KR102197993B1 - Data distribution storage apparatus and method using partitioned iteration code - Google Patents

Abstract

Disclosed is an apparatus and method for distributed storage of data using divisional repetition codes. The present invention provides a method of distributing and storing a plurality of messages in several data stores, allowing the plurality of messages to be distributed and stored using only a small number of data stores, and at the same time, at least among data stores in which messages are distributed and stored. The present invention relates to an efficient data distribution storage technique capable of preventing data loss for messages by allowing all of the plurality of messages to be loaded even if only the data stores of are selected and accessed.

Description

Data distribution storage device and method using division and repetition codes {DATA DISTRIBUTION STORAGE APPARATUS AND METHOD USING PARTITIONED ITERATION CODE}

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

In recent years, as the spread of the Internet is active, various Internet service providers have appeared. These Internet service providers provide e-mail services, blogs, cafes, and various services such as cloud data storage services.

In particular, in recent years, as the popularity of social network services has increased rapidly, Internet service providers providing social network services are also increasing very much.

These Internet service providers must store large amounts of data uploaded from multiple client terminals in their own data storage to provide email services, blogs, cafes, cloud services, and social network services. If it is lost in the storage, it cannot provide normal services to customers, so various efforts are being made to prevent data loss.

As part of such efforts, a method of backing up user messages that should be stored in data stores is being introduced. This data backup method is one of the most ideal methods for coping with data loss in that even if the original message is lost, the original message can be replaced by using the backed up message.

In this way, if one original message is stored in a plurality of data stores several times, when a data load command for the original message is applied, an arbitrary data store with the fastest operation speed among the plurality of data stores Since the message corresponding to the original message can be directly loaded from the source, it is possible to provide a faster data loading speed than the method of storing the original message in one data storage.

However, the existing data backup method only introduces a method of simply creating a copy message for the original message and storing it in a plurality of data stores without a separate data storage standard, so it is necessary to store multiple messages in the data storage. In this case, there is an inefficiency such as having to secure a large data storage space or have a large number of data storage.

Therefore, in relation to the method of distributing and storing a plurality of messages in several data stores, it is possible to distribute the plurality of messages even by using only a small number of data stores, and a plurality of data in which data is distributedly stored. There is a need for research on an efficient data distribution storage technique that allows all of the plurality of messages to be loaded even if only the minimum number of data stores among the stores are selected and accessed.

Korean Registered Patent Publication No. 10-1893566 (2018.08.30) Korean Patent Publication No. 10-1741186 (2017.05.29)

The present invention provides a method of distributing and storing a plurality of messages in several data stores, allowing the plurality of messages to be distributed and stored using only a small number of data stores, and at the same time, at least among data stores in which messages are distributed and stored. By allowing all of the plurality of messages to be loaded even if only the data storages of are selected and accessed, an efficient data distribution storage technique capable of preventing data loss for messages is provided.

In the data distribution storage device using the divisional repetition code according to an embodiment of the present invention, N-N is a natural number that satisfies the condition of N=tv, v is a natural number that satisfies the condition of v=6t-k, and k and t is a natural number of 2 or more-When a command for distributed storage to v data stores is applied for 2 messages, a distribution group generator that generates v message distribution groups corresponding to the v data stores, "0 G, which is a set for an additive group containing integers from "to "v-1" as elements-the number of elements in G is v-a plurality of k elements for Among the subsets, all remaining values calculated by dividing the difference between two different elements among the k elements included in each of the plurality of subsets by v represent "0" in the elements included in G. A first subset selection unit that selects t-1 first subsets that are found to match only one other element except for, and the t-1 first subsets are excluded from the plurality of subsets. Among the remaining subsets, all the remaining values calculated by dividing the difference between two different elements by v among k elements included in each of the remaining subsets are included as elements of their own subset. A second subset selection unit for selecting 2 subsets, distributing the N messages to the v messages based on k elements included in each of the t-1 first subsets and the second subset Create a sequence to allocate to each of the groups, and based on the sequence, for each of the v message distribution groups, the N messages tk-tk means the product of t and k-each distributed allocation-the v The messages distributedly allocated to the message distribution groups may overlap each other-the message distribution allocation unit and the tk messages allocated to each of the v message distribution groups Each of the v message distribution groups includes a data storage unit that separates and stores data in a data store corresponding to each message distribution group among the v data stores.

In addition, in the data distribution storage method using the divided repetition code according to an embodiment of the present invention, N-N is a natural number that satisfies the condition of N = tv, and v is a natural number that satisfies the condition of v = 6t-k, k and t are natural numbers greater than or equal to 2-when a command for distributed storage to v data stores is applied for 2 messages, generating v message distribution groups corresponding to the v data stores, "0" Among a plurality of subsets having k elements for G, which is a set for an addition group containing integers from "v-1" as elements-the number of elements included in G is v- All remaining values calculated by dividing the difference between two different elements by v among the k elements included in each of the plurality of subsets are one different from the elements included in G except "0" Selecting t-1 first subsets that are found to match only the element, the remaining subsets from among the remaining subsets in which the t-1 first subsets are excluded from the plurality of subsets Selecting one second subset in which all the remaining values calculated by dividing the difference between two different elements by v among the k elements included in each of the k elements are included as elements of their own subset, the t -A sequence for allocating the N messages to each of the v message distribution groups is generated based on k elements included in each of one first subset and the second subset, and the sequence Based on the N messages for each of the v message distribution groups, tk-tk means the product of t and k-distributed allocation by each-messages distributed distributedly allocated to the v message distribution groups may overlap each other Yes-step of, and tk messages allocated to each of the v message distribution groups, for each of the v message distribution groups, and each message distribution group among the v data stores And separately storing it in a data storage corresponding to.

The present invention provides a method of distributing and storing a plurality of messages in several data stores, allowing the plurality of messages to be distributed and stored using only a small number of data stores, and at the same time, at least among data stores in which messages are distributed and stored. By allowing all of the plurality of messages to be loaded even if only the data stores of are selected and accessed, it is possible to provide an efficient data distribution storage technique capable of preventing data loss for messages.

