KR20140108618A - Apparatus and method for storing data using scalable distributed index - Google Patents

Abstract

A device and a method to distribute and store data in at least one storage by using an expandable distributed index are disclosed. The at least one storage constitutes a tree structure, and each storage has a hierarchical key. Each storage responds to its own key and stores data which does not correspond to a key of a sub-storage thereof.

Description

[0001] APPARATUS AND METHOD FOR STORING DATA USING SCALABLE DISTRIBUTED INDEX [0002]

The following embodiments are directed to an apparatus and method for storing data.

An apparatus and method for distributing and storing data in one or more repositories using an expandable distributed index is disclosed.

In order to increase the data storage capacity, the 'vertical expansion' method and the 'horizontal expansion' method may be used. Vertical expansion refers to the use of more specialized equipment to store data. Horizontal expansion is a method of enlarging the data storage capacity by adding a device for storing data. Vertical expansion can not handle data that exceeds the capacity that the device can handle. Thus, in general, many companies that need to process large amounts of data, such as Internet companies, use horizontal scaling to process large amounts of data.

The Relational DataBase Management System (RDBMS) can not be horizontally extended due to its nature. Therefore, when an RDBMS is used for a large amount of data, the entire capacity of the RDBMS can be expanded through sharding (or data partitioning). That is, one or more RDBMS devices may be used that store different data using the same schema.

When one or more RDBMS devices are used, an application server that knows which RDBMS devices are available may be used, and the actual data location may be hidden by a separate middleware.

When sharding (or data partitioning) is used, the problem may arise that data is redistributed between one or more RDBMS devices when an RDBMS device is added to process a larger amount of data. Redistribution of such data can be done during operation of RDBMS devices. However, instant redemption may be slow because the redistribution of data can not be performed quickly. To distribute and store data, a distributed key-value DB can be used. The distributed key-value DB can distribute and store data by using a consistent hashing scheme. Consistent hashing schemes can have structural advantages in data expansion. That is, a data storage system using a coherent hashing scheme can increase the total amount of data that can be processed by adding a server. However, since this hashing scheme does not support the concept of size, it is vulnerable to range retrieval and can not process data more than two dimensions.

Data in a system using a distributed key-value DB is distributed by hashing. Therefore, the search results can not be obtained sequentially for the range search, and the keys within the search range must be inquired individually.

For example, in the RDBMS, a query such as' select * from foo where A '= 1 and A < = 10' may be used to search for ranges where the value of the field A is between 1 and 10. On the other hand, the key-value DB needs to inquire data from 1 to 10 one by one.

Therefore, a key-value DB using hashing can not have a spatial index for processing data of two or more dimensions. That is, when data in a specific space is to be processed, since data in the space is distributed and stored in a plurality of nodes (i.e., servers) of the key-value DB, A complete index can not be prepared.

One embodiment of the present invention may provide an apparatus and method for storing data using one or more repositories configured in a tree structure.

One embodiment of the present invention may provide an apparatus and method for determining a repository in which data will be stored using a hierarchical key.

According to one aspect of the present invention there is provided a method for managing a tree, comprising: storing one or more stores in a tree structure, wherein each store corresponds to a node in the tree; each of the one or more stores has a hierarchical key Wherein the repositories in the subtree rooted at any one of the one or more repositories store data corresponding to a first key of the first repository and the first key is one Wherein the subkeys are a concatenated key, the second key is a key of a second repository, and the second repository is a parent repository of the first repository.

Each of the one or more repositories may be a relational database device.

Each of the one or more repositories may include middleware that understands and processes indexes, keys, and instructions of the relational database.

The hierarchical key may be a string combined with a numeric character and a delimiter.

The data corresponding to the first key may mean data one of the prefixes of the key of the data is the same as the first key.

The prefix may be the first i subkeys among the n subkeys of the key of the data.

i may be 1 or more and n or less.

If one of the key prefixes of the data is the same as the key of the repository, the data may correspond to the key of the repository.

The first repository may store data that does not correspond to the key of the child repository of the first repository among the data corresponding to the key of the first repository.

