values of "1" existing in each sequence is always q, and q sequences randomly selected from among the N sequences are randomly selected, The number of code values of "1" existing in the result of performing the "OR"

, And the number of code values of "0"

Or less.

이고,

이므로, 상기 G에 포함되어 있는 (i₁, i₂)로 표현되는 2차원 원소들은 (0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2)가 될 수 있으며, a=1일 때, 상기 수학식 1에 따라 연산되는 상대 차집합

가 되며, 상대 차집합 D의 파라미터는 (3, 3, 3, 1)이 될 수 있다.For example, assuming that q = 3, the number of elements

ego,

Since the two-dimensional element represented by (i _1, i ₂₎ contained in the G are (0,0), (0,1), (0,2), (1,0), (1, 1 (1, 2), (2, 0), (2, 1), (2, 2)

, And the parameter of the relative difference set D may be (3, 3, 3, 1).

(1, 0), (1, 0), (1, 1), (1, 2), (2, 1 ", (1, 2), which are the three 2-dimensional elements included in the relative difference set D among (0, 0), (2, 1) Quot; 1 "and" 2 " are assigned to the two-dimensional elements included in G by assigning a code value of " 0 "to the remaining two- Quot; 1 "and " 0" code values of "100011000 "

가 생성되고, 여기서, 첫 번째 우코셋이

이기 때문에 G에 포함되어 있는 2차원 원소들인 (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2) 중

에 포함되는 3개의 2차원 원소인 (0, 1), (1, 0), (1, 2)에 대해 "1"의 코드 값을 할당하고,

에 포함되지 않는 나머지 2차원 원소들에 대해 "0"의 코드 값을 할당하는 방식으로 상기 G에 포함되어 있는 2차원 원소들에 대해 "1"과 "0"의 코드 값을 순차적으로 할당하게 되면, "010101000"이라는 9개의 "1"과 "0"의 코드 값으로 구성된 또 다른 시퀀스를 생성할 수 있다.(1, 0), (1, 1), (1, 2), (2, 0), (2, (Except for (0, 0)) of the relative difference set D according to Equation (2) for each of (1, 2)

Is generated, where the first right coset is < RTI ID = 0.0 >

(1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2)

A code value of "1" is assigned to (0, 1), (1, 0), (1, 2)

Quot; 1 "and" 0 "are sequentially allocated to the two-dimensional elements included in the G in such a manner that a code value of" 0 "is assigned to remaining two- Quot; 1 "and" 0 ", i.e., "010101000"

에 대해서 시퀀스를 생성하게 되면, 최종적으로 총 9개의 시퀀스들이 생성될 수 있고, 이때, 각 시퀀스들 내에 존재하는 "1"의 코드 값의 개수는 모두 3개로 일정하며, 상기 9개의 시퀀스들 중에서 서로 다른 3개의 시퀀스들을 랜덤하게 선택한 후 랜덤하게 선택된 3개의 시퀀스들을 모두 "OR"연산하게 되면, "OR" 연산이 수행된 결과 값에서 존재하는 "1"의 코드 값의 개수는 항상 6개 이상이 되며, "0"의 코드 값의 개수는 항상 3개 이하가 되는 특징을 가지게 된다.In this way,

A total of 9 sequences can be finally generated. At this time, the number of code values of "1" existing in each sequence is constant at three, and among the nine sequences, The number of code values of "1" existing in the result of performing the "OR" operation is always 6 or more when the other three sequences are randomly selected and all three randomly selected sequences are & And the number of code values of "0 " is always 3 or less.

In accordance with the above-described relative difference set and the sequence generation process from the right co-ordinate, the first sequence generator 114 calculates the relative difference set for the group Z _q for addition with the number of elements _q , ₁ relative to q two-dimensional elements included in the relative difference set D among the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G after calculating the relative difference set D of G, And a code value of "0" is assigned to remaining two-dimensional elements not included in the relative difference set D, as shown in (i ₁ , i ₂ ) By assigning code values of "1" and "0" to each of the two-dimensional elements, a first sequence composed of N "1" and "0" code values can be generated.

Then, the second sequence generator 115 calculates the right coset of the relative difference set D for all the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G, and outputs N-1 right (I ₁ , i ₂ ) contained in the G is included in each of the N-1 uxcosets of the two-dimensional elements represented by (i ₁ , i ₂ ) A code value of "1" is allocated to q two-dimensional elements which are not included in each of the N-1 ellipses, and a code value of "0" is assigned to remaining two-dimensional elements not included in each of the N-1 By assigning code values of "1" and "0" to each of the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G for each of the N-1 uxcosets, Lt; RTI ID = 0.0 > sequences. ≪ / RTI >

When the first sequence and the N-1 second sequences are generated, the sequence matching unit 116 may combine the first sequence and the N-1 second sequences for each of the N message distribution groups Match one by one.

For each of the N message distribution groups, the allocation unit 117 divides the N messages into q message groups based on N code values constituting each sequence matched to each of the N message distribution groups Distributed allocation.

In this case, according to an embodiment of the present invention, the message distribution assigning unit 112 may further include an index assigning unit 118 assigning an index value associated with the order information to each of the N messages.