1 is a diagram showing the structure of a data distribution storage device using divisional repetition codes according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating an operation of a data distribution storage device using divisional repetition codes according to an embodiment of the present invention.
3 is a flowchart illustrating a data distribution storage method using divided repetition codes according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. This description is not intended to limit the present invention to a specific embodiment, it is to be understood to include all changes, equivalents, and substitutes included in the spirit and scope of the present invention. While describing each drawing, similar reference numerals have been used for similar components, and unless otherwise defined, all terms used in the present specification including technical or scientific terms refer to common knowledge in the technical field to which the present invention belongs. It has the same meaning as commonly understood by someone who has it.

In this document, when a part "includes" a certain component, it means that other components may be further included rather than excluding other components unless otherwise stated. In addition, in various embodiments of the present invention, each component, functional blocks or means may be composed of one or more sub-components, and the electrical, electronic, and mechanical functions performed by each component are electronic. A circuit, an integrated circuit, and an application specific integrated circuit (ASIC) may be implemented with various known devices or mechanical elements, and may be implemented separately or two or more may be integrated into one.

On the other hand, the blocks of the attached block diagram and the steps in the flowchart are computer program instructions that are mounted on a processor or memory of equipment capable of processing data such as a general-purpose computer, a special-purpose computer, a portable notebook computer, and a network computer to perform specified functions. It can be interpreted as meaning. Since these computer program instructions can be stored in a memory provided in a computer device or in a memory readable by a computer, the functions described in the blocks in the block diagram or in the steps in the flowchart are produced as a product containing the instruction means to perform this. It could be. In addition, each block or each step may represent a module, segment, or part of code including one or more executable instructions for executing the specified logical function(s). In addition, it should be noted that in some alternative embodiments, functions mentioned in blocks or steps may be executed in a different order. For example, two blocks or steps shown in succession may be performed substantially simultaneously or may be performed in reverse order, and in some cases, some blocks or steps may be omitted.

1 is a diagram showing the structure of a data distribution storage device using divisional repetition codes according to an embodiment of the present invention.

Referring to FIG. 1, a data distribution storage device 110 using a divided repetition code according to the present invention includes a distribution group generation unit 111, a first subset selection unit 112, and a second subset selection unit 113. , A message distribution allocation unit 114 and a data storage unit 115.

When a distributed storage command to the v data stores 101, 102, 103 is applied to the N messages, the distribution group generator 111 corresponds to the v data stores 101, 102, 103. Create v message distribution groups.

Here, N is a natural number that satisfies the condition of N=tv, v is a natural number that satisfies the condition of v=6t-k, and k and t are natural numbers of 2 or more.

In addition, the message refers to a kind of data symbol to be stored in the v data stores 101, 102, and 103, and the N messages are different data symbols.

The first subset selection unit 112 includes G, which is a set for an additive group including integers from "0" to "v-1" as elements (the number of elements included in G is Among the plurality of subsets having k elements for v), all remaining values calculated by dividing the difference between two different elements by v among the k elements included in each of the plurality of subsets are the From the elements included in G, t-1 first subsets that are found to match only one element other than "0" are selected.

In connection, the process of selecting the t-1 first subsets will be described in more detail as follows.

라고 하자.First, when there is G, which is a set containing integers from "0" to "v-1" as elements, a plurality of subsets having k elements for G

Let's say.

Here, if the remaining value calculated by dividing the difference between two different elements by v among k elements included in each of the plurality of subsets B _i is D, then D is according to Equation 1 below. Is calculated. If the difference between the two different elements becomes a negative value, the difference between the two different elements may be a residual value divided by v for the negative value according to Equation 1 below.

In this case, the first subset selection unit 112 includes each of the remaining values D calculated according to Equation 1 among the plurality of subsets B _i except for "0" in the elements included in G. It is possible to select t-1 first subsets that are found to match only one other element.

For example, assuming t=3, k=3, v=15, "0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14" is an element When there is G, which is a set containing as, there may be a plurality of subsets of G containing three elements. Since each of the plurality of subsets includes 3 elements, there may be a total of 6 differences between the two different elements included in each subset. In this case, the first subset selection unit 112 When the difference between the two elements is divided by “15” according to the operation of Equation 1, the six remaining values calculated are “0, 1, 2, 3, 4, 5, 6, 7, which are included in G. Among the integers 8, 9, 10, 11, 12, 13, 14", subsets found to match only one integer excluding "0" may be selected as the first subset.

In the above example, "{0, 1, 4}" and "{0, 2, 9}" may exist as the first subset. Relatedly, in the case of "{0, 1, 4}", the operation value according to Equation 1 between two different elements may be "1, 3, 4, 11, 12, 14", and in this case, "1 , 3, 4, 11, 12, 14" each matches only one of the integers included in G, so the subset "{0, 1, 4}" is selected as one of the first subsets. Can be. And, in the case of "{0, 2, 9}", the operation value according to Equation 1 between two different elements may be "2, 6, 7, 8, 9, 13". In this case, "2, Since each of 6, 7, 8, 9, 13" matches only one of the integers included in G, a subset of "{0, 2, 9}" can also be selected as one of the first subsets. I can.

The second subset selection unit 113 is selected from among k elements included in each of the remaining subsets among the remaining subsets from which the t-1 first subsets are excluded from the plurality of subsets. A second subset is selected in which all remaining values calculated by dividing the difference between two different elements by v are included as elements of its own subset.

In relation to the second subset selection unit 113, the entire operation value according to Equation 1 between two elements in each subset among the remaining subsets excluding the t-1 first subsets is itself It is possible to select one second subset containing as elements of a subset of.

For example, as in the example described above, assuming that t=3, k=3, v=15, "0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, When it is assumed that there is G, which is a set including 13 and 14" as an element, the second subset selection unit 113 includes "{0, 1, 4}", "{0, 2," 9}", among the remaining subsets, a subset in which all of the calculation values according to Equation 1 between two elements are included as its own elements may be selected as the second subset.

In this example, "{0, 5, 10}" may exist as the second subset. Relatedly, in the case of "{0, 5, 10}", the operation value according to Equation 1 between two different elements may be "5, 5, 10, 5, 5, 10", and in this case, "5 , 5, 10, 5, 5, 10" are all contained in the elements of their own subset called "{0, 5, 10}", so the subset "{0, 5, 10}" It may be selected as a second subset.