Wherein the third storage is capable of moving data corresponding to a third key of the third storage from the first storage to the third storage when the third storage is added to the data storage device, Lt; / RTI &gt;

When the amount of storage of the first storage reaches a predefined criterion, the data storage device may perform the addition and the data movement.

The first repository may request one or more child repositories of the first repository for a first list of data having a key corresponding to the retrieval scope, May be merged into the first lists returned according to the request and returned as a result for the search range.

According to another aspect of the present invention there is provided a method of configuring one or more repositories in a tree structure, each repository corresponding to one node in the tree, And storing data corresponding to a first key of the first repository in repositories in a subtree rooted at any first one of the one or more repositories, Wherein the key is a key in which one or more subkeys are concatenated to a second key, the second key is a key of a second repository, and the second repository is a parent repository of the first repository.

The storing may include storing, in the first storage, data that does not correspond to a key of the child storage of the first storage among data corresponding to the key of the first storage.

Wherein the data storage method further comprises: adding a third storage to the one or more stores as a child storage of the first storage and transferring data corresponding to a third key of the third storage from the first storage to the third storage To the mobile terminal.

The adding and moving may be performed when the amount of storage in the first reservoir reaches a predefined criterion.

Wherein the one or more child repositories of the first repository are requested by the first repository, the first repositories of the one or more child repositories of the first repository, And returning a second list of data corresponding to the search range out of the data stored in the first store to the returned first lists as a result for the search range have.

An apparatus and method for storing data using one or more repositories configured in a tree structure are provided.

An apparatus and method for determining a repository for storing data using a hierarchical key is provided.

An apparatus and method for extending a storage into a tree structure and moving data to an extended storage as data is stored are provided.

There is provided an apparatus and method for delivering a query to a child store corresponding to a child node, and for merging and returning a list of data returned from the child store as a search result for a query.

1 shows a data storage device 100 according to an embodiment of the present invention.
2 is a diagram illustrating a process of adding a repository to the apparatus 100 according to an exemplary embodiment of the present invention.
FIG. 3 illustrates a range search for device 100 in accordance with one example of the present invention.
4 is a flowchart of a data storage method according to an embodiment of the present invention.
5 is a flow diagram of a method of extending an apparatus 100 in accordance with an embodiment of the present invention.
Figure 6 is a flow diagram of a range search of an apparatus 100 according to an example of the present invention.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to or limited by the embodiments. Like reference symbols in the drawings denote like elements.

Figure 1 illustrates a data storage device 100 according to one embodiment of the present invention.

The data storage device 100 (hereinafter abbreviated as device 100) may include one or more depots 110, 120, 130, 140 and 150. Here, "storage" means a physical or logical space for storing data. For example, a "repository" can be a relational database or file system, or a distributed cluster that is a collection of replicas for the same data.

According to an embodiment of the present invention, each repository has a hierarchical structure logically in the form of a tree. In other words, each repository corresponds to one node in the tree, and between any two repositories, a parent-child relationship or sibling relationship may be established depending on the relationship between the nodes corresponding to the two repositories.

Referring to FIG. 1, the first storage 110 is connected to the root node of the tree structure, the second storage 120 and the third storage 130 are connected to the right child node and the left child node of the first storage 110, Respectively. The first storage 110 is the parent storage of the second storage 120 and the third storage 130. The second storage 120 is the (left) child storage of the first storage 110. The second storage 120 is a sibling store of the third storage 130. The fourth storage 140 and the fifth storage 150 correspond to the left child node and the right child node of the third storage 130, respectively.

Each of the repositories 110, 120, 130, 140 and 150 may be assigned a key for identifying each repository. The key is configured in a form capable of expressing a hierarchical structure between the stores. For example, one key may comprise one or more sub-keys and a separator for separating the sub-keys. Here, the subkey is a value representing all the nodes in an arbitrary subtree. Referring to FIG. 1, "korea" is a key representing a subtree rooted at the first storage 110, and "seoul " is a key representing a subtree rooted at the third storage 130. 1, "korea", "kyeonggi", "seoul", "kangbuk" and "kangnam" are each a sub key, and one key such as "korea.seoul.kangnam" is used as a delimiter Respectively.