At this time, the allocating unit 117 allocates, for each of the N message distribution groups, an arrangement order of N code values constituting each sequence matched to each of the N message distribution groups, And then allocate the N messages to each of the N messages based on the code value of the array order corresponding to the index value assigned to the N messages.

In this case, according to an embodiment of the present invention, the allocation unit 117 allocates, for each of the N message distribution groups, N code values constituting each sequence matched to the N message distribution groups After identifying the code value of "1 " existing for q, an index value corresponding to the code value of the" 1 "code value among the N messages is allocated to each of the N message distribution groups Q messages can be distributed.

According to an embodiment of the present invention, the data distribution storage apparatus 110 using the relative difference set generated from the group having two-dimensional elements further includes a storage message extracting unit 119 and a data loading unit 120 .

The stored message extracting unit 119 extracts N messages stored in the N data stores 101, 102, and 103 and stores the N messages in the N Randomly selecting q ones of the N data stores (101, 102, 103) based on the command, and selecting from the q data stores each of the q store messages stored in each of the q data stores .

The data loading unit 120 loads the stored messages extracted from each of the q data repositories into the data corresponding to the load command for the N messages.

In this case, according to an embodiment of the present invention, the N messages may be Maximum Distance Separable (MDS) codes capable of recovering lost messages based on remaining messages when loss of some messages occurs .

Here, the MDS code is a code that can recover some messages lost based on remaining messages when a loss of some messages among a plurality of messages occurs, thereby performing MDS coding on a predetermined original message, Code can be used to recover lost messages when some of the messages changed to MDS code are lost. At this time, if the original message is to be generated again from the MDS code, MDS decoding can be performed to restore the original message. Therefore, when storing important messages, the MDS code is changed and stored, so that a case of loss of a message can be prepared in advance.

At this time, the data load unit 120 compares the index value assigned to the extracted stored messages with the index value assigned to the N messages when the q stored messages are extracted from each of the q data stores, If it is determined that there is a remaining message that is not included in the extracted stored messages among the N messages, restores the remaining message based on the extracted stored messages, And load the remaining messages into the data according to the load command for the N messages.

Hereinafter, the operation of the data distribution storage apparatus 110 using the relative difference set generated from the group having two-dimensional elements according to the present invention will be described in detail with reference to FIG. 2, for example.

FIG. 2 is a view for explaining the operation of the data distribution storage apparatus 110 using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention.

2, N data stores 101, 102 and 103 are connected to the data storage 1 211, the data storage 2 221, 3 231, Data Storage 4 241, Data Storage 5 251, Data Storage 6 261, Data Storage 7 271, Data Storage 8 281 and Data Storage 9 291, N = 9, it is assumed that the number of messages to be stored in the data repositories 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281 and 291) is nine.

At this time, when the distributed storage command to the data storage 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281, 291) is applied to nine messages, 9 message distribution groups 213, 223, 233, 243, 253, 263, 273, 283, 293 corresponding to 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281, 291) Can be generated.

The index assigning unit 118 included in the message distribution assigning unit 112 may assign an index value associated with the order information to the nine messages. When the index assigning unit 118 assigns the nine messages M ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ " As shown in FIG.

에 대해서, G가

로 정의되고, K가

으로 정의될 때, 상기 수학식 1에 따라 K에 상대적인 G의 상대 차집합 D를 연산(a=1임)할 수 있다.q = 3, and N = 9, the first sequence generator 114 generates a first sequence having three elements,

, G

And K is defined as

, A relative difference D of G relative to K can be calculated (a = 1) according to Equation (1) above.

이므로, 상기 G에 포함되어 있는 (i₁, i₂)로 표현되는 2차원 원소들은 (0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2)가 될 수 있으며, a=1일 때, 상기 수학식 1에 따라 연산되는 상대 차집합

가 되며, 상대 차집합 D의 파라미터는 (3, 3, 3, 1)이 될 수 있다.Relatedly,

Since the two-dimensional element represented by (i _1, i ₂₎ contained in the G are (0,0), (0,1), (0,2), (1,0), (1, 1 (1, 2), (2, 0), (2, 1), (2, 2)

, And the parameter of the relative difference set D may be (3, 3, 3, 1).

Then, the first sequence generator 114 generates 2-dimensional elements (0,0), (0,1), (0,2), (1,0), (1,1) (0, 0), (1, 1), and (2), which are included in the relative difference set D among (1, 2), (2, 0), 2 "included in the G is allocated in such a manner that a code value of" 1 "is allocated to the first difference value (1, 2) and a code value of" 0 "is assigned to remaining two-dimensional elements not included in the relative difference difference D, By assigning code values of "1" and "0" sequentially to the dimensional elements, a first sequence composed of nine code values "1" and "0" of "100011000" .

라는 8개의 우코셋들을 생성할 수 있다.The second sequence generation unit 115 generates the second sequence by using (0,0), (0,1), (0,2), (1,0), (1,1), (0, 0) is calculated for each of (1, 2), (2, 0), (2, 1) Ex)

Can be generated.

Then, the second sequence generator 115 sequentially generates 1s and 2s for the 2-dimensional elements included in the G based on the 3 2-dimensional elements included in each of the 8-ellipses, Quot; 0 "code value to generate eight second sequences by generating a sequence composed of nine code values of" 1 "and" 0 " for each of the eight co-coils.