As a result, when t=3, k=3, and v=15, the first subset selection unit 112 includes "{0, 1, 4}", "{0, 2" among a plurality of subsets for G. , 9}" may be selected, and the second subset selection unit 113 may select one second part of "{0, 5, 10}" among a plurality of subsets for G You can choose a set.

In this way, when the condition of v=6t-k is satisfied, a plurality of subsets having k elements for G, which is a set of addition groups containing integers from "0" to "v-1" as elements Among them, there are t-1 first subsets and one second subset, and the group of the t-1 first subsets and the one second subset is (v, k, 1) It may be referred to as a partial cyclic difference family having a parameter of.

The message distribution allocation unit 114 allocates the N messages to each of the v message distribution groups based on k elements included in each of the t-1 first subsets and the second subsets. A sequence is generated, and the N messages are distributedly allocated to each of the v message distribution groups based on the sequence, tk (tk means a product of t and k).

In this case, messages distributedly allocated to the v message distribution groups may overlap each other.

The data storage unit 115 corresponds to the tk messages allocated to each of the v message distribution groups for each of the v message distribution groups, and to each message distribution group among the v data stores 101, 102, 103 Separately stored in the data storage.

At this time, according to an embodiment of the present invention, the data distribution allocation unit 114 includes the first sequence generation unit 116, the second sequence generation unit 117, the sequence matching unit 118, and the allocation unit 119. Can include.

The first sequence generator 116 is based on the k elements included in each of the t-1 first subsets and the second subset, from "0" to "v-1" included in the G. By assigning code values of "1" and "0" in the order of "0" to "v-1" for integers from "to", each of the t-1 first subsets and the second subset is A first sequence consisting of a total of N code values of "1" and "0" by generating t partial sequences consisting of v corresponding code values of "1" and "0", and combining the t partial sequences (N=tv, and the first sequence is a combination of t partial sequences consisting of v code values of “1” and “0”, so the code values of “1” and “0” constituting the first sequence are A total of N) is generated.

In this case, the first sequence generation unit 116 is included in each of the t-1 first subsets and the second subsets among integers from "0" to "v-1" included in the G. A code value of "1" is assigned to integers that match the k elements, and matches k elements included in each of the t-1 first subsets and the second subset. Code values of "1" and "0" corresponding to the t-1 first subsets and the second subsets respectively by assigning a code value of "0" to the remaining integers The t partial sequences composed of may be generated.

In addition, the second sequence generator 117 performs a cyclic shift operation v-1 times in a predetermined direction for each of the t partial sequences composed of the v code values of "1" and "0". While performing, a second sequence is generated by combining the t partial sequences on which the cyclic shift operation is completed each time each cyclic shift operation is performed, thereby generating a total of v-1 second sequences.

In relation to the process of generating the first sequence and the v-1 second sequence, an example will be described as follows.

First, when it is said that a total of t subsets exist as the selection of the t-1 first subsets and the second subset is completed, each of the total t subsets is included in the G A code value of "1" is assigned to integers that match k elements included in each of the total t subsets of integers from "0" to "v-1", and a total of t If a code value of "0" is assigned to the remaining integers that do not match the k elements included in each of the subsets, t sequences consisting of v "1" and "0" code values are Can be obtained.

At this time, the t sequences are referred to as t subsequences, and when the t subsequences are arranged in a line and combined with each other, a total of N (N=tv) code values of “1” and “0” It is possible to obtain a first sequence consisting of.

And, while performing a cyclic shift operation v-1 times in a selected direction for each of the t partial sequences consisting of the v code values of "1" and "0", each time each cyclic shift operation is performed When a second sequence is generated by arranging and combining the t partial sequences on which the cyclic shift operation has been completed, a total of v-1 second sequences may be additionally obtained.

For example, assuming k=3, t=3, and v=6t-k, if v is 15, "0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 , 11, 12, 13, 14" as an element in G, the first subset includes "{0, 1, 4}" and "{0, 2, 9}", and the first In the second subset, "{0, 5, 10}" exists.

At this time, among the integers from "0" to "14" included in the G, the three subsets "{0, 1, 4}", "{0, 2, 9}", "{0," 5, 10}" Allocates a code value of "1" to integers that match the three elements included in each of the three subsets "{0, 1, 4}", "{0" , 2, 9}", "{0, 5, 10}" If the code value of "0" is assigned to the remaining integers that do not match the 3 elements in each, "110010000000000", " Three partial sequences of 101000000100000" and "100001000010000" can be obtained.

At this time, if three partial sequences of "110010000000000", "101000000100000" and "100001000010000" are arranged in a row and combined, a first sequence of "110010000000000101000000100000100001000010000" consisting of 45 code values of "1" and "0" Can be obtained.

And, while performing a cyclic shift operation 14 times in the right direction for the three partial sequences “110010000000000”, “101000000100000”, and “100001000010000”, three cyclic shift operations are completed each time each cyclic shift operation is performed. When a second sequence is generated by arranging and combining partial sequences in a line, a total of 14 second sequences may be additionally obtained.

Relatedly, if the cyclic shift operation is performed once in the right direction for the three subsequences of "110010000000000", "101000000100000" and "100001000010000", three types of "011001000000000", "010100000010000", and "010000100001000" Additional partial sequences can be obtained, and when three additional partial sequences, "011001000000000", "010100000010000", and "010000100001000", are arranged in a line with each other and combined, an additional sequence of 011001000000000010100000010000010000100001000" can be obtained. In this way When an additional sequence is acquired while performing a cyclic shift operation in the right direction for the three partial sequences 14 times, since the cyclic shift operation is performed 14 times, a total of 14 sequences can be additionally obtained.

In this way, when a sequence is generated, a total of 15 sequences can be finally generated, and at this time, the number of code values of "1" in each sequence is the same as 9.

In this way, when the generation of the first sequence and the v-1 second sequence is completed, the sequence matching unit 118 performs the first sequence and the v-1 second sequence for each of the v message distribution groups. Match them one by one.

And, the allocation unit 119, for each of the v message distribution groups, tk each of the N messages based on N code values constituting each sequence matched to each of the v message distribution groups. Distributed allocation.