Equation (1) is an example of a regular expression indicating such a hierarchical key.

Here, key represents a hierarchical key, and key is a string combined with alphanumeric and delimiter "."

In the case of using a normal alphabetic character, a number, and a delimiter ". &Quot;, the key can be expressed as Equation (1), but the present invention is not limited thereto and other characters may be used or other types of delimiters may be used. Also, the key does not necessarily have to be composed of a single string, but may be configured in the form of, for example, a linked list. Hereinafter, sub keys divided by delimiters are referred to as a first sub key, a second sub key, ..., and an nth sub key, respectively. Assuming that the level of the root node is 1, the key identifying the repository corresponding to the node located at level n will contain n subkeys.

On the other hand, the key of the repository corresponding to the root node may be null. In this case, the root node includes 0 subkeys, and the key identifying the repository corresponding to the node located at level n includes n-1 subkeys. For example, the hierarchical key may be a string generated by the regular expression of Equation (1) or an empty string.

In Figure 1, the key 115 assigned to the first storage 110 is "korea ". The key 115 has only the first sub key "korea ".

The key 125 assigned to the second storage 120 is "korea.gyeonggi ". The key 125 has a first sub key "korea" and a second sub key "gyeonggi ". The key 135 assigned to the third store 130 is "korea.seoul ". The key 135 has a first sub key "korea" and a second sub key "seoul ". The key 145 assigned to the fourth store 140 is "korea.seoul.kangbuk ". The key 135 has a first sub key "korea", a second sub key "seoul", and a third sub key "kangbuk". The key 155 assigned to the fifth repository 150 is "korea.seoul.kangnam ". The key 155 has a first sub key "korea", a second sub key "seoul", and a third sub key "kangnam".

The key assigned to the repository may be a key that is chained to one or more subkeys in the key of the repository's parent repository.

For example, the key assigned to the fourth storage 140 is a key in which the subkey "kangbuk" is cascaded to the key "korea.seoul" of the third storage 130, which is the parent storage of the fourth storage.

That is, the store p and store c parent store each other if the relationship between the child storage, if the storage p of the key k _p a first sub-key sk ₁₎ to (n includes sub key sk _n (where, n is at least 1 an integer), the storage key k _c a c a first sub-key sk ₁ to the n can include sk _n, and includes a first n + 1 sub-key sk _{(n + 1)} to the m-th sub-key sk _m have. Here, m is an integer of n + 1 or more.

According to one embodiment of the present invention, the device 100 may classify and store data according to the key of each repository. The data has a key. The key of the data is the identifier used in the classification scheme of the data. The key of the data may indicate the location on the classification system of the data. If the tree structure within the device 100 represents a classification scheme, the data may correspond to a particular repository (or a key of the repository) according to the key value of the data.

In a subtree rooted at a node representing a particular repository, repositories in the subtree store data corresponding to the key of the particular repository.

For example, in a subtree 160 that is the root of the third repository 130, the repositories 130, 140, and 150 in the subtree 160 are associated with the key 135 "korea.soul" As shown in FIG.

In addition, the subtree rooted at the fourth storage 140 includes only the fourth storage 140. Thus, the fourth storage 140 may store data corresponding to the key 145 "korea.seoul.kangbuk" of the fourth storage 140. [

In addition, the subtree rooted at the fifth storage 150 includes only the fifth storage 150. Thus, the fifth storage 150 may store data corresponding to the key 155 "korea.seoul.kangnam" of the fifth storage 150.

Here, the data corresponding to the key of the repository may mean data in which one of the prefixes of the key of the data is the same as the key of the above-mentioned repository.

The prefix of key x means that when the hierarchical key x contains n subkeys x ₁ , x ₂ , x ₃ , ..., x _n , the _first i of the subkeys of x (Hereinafter referred to as " key &quot;). For example, the prefixes of the hierarchical key "abc" may be "a &quot;," ab &quot;, and "abc &quot;.

For example, if the key of the data is "korea.seoul.kangnam" then "korea.seoul" is one of the prefixes of the above keys. The data thus corresponds to the key 115 of the first store 110 and the key 135 of the third store 130 and the key 145 of the fourth store 140, And the key 155 of the fifth storage 150.