이기 때문에 G에 포함되어 있는 2차원 원소들인 (0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2) 중

에 포함되는 3개의 2차원 원소들인 (0, 1), (1, 0), (1, 2)에 대해 "1"의 코드 값을 할당하고,

에 포함되지 않는 나머지 2차원 원소들에 대해 "0"의 코드 값을 할당하는 방식으로 상기 G에 포함되어 있는 2차원 원소들에 대해 순차적으로 "1"과 "0"의 코드 값을 할당함으로써, "010101000"(222)이라는 9개의 "1"과 "0"의 코드 값으로 구성된 시퀀스를 생성할 수 있다.In this regard, the second sequence generator 115 generates a second sequence

(1, 0), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2)

A code value of "1" is assigned to (0, 1), (1, 0), (1, 2)

Quot; 1 "and" 0 "are sequentially assigned to the two-dimensional elements included in the G in such a manner that a code value of" 0 "is assigned to remaining two- Quot; 1 "and" 0 ", i.e., "010101000" (222).

에 대해서도 시퀀스를 생성함으로써, "001110000"(232), "000100011"(242), "000010101"(252), "000001110"(262), "011000100"(272), "101000010"(282), "110000001"(292)라는 시퀀스들을 추가로 생성할 수 있다.In this way, the second sequence generator 115 generates the second

000010101 "252," 000001110 "262," 011000100 "272," 101000010 "282," 01000010 "Quot; 110000001 "(292).

This first sequence "100011000" 212 and eight second sequences "010101000" 222, "001110000" 232, "000100011" 242, "000010101" 252, "000001110" 262, "011000100" 272, "101000010" 282, and "110000001" 292 always have q "1" code values as described above. In this embodiment, q is set to 3 The code value of "1" contained in each of the nine sequences 212, 222, 232, 242, 252, 262, 272, 282, 292 is always three.

개 이상이 되고, "0"의 코드 값은 항상

개 이하가 되기 때문에 본 실시예에서는 q가 3이라고 가정하였으므로, 9개의 시퀀스들(212, 222, 232, 242, 252, 262, 272, 282, 292) 중 어느 3개의 시퀀스들 간에 "OR" 논리 연산이 수행되면, 그 연산 결과 값에 포함되어 있는 "1"의 코드 값은 항상 6개 이상이 되고, "0"의 코드 값은 항상 3개 이하가 된다.If the first sequence and the second sequence perform an "OR" logical operation between three different sequences, the code value of "1" contained in the result of the logical operation is always

, And the code value of "0 "

OR "logic between any three of the nine sequences 212, 222, 232, 242, 252, 262, 272, 282, 292 since q is assumed to be 3, When the operation is performed, the code value of "1" included in the operation result value is always 6 or more, and the code value of "0" is always 3 or less.

When the nine sequences 212, 222, 232, 242, 252, 262, 272, 282 and 292 are generated, the sequence matching unit 116 acquires the data sets 1 to 9 (211, 221, 231, 241, 9 sequences 212 and 222 for each of the nine message distribution groups 213, 223, 233, 243, 253, 263, 273, 283 and 293 corresponding to the message distribution groups 251, 261, 271, 281, , 232, 242, 252, 262, 272, 282, 292 can be matched one by one.

2, the sequence matching unit 116 may match 100011000 (212) to the message distribution group 1 (213), 010101000 (222) to the message distribution group 2 (223) , Match 001110000 (232) to message distribution group 3 233, match 000100011 (242) to message distribution group 4 243, match message distribution group 5 253 , Match 000001110 (262) to message distribution group 6 (263), 011000100 (272) to message distribution group 7 (273) , 101000010 (282) for message distribution group 8 (283), and 110000001 (292) for message distribution group 9 (293).

The allocator 117 then allocates nine message distribution groups 213, 223, 233, and 233 for each of nine message distribution groups 213, 223, 233, 243, 253, 263, M ₁ , m ₂ , m ₃ , m ₄ , and m ₅ , based on the nine code values constituting each sequence matched to each of the nine messages , m _6, m _7, m _8, m ₉ "a may be assigned, assignment section 117 of the nine message distribution group (213, 223, 233, 243, 253, 263, 273, 283, 293) M ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , and m ₉ , which are the nine code values that constitute each sequence, after comparing the index value assigned to each of 9 messages, and an index value assigned to the _{"m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9" M ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ "can be assigned to nine message distribution groups 213, 223, 233, 243, 253, 263, 273, 283 and 293.

At this time, the assigning unit 117 assigns each of the nine code values constituting each sequence matched to each of the nine message distribution groups 213, 223, 233, 243, 253, 263, 273, 283, Code value of "1" among the nine messages "m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ " 233, 243, 253, 263, 273, 283, and 293 to which the index values corresponding to the arrangement order for the message distribution groups 213, 223, 233, 243, 253, 263, 273, 283, and 293 are assigned.

, Assignment section 117, the nine message distribution group (213, 223, 233, 243, 253, 263, 273, 283, 293) on to which are nine message _{"m 1, m 2, m} 3 , respectively in connection , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ "are divided according to each message distribution group.

First, after for message distribution group 1 (213) from the allocation unit 117 100 011 000 212 Since the sequence of 100 011 000 212 matched checking the code of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 1 (213).