In this case, according to an embodiment of the present invention, the message distribution allocating unit 114 may further include an index allocating unit 120 that allocates an index value associated with the sequence number information for each of the N messages.

At this time, the allocator 119, for each of the v message distribution groups, an arrangement order of N code values constituting each sequence matched to each of the v message distribution groups and each of the N messages After comparing the index values allocated to the N messages, tk may be distributed and allocated to the N messages based on a code value in an array order that matches the index values allocated to the N messages.

At this time, according to an embodiment of the present invention, the allocator 119 for each of the v message distribution groups, among the N code values constituting each sequence matched to each of the v message distribution groups After checking the code value of "1" each of tk, for each of the v message distribution groups, an index value that matches the arrangement order of the code value of "1" among the N messages is allocated Tk messages can be distributedly allocated.

In this case, according to an embodiment of the present invention, the data distribution storage device 110 using the divisional repetition code may further include a storage message extraction unit 121 and a data load unit 122.

The stored message extracting unit 121 distributes and stores tk of the N messages in v data stores 101, 102, 103, and then when a load command for the N messages is applied, the load Tk stored messages stored in each of the tk data stores from each of the tk data stores randomly selected from among v data stores (101, 102, 103) based on the command Extract them.

The data load unit 122 loads tk stored messages extracted from each of the tk data stores as data according to a load command for the N messages.

In this case, according to an embodiment of the present invention, the N messages may be a Maximum Distance Separable (MDS) code that enables recovery of lost messages based on the remaining messages when some messages are lost. .

Here, the MDS code is a code capable of restoring some lost messages based on the remaining messages when some of the messages are lost. By performing MDS coding on a predetermined original message, MDS It can be used to restore the lost message when loss of some messages among messages changed to the MDS code after changing to the code occurs. In this case, if the original message is to be generated again from the MDS code, the original message may be restored by performing MDS decoding. Therefore, when important messages are stored, by changing them to MDS codes and storing them, it is possible to prepare in advance in case of loss of messages.

At this time, when tk stored messages are extracted from each of the tk data stores, the data load unit 122 compares an index value allocated to the extracted stored messages with an index value allocated to the N messages. Thus, if it is determined that there is a residual message not included in the extracted stored messages among the N messages, the residual message is restored based on the extracted stored messages, and the extracted stored messages and the restoration The remaining messages can be loaded as data according to the load command for the N messages.

Hereinafter, with reference to FIG. 2, the operation of the data distribution storage device 110 using the divided repetition code according to the present invention will be described in detail by way of example.

FIG. 2 is a diagram for explaining the operation of the data distribution storage device 110 using divided repetition codes according to an embodiment of the present invention.

First, assuming that k=3, t=3, v=6t-k, v=15, N=tv, and N=45, v data stores 101, 102, 103), as shown in Figure 2, data store 1 (201), data store 2 (202), data store 3 (203), data store 4 (204), data store 5 (205), data store 6 ( 206), Data Store 7(207), Data Store 8(208), Data Store 9(209), Data Store 10(210), Data Store 11(211), Data Store 12(212), Data Store 13(213) ), data storage 14(214), data storage 15(215), and data storage 1~15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210) as N=45 , 211, 212, 213, 214, 215), it is assumed that the number of messages to be stored is 45.

At this time, the distribution group generation unit 111 is a data storage 1 to 15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, for 45 messages) 215), 15 corresponding to data storage 1 to 15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215) Two message distribution groups 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415 can be created.

In addition, the index allocating unit 120 included in the message distribution allocating unit 114 may allocate an index value associated with the sequence number information to the 45 messages, and the index allocating unit 120 Assuming that index values from 1 to 45 are allocated in order, the 45 messages are "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14" , m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39 , m40, m41, m42, m43, m44, m45".

Since it is assumed that k=3, t=3, and v=15, the first subset selection unit 112 is "0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 Among a plurality of subsets having three elements for G, which is a set containing 12, 13, 14" as elements, two different elements among the three elements included in each of the plurality of subsets It is possible to select two first subsets in which all remaining values calculated by dividing the difference between them by 15 are found to match only one different element except for “0” in the elements included in G.

In addition, the second subset selection unit 113 is selected from among the three elements included in each of the remaining subsets among the remaining subsets from which the two first subsets are excluded from the plurality of subsets. One second subset may be selected in which all remaining values calculated by dividing the difference between two different elements by 15 are included as elements of the own subset.

At this time, the first sequence generation unit 116 is based on the three elements included in each of the two first subsets and the one second subset, from "0" to "" By assigning code values of "1" and "0" in the order of "0" to "14" for integers up to 14", 3 partial sequences consisting of 15 code values of "1" and "0" And, by combining the three partial sequences, a first sequence composed of a total of 45 code values of "1" and "0" may be generated.

In this regard, assume that the first subset is "{0, 1, 4}" and "{0, 2, 9}", and the second subset is "{0, 5, 10}".

At this time, the first sequence generating unit 116 is included in the G for each of "{0, 1, 4}", "{0, 2, 9}", and "{0, 5, 10}" Three elements included in each of "{0, 1, 4}", "{0, 2, 9}", and "{0, 5, 10}" among integers from "0" to "14" Allocates a code value of "1" to the integers that match with and is included in each of "{0, 1, 4}", "{0, 2, 9}", and "{0, 5, 10}" By assigning a code value of "0" to the remaining integers that do not match the three elements that exist, three subsequences of "110010000000000", "101000000100000" and "100001000010000" can be generated.

Then, the first sequence generation unit 116 arranges and combines three subsequences of "110010000000000", "101000000100000", and "100001000010000" in a row, so that a total of 45 code values of "1" and "0" The configured first sequence of 110010000000000101000000100000100001000010000 (301) may be obtained.

In addition, the second sequence generator 117 performs a cyclic shift operation 14 times in a selected direction of either the right or the left for the three partial sequences of "110010000000000", "101000000100000", and "100001000010000" ( In the present embodiment, a cyclic shift operation is performed in the right direction), and a second sequence is generated by combining three partial sequences in which the cyclic shift operation is completed each time each cyclic shift operation is performed. Second sequences 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315 may be generated.