For example, if the key of the data is "korea.seoul.kangnam.shinsa ", the prefixes of the keys include" korea ", " korea.seoul ", and " korea.seoul.kangnam ". The above data corresponds to the key 115 of the first storage 110, the key 135 of the third storage 130 and the key 155 of the fifth storage 150.

The data may be stored in a repository of one of the one or more repositories 110, 120, 130, 140 and 150 of the device 100 according to its key.

For example, the repository may store data that 1) corresponds to the key of the repository, and 2) does not correspond to the key of the repository's child repository.

The data corresponding to the repository p may correspond to the child repository c of the repository p. If the data corresponds to repository p and repository c, then the data may be stored in one of the repositories in the subtree rooted at repository c, rather than repository p.

If the data is "korea.seoul.kangnam.shinsa" or" korea.seoul.kangnam.shinsa.1&quot ;, the data is stored in the first storage 110, the third storage 130, . The data may be stored in the fifth storage 150.

If the key of the data is "korea.gangwon ", the data corresponds to the first storage 110. The data may be stored in the first storage 110.

When a storage method such as the one described above is used, the repository in which the actual data is located can be retrieved by the hierarchical key. That is, the repository where the actual data is located can be retrieved by the entire hierarchical key or the prefix of the key.

Hereinafter, an example in which data with a key of data "korea.seoul.kangnam.shinsa" is searched will be described. The key 115 of the first storage 110 corresponding to the root node is compared with the key of the data. One of the prefixes of the keys of the data, "korea" is the same as the key 115 of the first store 110. Also, the third storage 130 and the first storage 110 are in a child-parent relationship with each other. &Quot; korea.seoul ", which is one of the prefixes of the keys of the data, is the same as the key 135 of the third storage 130. The fifth storage 150 and the first storage 110 are in a child-parent relationship with each other. One of the prefixes of the keys of the data, "korea.seoul.kangnam" is the same as the key 155 of the fifth storage 150. Thus, the data whose key is "korea.seoul.kangnam.shinsa" can be retrieved in the fifth storage 150 via the first storage 110 and the third storage 130. [

Hereinafter, an example in which data having a key of "korea.jejudo" is searched will be described. One of the prefixes of the keys of the data, "korea" is the same as the key 115 of the first store 110. The prefixes of the keys of the data, "korea" and "korea.jejudo", are not the same as the keys 125 of the second storage 120 and the keys 135 of the third storage 130. Thus, data with the key "korea.jejudo" can be retrieved in the first storage 110.

2 is a diagram illustrating a process of adding a repository to the apparatus 100 according to an exemplary embodiment of the present invention.

The device 100 described above with reference to Figure 1 may be considered to have added storage to store data corresponding to the key "korea" of the store.

In the initial state 210, the device 100 has only the first storage 110.

The key 115 of the first store 110 is "korea ". Thus, data whose key is "korea ", data whose key starts with " korea.gyeonggi ", and data whose key starts with" korea.seoul "

As the device 100 is operated, as more data (e.g., "korea.seoul") with a particular prefix (i.e., starting with a particular string) You can add a repository that you have. This addition means that the tree structure of the apparatus 100 is expanded. That is, the above expansion means that a new storage (or a node representing a new storage) is added to the tree structure of the device 100.

In the state 220 where the data is stored, the first storage 110 stores one or more data 260 corresponding to the key 115.

If the first storage 110 can not process all of the one or more data 260, the addition of new storage may be required. For example, the device 100 may be extended via the following states 230, 240, and 250.

The device 100 may create a third storage 130 that is a child storage of the first storage 110, as shown in the node creation state 230. [

That is, the data corresponding to the third key 135 (i.e., data whose data key starts with "korea.seoul ") can be separated separately from the third storage 130, 130).

In the additional notification state 240, the newly created third repository 130 may notify the first repository 110, which is its parent repository, of the key to be processed (i.e., the third key 135) . The above notification informs the addition of the repository.