2, since the code value of "1" is arranged at positions 1, 5, and 6 in the 100011000 (212), the allocation unit 117 allocates the message distribution group 1 (213) M ₁ , m ₅ , and m ₆ "to which index values corresponding to positions 1, 5, and 6 are assigned to the indexes" ₁ "," ₅ ", and"

Then, after for message distribution group 2 (223) from the allocation unit 117 010 101 000 222 Since the sequence of 010 101 000 222 matched verify the code value of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 2 (223).

2, since the code value of "1" is arranged at positions 2, 4 and 6 in the 010101000222, the allocation unit 117 allocates the message distribution group 2 223, Quot; m ₂ , m ₄ , m ₆ "to which the index values corresponding to the positions ₂ , ₄ , and ₆ are assigned.

Then, after for message distribution group 3 (233) from the allocation unit 117 001 110 000 232 Since the sequence of 001 110 000 232 matched verify the code value of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 3 (233).

2, since the code value of "1" is arranged at positions 3, 4, and 5, the allocation unit 117 allocates the message distribution group 3 233, Quot; m ₃ , m ₄ , m ₅ "to which index values corresponding to positions 3, 4, and 5 are assigned to the indexes.

Then, after for message distribution group 4 (243) from the allocation unit 117 000 100 011 242 Since the sequence of 000 100 011 242 matched verify the code value of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 4 (243).

2, since the code value of "1" is arranged at positions 4, 8, and 9 in the 000100011 (242), the allocation unit 117 allocates the message distribution group 4 (243) Quot; m ₄ , m ₈ , m ₉ "to which the index values corresponding to positions ₄ , ₈ , and ₉ are assigned.

Then, after for messages distributed Group 5 253 from the allocation unit 117 000 010 101 252 Since the sequence of 000 010 101 252 matched verify the code value of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 5 (253).

2, since the code value of "1" is arranged at positions 5, 7, and 9 in the 000010101 252, the allocation unit 117 allocates the message distribution group 5 253, Quot; m ₅ , m ₇ , m ₉ "to which index values corresponding to positions ₅ , ₇ , and ₉ are assigned.

Then, after for message distribution group 6 (263) from the allocation unit 117 000 001 110 262 Since the sequence of 000 001 110 262 matched checking the code of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 6 (263).

2, since the code value of "1" is arranged at positions 6, 7 and 8 in the 000001110 262, the allocation unit 117 allocates the message distribution group 6 263, Quot; m ₆ , m ₇ , m ₈ "to which the index values corresponding to the positions ₆ , ₇ , and ₈ are assigned.

Then, after for messages distributed Group 7 (273) from the allocation unit 117 011 000 100 272 Since the sequence of 011 000 100 272 matched verify the code value of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 7 (273).

2, since the code value of "1" is arranged at positions 2, 3, and 7 in 011000100272, the allocation unit 117 allocates the message distribution group 7 273, M ₂ , m ₃ , m ₇ "to which the index values corresponding to the positions ₂ , ₃ , and ₇ are assigned can be assigned.

Then, after for message distribution group 8 (283) from the allocation unit 117 101 000 010 282, because the sequence of 101 000 010 282 matched verify the code value of "1", which are the nine message "m _{1, m 2, m 3, m 4} , m 5, m 6, m 7, m 8, m 9 "from" 1 "distributed three messages message with the index value to correspond with the arrangement order of the code value is assigned to the Group 8 (283).

2, since the code value of "1" is arranged at positions 1, 3 and 8 in the 101000010282, the allocation unit 117 allocates the message distribution group 8 (283) Quot; m ₁ , m ₃ , m ₈ "to which index values corresponding to positions ₁ , ₃ ,

Finally, with respect to the message distribution group 9 (293) a sequence of 110 000 001 292 matched because assignment unit 117, which are the nine message after checking the code of "1" from 110 000 001 292 "m _{1 , m 2, m 3, m} 4, m 5, m 6, m 7, m 8, m 9 match the arrangement order of the code values of 1 "" from the "three messages messages that are assigned index values for Can be allocated to the distribution group 9 (293).

2, since the code value of "1" is arranged at positions 1, 2, and 9 in the 110000001 (292), the allocation unit 117 allocates the message distribution group 9 (293) M ₁ , m ₂ , m ₉ "to which the index values corresponding to the positions ₁ , ₂ , and ₉ are assigned.

In this way, the nine messages "m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ " are all nine message distribution groups 213, 223, 233, The data storage unit 113 stores nine message distribution groups 213, 223, 233, 243, 253, 263, 273, and 283 293 are allocated to the data stores 1 to 9 (211, 221, 231, and 231) by nine message distribution groups 213, 223, 233, 243, 253, 263, 241, 251, 261, 271, 281, and 291) in the data store corresponding to each message distribution group.

The data storage unit 113 can store the three messages "m ₁ , m ₅ , m ₆ " assigned to the message distribution group 1 213 in the corresponding data store 1 (211) have.

The data storage unit 113 may store the three messages "m ₂ , m ₄ , m ₆ " assigned to the message distribution group 2 223 in the data store 2 221 as the corresponding data store .

The data storage unit 113 may store the three messages "m ₃ , m ₄ , m ₅ " assigned to the message distribution group 3 233 in the corresponding data store 3 231 .