This first sequence 11001000000001010000100000 (301) and 14 second sequences (302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315) are as described above. , It always has tk code values of "1". In this embodiment, since t=3, k=3 is assumed, 15 sequences (301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315) There are always 9 code values of "1" included in each.

Thus, when 15 sequences 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315 are generated, the sequence matching unit 118 is a data storage 15 message distribution groups (401, 402, 403) each corresponding to 1 to 15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215) , 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415) 15 sequences (301, 302, 303, 304, 305, 306, 307, 308, 309) , 310, 311, 312, 313, 314, 315) can be matched one by one.

Then, the allocation unit 119 for each of the 15 message distribution groups (401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415), 45 codes constituting each sequence matched to each of 15 message distribution groups (401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415) Based on the value, the 45 messages, "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" Can be assigned.

Specifically, the allocation unit 119 is matched to each of the 15 message distribution groups 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415 The order of the arrangement of 45 code values constituting each sequence and the 45 messages such as "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, After comparing the index values assigned to each of m41, m42, m43, m44, m45", the 45 messages such as "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12" , m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37 , m38, m39, m40, m41, m42, m43, m44, m45" based on the code value of the sequence matching the index value assigned to the 45 messages "m1, m2, m3, m4, m5," m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" to 15 message distribution groups (401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415) Can be assigned.

At this time, the allocation unit 119 is matched to each of the 15 message distribution groups 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415 After checking the code value of "1", which is 9 out of 45 code values constituting each sequence, the 45 messages such as "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11" , m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36 , m37, m38, m39, m40, m41, m42, m43, m44, m45" of which 9 messages are assigned an index value that matches the arrangement order of the code value of "1" in 15 message distribution groups ( 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415).

For example, since the sequence of 110010000000000101000000100000100001000010000 (301) is matched for the message distribution group 1 (401), the allocation unit 119 checks the code value of "1" from 110010000000000101000000100000100001000010000 (301), and then the 45 messages "m1," m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45. Nine messages to which matching index values are allocated may be allocated to message distribution group 1 (401).

Relatedly, in 110010000000000101000000100000100001000010000(301), the code value of "1" is arranged at positions 1, 2, 5, 16, 18, 25, 31, 36, and 41, so the allocation unit ( 119) corresponds to positions 1, 2, 5, 16, 18, 25, 31, 36 and 41 for message distribution group 1 401, as shown in FIG. "M1, m2, m5, m16, m18, m25, m31, m36, m41" to which an index value is assigned can be allocated.

In this way, the allocator 119 for each of the 15 message distribution groups 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415 As shown in FIG. 2, each sequence matching each message distribution group (401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415) Based on the 45 messages, "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" to 9 Each can be distributed and allocated.

Thus, the 45 messages, "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" are all When the distribution allocation of 9 messages to 15 message distribution groups (401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415) is completed, the data storage unit ( 115) is 15 messages allocated to each of the 15 message distribution groups (401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415). Data storage 1~15 (201, 202, 203, 204, 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415) 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215) can be stored separately in the data storage corresponding to each message distribution group.

For example, the data storage unit 115 stores 9 messages allocated to the message distribution group 1 401, "m1, m2, m5, m16, m18, m25, m31, m36, m41", which is a corresponding data storage. Can be stored in storage 1 (201).

In this way, the data storage unit 115 has 15 message distribution groups 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415, respectively. Data storage 1~15 (for each of 15 message distribution groups (401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415) allocated 9 messages 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215) among the message distribution groups can be stored separately in the data storage corresponding to each message distribution group.

Thus, the 45 messages, "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" data After distributed storage in storage 1~15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215), the 45 messages, "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, When the load command for m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" is applied, the save message extractor (121) is the 9 data stores with the lowest data traffic among data stores 1 to 15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215) By selecting only randomly, it is possible to extract nine pieces of stored data stored in each of nine data stores, and the data load unit 122 includes the 45 pieces of data based on the stored data extracted through the nine data stores. Messages "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24 , m25, m26, m27, m28, m29, m30, m31, m32, Most messages among m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" can be loaded quickly.

At this time, if the 45 messages such as "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21 , m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" In the case of the MDS code, the data load unit 122 randomly selects data storage 1 to 15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215). When 9 stored messages are extracted from the 9 data stores selected accordingly, the index values allocated to the extracted stored messages and the 45 messages such as "m1, m2, m3, m4, m5, m6, m7, m8" , m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33 , m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" are compared, and the result of the comparison, the 45 messages of "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, Among m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45", there are residual messages that are not included in the extracted stored messages If it is confirmed that the stored messages are extracted, the remaining The messages are restored and the extracted stored messages and the restored residual messages are converted to the 45 messages of “m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14. , m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39 , m40, m41, m42, m43, m44, m45" can be loaded as data according to the load command.

In relation to the nine randomly selected data stores, data store 1 (201), data store 2 (202), data store 5 (205), data store 7 (207), data store 8 (208), data store 10 Speaking of (210), data storage 13 (213), data storage 14 (214), and data storage 15 (215), the storage message extraction unit 121 from the data storage 1 (201) "m1, m2, m5, m16 , m18, m25, m31, m36, m41" can be extracted, and "m2, m3, m6, m17, m19, m26, m32, m37, m42" can be extracted from data storage 2(202), and data "M5, m6, m9, m20, m22, m29, m35, m40, m45" can be extracted from storage 5(205), and "m7, m8, m11, m16, m22, m24" from data storage 7(207) , m32, m33, m42" can be extracted, and "m8, m9, m12, m17, m23, m25, m33, m34, m43" can be extracted from data storage 8 (208), and data storage 10 (210 ) "M10, m11, m14, m19, m25, m27, m35, m36, m45" can be extracted, and "m2, m13, m14, m22, m23, m28, m33, m38" from data storage 13(213) , m43" can be extracted, "m3, m14, m15, m23, m24, m29, m34, m39, m44" can be extracted from data storage 14(214), and "m1" from data storage 15(215) , m4, m15, m24, m25, m30, m35, m40, m45" can be extracted.

For this reason, the storage message extraction unit 121 is, in the end, data storage 1 (201), data storage 2 (202), data storage 5 (205), data storage 7 (207), data storage 8 (208), data storage From 10(210), data store 13(213), data store 14(214), data store 15(215) "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12 , m13, m14, m15, m16, m17, m18, m19, m20, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38 , m39, m40, m41, m42, m43, m44, m45" can be extracted.