In the data movement state 250, the first storage 110 receiving the notification moves the data 270 corresponding to the notified third key 135 among the data stored therein to the third storage 130 . For example, the first storage 110 receiving the notification can copy the data 270 corresponding to the notified third key 135 among the data stored in the first storage 110 to the third storage 130. When the above copying is completed, the first storage 110 does not need to duplicate the data held by the child storage (i.e., the third storage 130). Accordingly, the first storage 110 may delete the data copied to the third storage 130 among the data stored therein. After the deletion, the first storage 110 stores data not corresponding to the third key 135 among the data corresponding to the first key 115.

The amount of storage in the first storage 110 may be reduced by moving data 270 corresponding to the third key 135 of the third storage from the first storage 110 to the third storage 130. [

The device 100 may be configured to generate the third storage 130 and the third storage 130 when the amount of storage in the first storage 110 reaches a predefined criterion by inserting new data into the first storage 110 130 in accordance with the present invention.

Of the above extensions, the device 100 may not be interrupted (partially or entirely). In addition, data can also be automatically replicated by the above extension.

Of the above extensions, new incoming data corresponding to the third key 135 is delivered to the third storage 130 corresponding to the child node generated by the extension.

The reduction of the device 100 (i.e., the nodes of the tree being deleted) may proceed in the reverse order of the expansion of the device 100 described above.

FIG. 3 illustrates a range search for device 100 in accordance with one example of the present invention.

A query can be used to query data that meets a specific condition. A query may be a statement that queries a list of data having a key corresponding to a particular search scope.

All or a portion of the repositories 110, 120, 130, 140 and 150 of the device 100 may retrieve data for the query. By merging the results retrieved by all or a portion of the repositories 110, 120, 130, 140 and 150, the unsaved data can also be retrieved from the particular repository 110, 120, 130, 140 or 150 have.

In the query provisioning state 310, a query may be provided to any of the repositories 110, 120, 130, 140 or 150.

In this example, a case where a query is provided to the first storage 110 corresponding to the root node will be described.

The first storage 110 can determine whether or not its first key 115 corresponds to the query statement by analyzing the transmitted query statement. Here, the correspondence of the first key 115 to the query may mean that there may be data that can be included in the search range of the query query among the data corresponding to the first key.

For example, if the inquiry inquiry range is from "usa.ar" to "usa.ca", the data corresponding to the first key 115 "korea" can not be included in the above search range. Therefore, the first key 115 does not correspond to the query statement.

If there is no data corresponding to the query, the first store 110 may return null for the query statement or may not perform any further processing.

In the query delivery state 320, the first store 110 may pass a query to the second store 120 and the third store 130, which are its child repositories. That is, the first repository 110 may request one or more child repositories (i.e., the second repository 120 and the third repository 130) with a list of data having a key corresponding to the search scope.

The second store 120 and the third store 130 may also pass query messages to their child repositories. That is, the transfer of the query can be performed hierarchically. (Not shown)

In the list return state 330, lists of data having keys corresponding to the search ranges from one or more child stores (e.g., second store 120 and third store 130) of the first store 110 Can be returned.

In addition, when the query transmission is performed hierarchically, the second storage 120 and the third storage 130 also receive lists of data having a key corresponding to the search range from their respective one or more child repositories .

In the list merge and return search results state 340, the first store 110 may return a list of merged data as a result of the search term's search scope.

The list of data returned by the second storage 120 and the third storage 130 may also be a list of merged data. That is, the return of the list of merged data can be performed hierarchically.

The first storage 110 may merge a second list of data stored in the first storage 110 corresponding to the search range into the first lists returned and the merged list generated by the merge As a result for the search range of the search term.

As described above, according to an example of the present invention, range search and spatial index can be supported without using a key-value DB or hashing.

4 is a flowchart of a data storage method according to an embodiment of the present invention.

At step 410, one or more repositories are organized into a tree structure.

Each of the one or more repositories may be an RDBMS.

At step 420, a hierarchical key is assigned to each of the one or more repositories. A hierarchical key may have zero or more subkeys.

The hierarchical key may be a string generated by the regular expression of Equation (1) or an empty string.

At step 430, data corresponding to the first key of the first repository is stored in repositories in the server tree rooted at the first node representing any of the one or more repositories.

The first key is a key in which one or more subkeys are cascaded to a second key of a second repository representing the parent node of the first node.