The data storage unit 113 may store in which are three messages that are assigned to the message distribution group _{4 (243) "m 4,} m 8, m 9" data store of the data store 4 241 corresponding to the .

The data storage unit 113 may store the three messages "m ₅ , m ₇ , m ₉ " assigned to the message distribution group 5 253 in the corresponding data store 5 (251) .

The data storage unit 113 may store the three messages m ₆ , m ₇ , m ₈ assigned to the message distribution group 6 263 in the corresponding data store 6 (261) .

The data storage unit 113 may store the three messages "m ₂ , m ₃ , m ₇ " assigned to the message distribution group 7 273 in the corresponding data store 7 (271) .

The data storage unit 113 may store the three messages "m ₁ , m ₃ , m ₈ " assigned to the message distribution group 8 (283) in the corresponding data repository 8 (281) .

Finally, the data store 113 may store the three messages "m ₁ , m ₂ , m ₉ " assigned to the message distribution group 9 (293) in the corresponding data store 9 (291) have.

The nine messages m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ and m ₉ are stored in the data stores 1 to 9 211, 221, 231, 241, 251, 261, 271, 281, 291) when the distributed storage is complete, which are later the nine message _{"m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9 to perform a data load on "the nine messages, which are as the data load instruction is for _{the" m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9 " , Messages are stored in only the three data stores after accessing only three data stores out of the data stores 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281 and 291) When the nine messages, which are _{"m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9" at least six or more messages are presented in both extracts.

This is because the data stores 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281 and 291) The nine messages "m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ " are distributed and stored according to the code value of "1" The number of the three sequences 212, 222, 232, 242, 252, 262, 272, 282, 292 of the nine sequences 212, 222, 232, 242, Since the code value of "1" always comes out to 6 or more, if the "OR" logical operation is performed between the sequences, the data stores 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281 and 291 M ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₈ , m ₉ "at least six or more messages are those that are secured to both.

This results, which includes the nine message _{"m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9" by the data store 1 to 9 (211, 221, 231, 241 M ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , and m ₉ after the distributed storage is completed The storage message extracting unit 119 extracts the storage message from the data storage 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281, 291) The data loader 120 can extract three pieces of stored data stored in the three data stores, and the data loader 120 can load the nine messages M ₆ , m ₇ , m ₈ , m ₉ ", which are" m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ ".

At this time, if the nine messages "m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ " (9, 6) MDS code configured to be able to restore all of ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ " If three stored messages are extracted from three randomly selected data stores among the data stores 1 to 9 (211, 221, 231, 241, 251, 261, 271, 281 and 291) , which are the index value and the nine messages in _{"m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9" the nine messages by comparing the index value assigned to the which are when it is identified that the remaining messages that are not included in the the extracted stored messages of the _{"m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9" presence Based on the extracted storage messages, W, which are messages recovered by the extracted storing messages and nine messages wherein the said repair residual message _{"m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m 8, m 9 "To < / RTI >

In this regard, if the three randomly selected data stores are referred to as data storage 3 231, data storage 5 251 and data storage 7 271, the storage message extracting unit 119 extracts, from the data storage 3 231 _{"m 3, m 4, m} 5" a and to extract, from the data storage _{5 (251) "m 5,} m 7, m 9" a and to extract, from the data store 7 (271) "m _2, m ₃ , m ₇ "can be extracted.

Therefore, the stored message extraction section 119 from the end, the data storage 3 231, data store 5, 251 and the data store _{7 (271) "m 2,} m 3, m 4, m 5, m 7, m ₉ "can be extracted.

At this time, the data load unit 120 stores the index values assigned to the _nine messages "m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₆ , m ₇ , m ₈ , m ₉ " extracted _{"m 2, m 3, m} 4, m 5, m 7, m 9" above, which are the nine message "m _1, m _2, and the index value assigned to the comparison with each other m _3, m _4, m M ₂ , m ₃ , m ₄ , m ₅ , m ₇ , m ₉ "among the extracted messages" m ₅ , m ₆ , m ₇ , m ₈ , m ₉ ".

In this embodiment, since "m ₁ , m ₆ , m ₈ " is the remaining message, the data loading unit 120 checks (m ₁ , m ₆ , m ₈ ) M ₁ , m ₆ , m ₈ "can be restored from" m ₂ , m ₃ , m ₄ , m ₅ , m ₇ , m ₉ "

Consequently, the data loading unit 120 stores the restored "m ₁ , m ₆ , m _{8 &} quot ;, the data stored in the data storage 3 231, the data storage 5 251 and the data storage 7 271, _{m 2, m 3, m 4} , m 5, m 7, m 9 " a, which are, of the nine messages by _{loading" m 1, m 2, m} 3, m 4, m 5, m 6, m 7, m ₈ , m ₉ "can be loaded.

As a result, the data distribution storage device 110 using the relative difference set generated from the group having two-dimensional elements according to the present invention stores the nine messages "m ₁ , m ₂ , m ₃ , m ₄ , m ₅ , m ₂ , m ₃ , m ₆ , m ₇ , m ₈ , m ₉ "in accordance with the first sequence and the second sequences generated based on the relative difference set generated from the group having the two- 221, 231, 241, 251, 261, 271, 281, and 291. Thus, when data is loaded later, only three data stores can be accessed and data can be restored. The efficiency of utilization of system resources in the system can be improved.