At this time, the data load unit 122 includes the 45 messages "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18" , m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43 , m44, m45" and the extracted "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17," m18, m19, m20, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, The index values allocated to m44, m45" are compared with each other, and the 45 messages such as "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15", m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45", the extracted "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18" , m19, m20, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44 , you can check the remaining messages not included in "m45".

In relation to this, since "m21" is a residual message in this embodiment, the data load unit 122 checks "m21" as a residual message and then MDS-coded "m1, m2, m3, m4, m5, m6, m7" , m8, m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33 , m34, m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" from "m21" can be restored.

Accordingly, in the end, the data load unit 122 is the restored "m21" and data storage 1 (201), data storage 2 (202), data storage 5 (205), data storage 7 (207), data storage 8 (208), data store 10(210), data store 13(213), data store 14(214), data store 15(215) extracted from "m1, m2, m3, m4, m5, m6, m7, m8 , m9, m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34 , m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" by loading the 45 messages, "m1, m2, m3, m4, m5, m6, m7, m8, m9" , m10, m11, m12, m13, m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34 , m35, m36, m37, m38, m39, m40, m41, m42, m43, m44, m45" can all be loaded.

As a result, the data distribution storage device 110 using the divided repetition code according to the present invention includes the 45 messages "m1, m2, m3, m4, m5, m6, m7, m8, m9, m10, m11, m12, m13". , m14, m15, m16, m17, m18, m19, m20, m21, m22, m23, m24, m25, m26, m27, m28, m29, m30, m31, m32, m33, m34, m35, m36, m37, m38 , m39, m40, m41, m42, m43, m44, m45" in accordance with the first sequence and the second sequence generated based on the partial cyclic difference, data storage 1 to 15 (201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215) by distributing 9 units each, so when the data is loaded later, data can be restored even if only 9 data stores are accessed. You can apply to do it.

3 is a flowchart illustrating a data distribution storage method using divided repetition codes according to an embodiment of the present invention.

In step S310, v for N messages (N is a natural number that satisfies the condition of N=tv, v is a natural number that satisfies the condition of v=6t-k, and k and t are natural numbers of 2 or more). When a command for distributed storage to 2 data stores is applied, v message distribution groups corresponding to the v data stores are created.

In step S320, k elements for G (the number of elements included in G is v), which is a set for an addition group including integers from "0" to "v-1" as elements Among the plurality of subsets, all the remaining values calculated by dividing the difference between two different elements by v among k elements included in each of the plurality of subsets are " T-1 first subsets that are found to match only one element other than 0" are selected.

In step S330, among the remaining subsets from which the t-1 first subsets are excluded from the plurality of subsets, between two different elements among k elements included in each of the remaining subsets A second subset is selected in which all remaining values calculated by dividing the difference by v are included as elements of its own subset.

In step S340, a sequence for allocating the N messages to each of the v message distribution groups based on k elements included in each of the t-1 first subsets and the second subsets And, based on the sequence, distributed allocation of the N messages for each of the v message distribution groups (tk means the product of t and k) by tk (distributed allocation to the v message distribution groups) Messages can be duplicated with each other).

In step S350, tk messages allocated to each of the v message distribution groups are separately stored for each of the v message distribution groups in a data storage corresponding to each message distribution group among the v data stores. .

At this time, according to an embodiment of the present invention, in step S340, "included in G" based on k elements included in each of the t-1 first subsets and the second subsets By assigning code values of "1" and "0" in the order of "0" to "v-1" for integers from 0" to "v-1", the t-1 first subsets and T partial sequences consisting of v code values of “1” and “0” corresponding to each of the second subsets are generated, and the t partial sequences are combined to form a total of N “1” and “0”. A first sequence consisting of code values (N=tv, and the first sequence is a combination of t partial sequences consisting of v code values of “1” and “0”, and thus “1” constituting the first sequence and Generating a total of N code values of "0"), performing a cyclic shift operation in a direction selected for each of the t partial sequences consisting of the v code values of "1" and "0" -Generating a total of v-1 second sequences by combining the t partial sequences for which the cyclic shift operation is completed each time each cyclic shift operation is performed while performing -1 time, and the v Matching the first sequence and the v-1 second sequences one by one for each of the message distribution groups, and matching each of the v message distribution groups for each of the v message distribution groups. It may include the step of distributedly allocating tk of the N messages based on the N code values constituting each sequence.

In this case, according to an embodiment of the present invention, the step of generating the first sequence includes the t-1 first subsets of integers from "0" to "v-1" included in the G. And a code value of "1" to integers that match k elements included in each of the second subsets and the t-1 first subsets and the second subsets respectively By assigning a code value of "0" to the remaining integers that do not match the k elements included, the v number of "s" corresponding to each of the t-1 first subsets and the second subsets The t partial sequences composed of code values of 1" and "0" may be generated.

In addition, according to an embodiment of the present invention, step S340 may further include allocating an index value associated with sequence number information for each of the N messages.

In this case, the step of distributing and allocating tk of the N messages based on the N code values comprises configuring each sequence matched to each of the v message distribution groups for each of the v message distribution groups. The N messages are based on a code value in an array order that matches the index value assigned to the N messages after comparing the arrangement order of the N code values with the index value assigned to each of the N messages. It can be distributedly allocated by tk.

Further, according to an embodiment of the present invention, the step of distributing tk each of the N messages based on the N code values includes, for each of the v message distribution groups, each of the v message distribution groups. After checking the code value of "1", which is present by tk among the N code values constituting each sequence matched to, for each of the v message distribution groups, the "1" of the N messages It is possible to distribute allocation of tk messages with index values that match the order of the code values.

At this time, according to an embodiment of the present invention, in the data distribution storage method using the divisional repetition code, after tk of the N messages are distributed and stored in the v data stores, a load command for the N messages is performed. When authorized, randomly selects tk data stores among the v data stores based on the load command, and receives tk store messages stored in each of the tk data stores from each of the tk data stores. It may further include extracting and loading stored messages extracted by tk from each of the tk data stores as data according to a load instruction for the N messages.