The data corresponding to the first key means that one of the prefixes of the key of the data is the same as the first key.

If one of the prefixes of the key of the data is the same as the key of the repository, the data can be viewed as corresponding to the repository.

The data corresponding to the third storage representing the child nodes of the first node among the data corresponding to the first storage is stored in the third storage and the data not corresponding to the third storage can be stored in the first storage. Thus, step 430 may include storing, in the first repository, data that does not correspond to the repository representing the child node of the first node, among the data corresponding to the first repository.

The technical contents according to one embodiment of the present invention described above with reference to Figs. 1 to 3 can be applied to this embodiment as it is. Therefore, a more detailed description will be omitted below.

5 is a flow diagram of a method of extending an apparatus 100 in accordance with an embodiment of the present invention.

Steps 510, 515, 520, 530, and 540, described below, may be included in step 430.

In step 510, it is checked whether the amount of storage in the first reservoir has reached a predefined criterion.

If the amount of storage in the first reservoir has reached a predefined criterion, then step 520 is performed, otherwise the procedure ends.

At step 520, a third repository is created which is a child repository of the first repository.

In step 530, the third repository notifies the first repository of its third key.

At step 540, the data corresponding to the third key of the third repository moves from the first repository to the third repository.

Step 540 includes the steps of 1) copying the data corresponding to the third key of the third repository from the first repository to the third repository, and 2) copying the data of the first repository to the third repository Step &lt; / RTI &gt;

The technical contents according to one embodiment of the present invention described above with reference to Figs. 1 to 4 can be directly applied to this embodiment as well. Therefore, a more detailed description will be omitted below.

Figure 6 is a flow diagram of a range search of an apparatus 100 according to an example of the present invention.

At step 610, a query is provided to the first repository.

The query is a statement that requests the repository to have data with a key corresponding to the scope of the query in the query.

At step 620, the query statement is passed to the child repositories of the first repository. That is, the first repository requests one or more child repositories for a first list of data having a key corresponding to the search scope.

Step 620 may be performed recursively. The child repository of the first repository that has received the query may redirect the query to its one or more child repositories.

In step 630, one or more repositories return the first listings.

In step 640, a merged list is generated by merging a second list of data having a key corresponding to the search range out of the stored data of the first repository into the first lists returned.

Steps 630 and 640 may be performed recursively. The child repository of the first repository may be returned a third list of data corresponding to the retrieval scope from its one or more child repositories. The child repository of the first repository may merge the third list with the first list and return the merged first list to the first repository.

At step 650, the merged list is returned as a result of the search range.

The technical contents according to one embodiment of the present invention described above with reference to Figs. 1 to 5 may be applied to this embodiment as it is. Therefore, a more detailed description will be omitted below.

The method according to an embodiment of the present invention can be implemented in the form of a program command which 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 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.

While the invention has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. This is possible.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the equivalents of the claims, as well as the claims.

100: Data storage device
110: First storage
115: First key

Claims (5)

Comprising one or more repositories organized in a tree structure, each repository corresponding to one node in the tree,
Each of the one or more repositories is assigned a hierarchical key with zero or more subkeys,
Wherein the repositories in the subtree rooted at any of the one or more repositories store data corresponding to the first key of the first repository,
Wherein the first key is a key in which one or more subkeys are cascaded to a second key, the second key is a key of a second repository, and the second repository is a parent repository of the first repository. The method according to claim 1,
Wherein each of the one or more repositories is a relational database device. 3. The method of claim 2,
Wherein each of the one or more repositories includes middleware that processes indexes, keys, or instructions of the relational database. Configuring one or more repositories in a tree structure, each repository corresponding to one node in the tree;
Assigning a hierarchical key with zero or more subkeys to each of the one or more repositories; And
Storing data corresponding to a first key of the first repository in repositories in a subtree rooted at any one of the one or more repositories
Lt; / RTI &gt;
Wherein the first key is a key in which one or more sub keys are concatenated to a second key, the second key is a key of a second repository, and the second repository is a parent repository of the first repository. A computer-readable recording medium storing a program for performing the data storing method of claim 4.

Images