3 is a flowchart illustrating a data distribution storage method using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention.

In step S310, when a distributed storage command to N data stores is applied to N (N is a natural number satisfying the condition of N = q ² and q is a prime number of odd number) messages, And creates N message distribution groups corresponding to the data stores.

In step S320, q messages are distributedly allocated to the N message distribution groups (messages distributed to the N message distribution groups may overlap each other).

In step S330, q messages allocated to each of the N message distribution groups are divided into N data distribution groups and stored in a data store corresponding to each message distribution group among the N data storage groups do.

으로 정의됨)에 상대적인 G(G는

로 정의됨)의 상대 차집합 D(상기 상대 차집합 D는

으로 정의되고, a=1이며, (q, q, q, 1)의 파라미터를 가짐)를 연산하고, 상기 상대 차집합 D에 포함되는 q개의 2차원 원소들에 기초하여 상기 G에 포함되어 있는 (i₁, i₂)(i₁, i₂는 모두 Z_q에 포함되는 원소임)로 표현되는 2차원 원소들에 대해 "1"과 "0"의 코드 값을 하나씩 할당함으로써, N개의 "1"과 "0"의 코드 값으로 구성된 제1 시퀀스를 생성하는 단계, 상기 G에 포함되어 있는 (i₁, i₂)로 표현되는 모든 2차원 원소들에 대해서 상기 상대 차집합 D의 우코셋을 연산하여 N-1개의 우코셋들을 생성한 후, 상기 N-1개의 우코셋들 각각에 포함되어 있는 q개의 2차원 원소들을 기초로 상기 G에 포함되어 있는 (i₁, i₂)로 표현되는 2차원 원소들에 대해 "1"과 "0"의 코드 값을 하나씩 할당하여 N개의 "1"과 "0"의 코드 값으로 구성된 시퀀스를 상기 N-1개의 우코셋들 각각에 대해 생성함으로써, N-1개의 제2 시퀀스들을 생성하는 단계, 상기 N개의 메시지 분산 그룹들 각각에 대해 상기 제1 시퀀스와 상기 N-1개의 제2 시퀀스들을 하나씩 매칭시키는 단계 및 상기 N개의 메시지 분산 그룹들 각각에 대해, 상기 N개의 메시지 분산 그룹들 각각에 매칭되어 있는 각 시퀀스를 구성하는 N개의 코드 값에 기초하여 상기 N개의 메시지들을 q개씩 분산 할당하는 단계를 포함할 수 있다.At this time, according to an embodiment of the present invention, in step S320, for a group Z _q for the addition of the number of elements _q , K

Relative to G (G)

(The relative difference set D is defined as < RTI ID = 0.0 >

(Q, q, q, 1), a = 1 and has parameters of (q, q, q, 1) quot; 1 "and" 0 "are assigned to the two-dimensional elements represented by (i ₁ , i ₂ ) (i ₁ and i ₂ are elements included in Z _q ) (I ₁ , i ₂ ) included in the G, and generating a first sequence composed of code values of "1" and "0" (I ₁ , i ₂ ) included in the G based on q 2-dimensional elements included in each of the N-1 uxcosets, Quot; 1 " and " 0 "are assigned to each of the N-1 code elements and the N-1 code elements, Generating N-1 second sequences by matching the first sequence and the N-1 second sequences one by one for each of the N message distribution groups, For each of the groups, distributing the N messages in a q-basis based on N code values constituting each sequence matched to each of the N message distribution groups.

According to an embodiment of the present invention, the step of generating the first sequence may include: q (q i, q i) contained in the relative difference set D among the two-dimensional elements represented by (i ₁ , i ₂ ) 1 "and" 0 "are assigned to the remaining two-dimensional elements not included in the relative difference set D, by allocating a code value of"Quot; 0 "code value, and generating the N-1 second sequences may include generating, for each of the N-1 uxcodes, _1, i ₂₎ two-dimensional element from among the N-1 of the right koset each q of the two-dimensional assigned a code value of "1" for the elements, and wherein the N-1 of the right koset included in, represented by Quot; 0 "code by assigning a code value of" 0 "to the remaining two-dimensional elements not included in each of the N & 1 < / RTI >

According to an embodiment of the present invention, the step S320 may further comprise the step of assigning an index value associated with the order information to each of the N messages.

The step of distributing the N messages based on the N code values may include arranging each sequence matched to each of the N message distribution groups for each of the N message distribution groups The N number of messages is compared with the index value assigned to each of the N messages, and the N number of messages are compared with the index values assigned to the N messages, Can be distributed by q.

In this case, according to an embodiment of the present invention, the step of distributing the N messages based on the N code values may include allocating, for each of the N message distribution groups, the N message distribution groups 1 "among the N number of code values constituting each sequence matched with the code value of " 1 ", and for each of the N message distribution groups, Q messages with allocated index values matching the array order of code values can be distributed.

According to another aspect of the present invention, there is provided a data distribution storage method using a relative difference set generated from a group having two-dimensional elements, the method comprising: after q pieces of N messages are dispersedly stored in the N data stores, Randomly selecting q of the N data stores based on the load command when the load command for N messages is granted, and selecting, from each of the q data stores, the q data stores And loading the stored messages extracted from each of the q data stores into the data corresponding to the load command for the N messages, .