At this time, according to an embodiment of the present invention, the N messages are MDS codes capable of recovering lost messages based on the remaining messages when some messages are lost, and the loading step includes the tk data stores When tk stored messages are extracted from each, the index value allocated to the extracted stored messages and the index value allocated to the N messages are compared and included in the extracted stored messages among the N messages. If it is determined that there is a residual message that is not already present, the residual message is restored based on the extracted stored messages, and the extracted stored messages and the restored residual message are loaded according to the load command for the N messages. Can be loaded with data.

In the above, a data distribution storage method using divided repetition codes according to an embodiment of the present invention has been described with reference to FIG. 3. Here, the data distribution storage method using the divisional repetition code according to an embodiment of the present invention may correspond to the configuration of the operation of the data distribution storage device 110 using the divisional repetition code described with reference to FIGS. 1 and 2. Therefore, a more detailed description thereof will be omitted.

The data distribution storage method using divisional repetition codes according to an embodiment of the present invention may be implemented as a computer program stored in a storage medium for execution through combination with a computer.

In addition, the data distribution storage method using the divisional repetition code according to an embodiment of the present invention may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. 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 specially designed and configured for the present invention, or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -A hardware device specially configured to store and execute program instructions such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of the program instructions include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, in the present invention, specific matters such as specific components, etc., and limited embodiments and drawings have been described, but this is provided only to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments. , If a person of ordinary skill in the field to which the present invention belongs, various modifications and variations are possible from these descriptions.

Therefore, the spirit of the present invention is limited to the described embodiments and should not be defined, and all things that are equivalent or equivalent to the claims as well as the claims to be described later fall within the scope of the spirit of the present invention. .

110: Data distributed storage device using divisional repetition code
111: distribution group generation unit 112: first subset selection unit
113: second subset selection unit 114: message distribution allocation unit
115: data storage unit 116: first sequence generation unit
117: second sequence generation unit 118: sequence matching unit
119: allocation unit 120: index allocation unit
121: stored message extraction unit 122: data load unit
101, 102, 103: data stores

Claims (16)

N-N is a natural number that satisfies the condition of N=tv, v is a natural number that satisfies the condition of v=6t-k, and k and t are natural numbers greater than or equal to 2-v data stores for 2 messages A distribution group generating unit that generates v message distribution groups corresponding to the v data stores when the distributed storage command of is applied;
G, which is a set for an additive group containing integers from "0" to "v-1" as elements-the number of elements in G is v-has k elements for Among the plurality of subsets, all remaining values calculated by dividing the difference between two different elements by v among the k elements included in each of the plurality of subsets are "0" in the elements included in G. A first subset selection unit for selecting t-1 first subsets that are found to match only one element other than "";
Of the remaining subsets excluding the t-1 first subsets from the plurality of subsets, the difference between two different elements among k elements included in each of the remaining subsets is divided by v A second subset selection unit for selecting one second subset in which all remaining values to be calculated are included as elements of its own subset;
A sequence for allocating the N messages to each of the v message distribution groups is generated based on k elements included in each of the t-1 first subsets and the second subset, and the Based on the sequence, the N messages for each of the v message distribution groups are tk-tk means the product of t and k-each distributed allocation-messages distributed distributedly allocated to the v message distribution groups overlap each other Can be-a message distribution allocation unit; And
A data storage unit that separates and stores tk messages allocated to each of the v message distribution groups for each of the v message distribution groups, in a data store corresponding to each message distribution group among the v data stores
Data distributed storage device using a divided repeat code comprising a. The method of claim 1,
The message distribution allocation unit
"0" for integers from "0" to "v-1" included in the G based on k elements included in each of the t-1 first subsets and the second subset. By assigning code values of "1" and "0" in the order of "from "v-1", v "1" and "" corresponding to each of the t-1 first subsets and the second subset. A first sequence consisting of a total of N code values of "1" and "0" by generating t partial sequences consisting of a code value of 0" and combining the t partial sequences-N=tv, and the first Since the sequence is a combination of t partial sequences consisting of v code values of “1” and “0”, the code values of “1” and “0” constituting the first sequence become N total 1 sequence generator;
When each cyclic shift operation is performed while performing a cyclic shift operation v-1 times in a predetermined direction for each of the t partial sequences consisting of the v code values of “1” and “0” A second sequence generator configured to generate a total of v-1 second sequences by combining the t partial sequences for each cyclic shift operation is completed to generate a second sequence;
A sequence matching unit for matching the first sequence and the v-1 second sequences one by one for each of the v message distribution groups; And
For each of the v message distribution groups, an allocation unit for distributing and allocating tk of the N messages based on N code values constituting each sequence matched to each of the v message distribution groups
Data distributed storage device using a divisional repetition code comprising a. The method of claim 2,
The first sequence generator
Integers corresponding to k elements included in each of the t-1 first subsets and the second subsets among integers from "0" to "v-1" included in G A code value of "1" is assigned to and, and the remaining integers that do not match k elements included in each of the t-1 first subsets and the second subset are "0". Partitioning to generate the t partial sequences consisting of the v code values of "1" and "0" corresponding to the t-1 first subsets and the second subset by assigning a code value Data distribution storage device using repetition code. The method of claim 3,
The message distribution allocation unit
An index allocator that allocates an index value associated with sequence number information for each of the N messages
Including more,
The allocation unit
For each of the v message distribution groups, an array order of N code values constituting each sequence matched to each of the v message distribution groups and an index value assigned to each of the N messages are compared The data distribution storage device using a divisional repetition code for distributing and allocating the N messages by tk based on a code value in an arrangement order that matches the index value allocated to the N messages. The method of claim 4,
The allocation unit
For each of the v message distribution groups, after checking a code value of "1" each of tk among N code values constituting each sequence matched to each of the v message distribution groups, the v number of Data distribution storage device using divisional repetition code for distributing and allocating tk messages to which an index value matching the order of the code value of "1" among the N messages is allocated to each of the message distribution groups . The method of claim 5,
After tk of the N messages are distributed and stored in the v data stores, if a load command for the N messages is applied, tk data of the v data stores based on the load command A storage message extraction unit which randomly selects storages and extracts tk storage messages stored in each of the tk data storages from each of the tk data storages; And
A data load unit that loads tk stored messages extracted from each of the tk data stores as data according to a load instruction for the N messages
Data distributed storage device using a divisional repetition code further comprising a. The method of claim 6,
The N messages are a Maximum Distance Separable (MDS) code that enables recovery of lost messages based on the remaining messages when some messages are lost,
The data load unit
When tk stored messages are extracted from each of the tk data stores, the index value allocated to the extracted stored messages is compared with the index value allocated to the N messages, and the extraction of the N messages If it is confirmed that there is a residual message not included in the stored messages, the residual message is restored based on the extracted stored messages, and the extracted stored messages and the restored residual message are converted to the N messages. A data distribution storage device using a divisional repeat code that loads data according to a load instruction for In the distribution group generator, N-N is a natural number that satisfies the condition of N=tv, v is a natural number that satisfies the condition of v=6t-k, and k and t are natural numbers of 2 or more-v for 2 messages Generating v message distribution groups corresponding to the v data stores when a command for distributed storage to the v data stores is applied;
The first subset selection unit, G, which is a set for an additive group including integers from "0" to "v-1" as elements-The number of elements included in G is v- Among a plurality of subsets having k elements for, G includes all remaining values calculated by dividing the difference between two different elements by v among k elements included in each of the plurality of subsets. Selecting t-1 first subsets that are found to match only one element other than "0" from among the elements in the set;
The second subset selection unit comprises two different subsets of k elements included in each of the remaining subsets among the remaining subsets from which the t-1 first subsets are excluded from the plurality of subsets. Selecting one second subset in which all remaining values calculated by dividing the difference between the elements by v are included as elements of the own subset;
A sequence for the message distribution allocation unit to allocate the N messages to each of the v message distribution groups based on k elements included in each of the t-1 first subsets and the second subsets And, based on the sequence, for each of the v message distribution groups, tk-tk means the product of t and k-distributed allocation by each-distributed allocation to the v message distribution groups -The messages that are used may be duplicated with each other; And
Separately storing, by a data storage unit, tk messages allocated to each of the v message distribution groups, for each of the v message distribution groups, in a data store corresponding to each message distribution group among the v data stores
Data distributed storage method using a divided repetition code including a. The method of claim 8,
Distributing and allocating tk of the N messages to each of the v message distribution groups
The first sequence generation unit includes a range of "0" to "v-1" included in the G based on k elements included in each of the t-1 first subsets and the second subset. By assigning code values of "1" and "0" in the order of "0" to "v-1" for integers, v corresponding to each of the t-1 first subsets and the second subsets A first sequence consisting of a total of N code values of "1" and "0" by generating t partial sequences consisting of "1" and "0" code values, and combining the t partial sequences-N= tv, and since the first sequence is a combination of t partial sequences consisting of v code values of “1” and “0”, a total of N code values of “1” and “0” constituting the first sequence To create-
The second sequence generator performs a cyclic shift operation v-1 times in a predetermined direction for each of the t partial sequences consisting of the v code values of "1" and "0", Generating a total of v-1 second sequences by combining the t partial sequences on which the cyclic shift operation is completed each time a shift operation is performed, thereby generating a second sequence;
A sequence matching unit matching the first sequence and the v-1 second sequences one by one for each of the v message distribution groups; And
Allocating, for each of the v message distribution groups, distributedly allocating tk of the N messages based on N code values constituting each sequence matched to each of the v message distribution groups.
Data distributed storage method using a divided repeat code including a. The method of claim 9,
Generating the first sequence comprises:
The first sequence generation unit includes k numbers of t-1 first subsets and k included in each of the second subsets among integers from "0" to "v-1" included in the G A code value of "1" is allocated to integers that match elements, and the remaining integers that do not match k elements included in each of the t-1 first subsets and the second subsets By assigning a code value of "0" to the t-1, the t consisting of the v code values of "1" and "0" corresponding to each of the t-1 first subsets and the second subsets Data distributed storage method using divisional repetition codes that generate four partial sequences. The method of claim 10,
Distributing and allocating tk of the N messages to each of the v message distribution groups
Allocating, by an index allocator, an index value associated with sequence number information for each of the N messages
Including more,
Distributed allocation of the N messages by tk based on the N code values,
The allocation unit, for each of the v message distribution groups, is assigned to each of the N messages and an arrangement order of N code values constituting each sequence matched to each of the v message distribution groups. A data distribution storage method using a divisional repetition code in which tk of the N messages are distributed and allocated based on a code value in an array order that matches the index value allocated to the N messages after comparing index values. The method of claim 11,
Distributed allocation of the N messages by tk based on the N code values,
After the allocating unit, for each of the v message distribution groups, checks a code value of "1" each of tk among N code values constituting each sequence matched to each of the v message distribution groups , For each of the v message distribution groups, using a divisional repetition code for distributing and allocating tk messages to which an index value matching the arrangement order of the code value of "1" among the N messages is allocated. Data distributed storage method. The method of claim 12,
After tk of the N messages are distributed and stored in the v data stores, when a load command for the N messages is applied, a storage message extractor may perform the v data stores based on the load command. Randomly selecting tk data stores among the tk data stores, and extracting tk stored messages stored in each of the tk data stores from each of the tk data stores; And
Loading, by a data loading unit, stored messages extracted from each of the tk data stores as data according to a load instruction for the N messages.
Data distributed storage method using a divisional repetition code further comprising. The method of claim 13,
The N messages are a Maximum Distance Separable (MDS) code that enables recovery of lost messages based on the remaining messages when some messages are lost,
The loading step
When the data load unit extracts tk stored messages from each of the tk data stores, compares the index value allocated to the extracted stored messages with the index value allocated to the N messages, If it is determined that there is a residual message not included in the extracted stored messages among messages, the residual message is recovered based on the extracted stored messages, and the extracted stored messages and the recovered residual message are Data distributed storage method using divisional repetition codes for loading data according to a load command for the N messages. A computer-readable recording medium having a computer program recorded thereon for executing the method of any one of claims 8 to 14 through combination with a computer. A computer program stored in a storage medium for executing the method of any one of claims 8 to 14 through combination with a computer.

Images