According to an embodiment of the present invention, the N messages are MDS codes capable of recovering lost messages based on remaining messages when a loss of some messages occurs, When the stored messages are extracted from each of the N messages, the index value assigned to the extracted stored messages is compared with the index value assigned to the N messages to be included in the extracted stored messages among the N messages If it is determined that there is a remaining message that is not yet available, restores the remaining message based on the extracted stored messages, and stores the extracted stored messages and the restored remaining message according to a load command for the N messages You can load it as data.

The data distribution storage method using the relative difference set generated from the group having two-dimensional elements according to an embodiment of the present invention has been described above with reference to FIG. Here, a data distribution storage method using a relative difference set generated from a group having a two-dimensional element according to an embodiment of the present invention is a method for storing a relative difference set using a relative difference generated from a group having two- And the operation of the data distributed storage device 110 using the set can be corresponded to the configuration of the data distributed storage device 110, so a detailed description thereof will be omitted.

A data distribution storage method using a relative difference set generated from a group having two-dimensional elements according to an embodiment of the present invention can be implemented by a computer program stored in a storage medium for execution through a combination with a computer.

In addition, the data distribution storage method using the relative difference set generated from the group having two-dimensional elements according to an embodiment of the present invention may be implemented in a form of a program command that can be executed through various computer means, Lt; / RTI > The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

As described above, the present invention has been described with reference to particular embodiments, such as specific elements, and specific embodiments and drawings. However, it should be understood that the present invention is not limited to the above- And various modifications and changes may be made thereto by those skilled in the art to which the present invention pertains.

Accordingly, the spirit of the present invention should not be construed as being limited to the embodiments described, and all of the equivalents or equivalents of the claims, as well as the following claims, belong to the scope of the present invention .

110: Data Distributed Storage using Relative Difference Generated from Groups with Two Dimensional Elements
111: distributed group generation unit 112: message distribution assignment unit
113: Data storage unit 114: First sequence generator
115: second sequence generation unit 116: sequence matching unit
117: allocation unit 118: index allocation unit
119: storage message extracting unit 120: data loading unit
101, 102, 103: N data stores

Claims (16)

N - N is a natural number satisfying the condition of N = q ² , and q is a prime number of odd number. When a distributed storage command is applied to N data stores for the messages, A distribution group generation unit for generating N message distribution groups;
And distributing the N messages to each of the N message distribution groups by q, wherein messages distributedly distributed to the N message distribution groups may overlap each other; And
A data storage unit for separating and storing q messages allocated to each of the N message distribution groups into N data distribution groups and a data storage corresponding to each message distribution group among the N data storage groups,
Lt; / RTI >
The message distribution assignment unit
For a group Z _q of additions with a q number of elements, K - K is

Defined as - relative to G-G

- the relative difference set D - the relative difference set D

(Q, q, q, 1), a = 1 and having a parameter of (q, q, q, 1) _{(i 1, i 2) -} i 1, i 2 are both circle hydrogen contained in Z _q -, by "1" and "0" is assigned a code value of one for the two-dimensional element, represented by, N of " 1 "and" 0 "code values;
The right coset of the relative difference set D is calculated for all the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G to generate N-1 right co-sets, 1 "and "0" for two-dimensional elements represented by (i ₁ , i ₂ ) included in G based on q two-dimensional elements included in each of the N- Generating a second sequence by generating N-1 second sequences by generating a sequence consisting of N code values of "1" and "0 "part;
A sequence matching unit for matching the first sequence and the N-1 second sequences one by one for each of the N message distribution groups; And
For each of the N message distribution groups, an allocation unit for distributing the N messages by q, based on N code values constituting each sequence matched to each of the N message distribution groups,
And a relative difference set generated from the group having two-dimensional elements including the two-dimensional elements. delete The method according to claim 1,
The first sequence generator
Assigns a code value of "1" to q two-dimensional elements included in the relative difference set D among the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G, By assigning a code value of "0 " to remaining two-dimensional elements not included in the set D, generates the first sequence composed of N code values of" 1 &
The second sequence generator
And for each of the N-1 uxcosets, q q 2-dimensional elements contained in each of the N-1 uxcosets among the two-dimensional elements represented by (i ₁ , i ₂ ) Quot; 1 "and" 1 "are assigned to the remaining two-dimensional elements not included in each of the N-1 &Quot; 0 "and a code value of " 0 ". The method according to claim 1,
The message distribution assignment unit
An index assigning unit for assigning an index value associated with the order information to each of the N messages,
Further comprising:
The assigning unit
For each of the N message distribution groups, an arrangement order of N code values constituting each sequence matched to each of the N message distribution groups is compared with an index value assigned to each of the N messages. And a data using a relative difference set generated from a group having a two-dimensional element for distributing the N messages by q, based on a code value of an array order corresponding to an index value assigned to the N messages Distributed storage. 5. The method of claim 4,
The assigning unit
For each of the N message distribution groups, a code value of "1 " existing in q among N code values constituting each sequence matched to each of the N message distribution groups is checked, For each of the message distribution groups, q messages having a two-dimensional element for distributing q messages allocated with an index value corresponding to the code value of the "1" Data distributed storage using relative difference sets. 6. The method of claim 5,
Wherein when a load command for the N messages is applied after the N messages are distributedly distributed in the N data stores, q data of the N data stores based on the load command, A store message extractor for randomly selecting stores and for extracting q stored messages stored in each of the q data stores from each of the q data stores; And
A data load unit for loading stored messages extracted from each of the q data stores into data corresponding to a load command for the N messages,
And a second difference element having a second difference value. The method according to claim 6,
The N messages are Maximum Distance Separable (MDS) codes capable of recovering lost messages based on remaining messages when loss of some messages occurs,
The data loading unit
When the stored messages are extracted from each of the q data stores, the index value assigned to the extracted stored messages is compared with the index value assigned to the N messages, Recovering the stored messages based on the extracted stored messages and storing the extracted stored messages and the recovered residual messages in the N messages Dimensional element loaded with data according to a load instruction to the data storage device. N - N is a natural number satisfying the condition of N = q ² , and q is a prime number of odd number. When a distributed storage command is applied to N data stores for the messages, Creating N message distribution groups;
Distributing the N messages for each of the N message distribution groups by q; messages distributedly distributed to the N message distribution groups may overlap each other; And
Separating and storing q messages allocated to each of the N message distribution groups into a data store corresponding to each message distribution group among the N message storage groups for each of the N message distribution groups
Lt; / RTI >
The step of distributing the N messages by q for each of the N message distribution groups
For a group Z _q of additions with a q number of elements, K - K is

Defined as - relative to G-G

- the relative difference set D - the relative difference set D

(Q, q, q, 1), a = 1 and having a parameter of (q, q, q, 1) _{(i 1, i 2) -} i 1, i 2 are both circle hydrogen contained in Z _q -, by "1" and "0" is assigned a code value of one for the two-dimensional element, represented by, N of " Generating a first sequence of code values of " 1 "and" 0 ";
The right coset of the relative difference set D is calculated for all the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G to generate N-1 right co-sets, 1 "and "0" for two-dimensional elements represented by (i ₁ , i ₂ ) included in G based on q two-dimensional elements included in each of the N- Generating N-1 second sequences by generating a sequence of code values of N "1" and "0 " by assigning one code value to each of the N-1 uxcodes;
Matching the first sequence and the N-1 second sequences one by one for each of the N message distribution groups; And
Distributing, for each of the N message distribution groups, q messages of N messages based on N code values constituting each sequence matched to each of the N message distribution groups
And a second set of two-dimensional elements including the second set of elements. delete 9. The method of claim 8,
The step of generating the first sequence
Assigns a code value of "1" to q two-dimensional elements included in the relative difference set D among the two-dimensional elements represented by (i ₁ , i ₂ ) included in the G, By assigning a code value of "0 " to remaining two-dimensional elements not included in the set D, generates the first sequence composed of N code values of" 1 &
Wherein generating the N-1 second sequences comprises:
And for each of the N-1 uxcosets, q q 2-dimensional elements contained in each of the N-1 uxcosets among the two-dimensional elements represented by (i ₁ , i ₂ ) Quot; 1 "and" 1 "are assigned to the remaining two-dimensional elements not included in each of the N-1 &Quot; 0 "and a code value of " 0 " to generate N-1 second sequences. 9. The method of claim 8,
The step of distributing the N messages by q for each of the N message distribution groups
Assigning an index value associated with the order information to each of the N messages
Further comprising:
The step of distributing the N messages by q, based on the N code values,
For each of the N message distribution groups, an arrangement order of N code values constituting each sequence matched to each of the N message distribution groups is compared with an index value assigned to each of the N messages. And a data using a relative difference set generated from a group having a two-dimensional element for distributing the N messages by q, based on a code value of an array order corresponding to an index value assigned to the N messages Distributed storage method. 12. The method of claim 11,
The step of distributing the N messages by q, based on the N code values,
For each of the N message distribution groups, a code value of "1 " existing in q among N code values constituting each sequence matched to each of the N message distribution groups is checked, For each of the message distribution groups, q messages having a two-dimensional element for distributing q messages allocated with an index value corresponding to the code value of the "1" A method of data distribution storage using relative difference sets. 13. The method of claim 12,
Wherein when a load command for the N messages is applied after the N messages are distributedly distributed in the N data stores, q data of the N data stores based on the load command, Randomly selecting repositories and extracting q stored messages stored in each of the q data repositories from each of the q data repositories; And
Loading the stored messages extracted from each of the q data stores into q corresponding to a load command for the N messages
And a second set of two-dimensional elements further comprising a second set of two-dimensional elements. 14. The method of claim 13,
The N messages are Maximum Distance Separable (MDS) codes capable of recovering lost messages based on remaining messages when loss of some messages occurs,
The loading step
When the stored messages are extracted from each of the q data stores, the index value assigned to the extracted stored messages is compared with the index value assigned to the N messages, Recovering the stored messages based on the extracted stored messages and storing the extracted stored messages and the recovered residual messages in the N messages Dimensional element loaded with data according to a load instruction for the data to be distributed. A computer-readable recording medium recording a program for performing the method of any one of claims 8, 10, 11, 12, 13, A computer program stored in a storage medium for executing the method of any one of claims 8, 10, 11, 12, 13 or 14 in combination with a computer.

Images