JP2014099037A - Database management system and database management method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a DB management system and a DB management method which can suppress update processing of unnecessary data and concentration of update processing loads on one processing server.SOLUTION: A DB management system includes: a plurality of second DBs for storing corresponding data that is either duplication of object data or processed data of object data with respect to the object data inside a plurality of pieces of data in a first DB; storage means for storing the second DB identifying information for identifying any of the plurality of second DBs and object data identifying information for identifying the object data necessary for the identified second DB; control means for creating update processing data including updated object data when one of the object data identified by the object data identifying information is updated in update of the plurality of pieces of data in the first DB; and a processing server for performing update processing that reflects update of the object data in the first DB in corresponding data of the second DB identified by the second DB identifying information by using update processing data.

Description

  The present invention relates to a database management system and a database management method.

  A replication technology that creates a replica of a master DB (database) (hereinafter referred to as a “replica”), reflects the data update in the master DB to the replica, and matches the data in the master DB with the data in the replica. It is known (see Patent Document 1).

  In the replication technique, in order to match the data in the master DB with the data in the replica, all the data updates in the master DB are reflected in the replica.

International Publication No. 2004/090726

  There is a movement to combine databases managed separately at a plurality of locations into one database (hereinafter referred to as “integrated database”) from the viewpoint of information sharing.

  For example, it is assumed that a database that is individually owned by each department in a certain company is integrated into an integrated database.

  In this case, it is conceivable that a replica of the integrated database is provided in each department, and replication is performed between the integrated database and the replica.

  However, even if an integrated database is created, the data required by each department is often data of some items of the integrated database, not data of all items of the integrated database.

  For this reason, when a replica of the integrated database is created in each department and replication is performed between the integrated database and the replica, there is a problem that data update processing is performed on unnecessary items in the replica.

  The objective of this invention is providing the database management system and database management method which can solve the said subject.

The database management system of the present invention includes:
A first database for storing a plurality of data;
For a plurality of data in the first database, a plurality of second data for storing corresponding data that is either a copy of the target data or processed data obtained by performing a predetermined process on the target data. Two databases,
Second database specifying information for specifying any one of the plurality of second databases, and target data specifying information for specifying the target data necessary for the second database specified by the second database specifying information Storage means for storing
In the first database, when at least one of the plurality of data is updated, it is determined whether one of the target data specified by the target data specifying information in the storage unit is updated, Control means for creating update processing data including the updated target data when the target data is updated;
A processing server that uses the update processing data to execute an update process that reflects the update of the target data in the first database in the corresponding data in the second database specified by the second database specifying information And including.

The database management method of the present invention includes:
Any one of a first database that stores a plurality of data, a target data among a plurality of data in the first database, a copy of the target data, and processed data obtained by performing a predetermined process on the target data A plurality of second databases for storing corresponding data, second database specifying information for specifying any one of the plurality of second databases, and a second database specified by the second database specifying information A database management method performed by a database management system including: storage means for storing target data specifying information for specifying the target data necessary for control; a control means; and a processing server,
When the control unit updates at least one of the plurality of data in the first database, one of the target data specified by the target data specifying information in the storage unit is updated. And when the target data is updated, create update processing data including the updated target data,
An update process in which the processing server reflects the update of the target data in the first database to the corresponding data in the second database specified by the second database specifying information using the update processing data. ,Run.

  According to the present invention, it is possible to suppress unnecessary data update processing.

It is a figure showing database management system 100 of one embodiment of the present invention. It is the figure which showed an example of the data in replication origin DB101. It is the figure which showed an example of the data in replication destination DB21. It is the figure which showed an example of the data in replication destination DB22. It is the figure which showed an example of the management information 103a. It is the figure which showed an example of the update data information 104a. It is a figure for demonstrating the condition where each processing server started the process. It is a figure showing the process of the update data information by a processing server. It is the figure which showed an example of the operation rate (operation degree) of the process servers 31-33, and the update data information in the update data storage part 104. FIG. It is a figure showing the substitution process of the update data information by a processing server.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing a database management system 100 according to an embodiment of the present invention.

  In FIG. 1, the database management system 100 includes a management system 1, replication destination DBs 21 and 22, and processing servers 31 to 3i. Note that i is an integer of 1 or more.

  The management system 1 includes a replication source DB 101, an update unit 102, a storage unit 103, an update data storage unit 104, a server load confirmation unit 105, and a control unit 106.

  The replication source DB 101 is an example of a first database.

  The replication source DB 101 stores data of a plurality of items, for example, data of m items (m is an integer of 3 or more) items.

  FIG. 2 is a diagram illustrating an example of data in the replication source DB 101.

  In FIG. 2, the replication source DB 101 stores tables TA to TD.

  The table TA includes, as items, ID information (user ID) A1 and a password A2. In table TA, data A1D of ID information A1 and data A2D of password A2 are stored in association with each other for each data A1D of ID information A1. The ID information A1 is information for specifying the data A1D. The password A2 is information for specifying the data A2D.

  The table TB has a product name B1 and a price B2 as items. In the table TB, for each data B1D of the product name B1, data B1D of the product name B1 and data B2D of the price B2 are stored in association with each other. The product name B1 is information for specifying the data B1D. The price B2 is information for specifying the data B2D.

  The table TC has address information C1 and a telephone number C2 as items. In the table TC, the data C1D of the address information C1 and the data C2D of the telephone number C2 are stored in association with each other for each data C1D of the address information C1. The address information C1 is information for specifying the data C1D. The telephone number C2 is information for specifying the data C2D.

  The table TD includes a store name D1 and a stock quantity D2 as items. In the table TD, for each data D1D of the store name D1, the data D1D of the store name D1 and the data D2D of the stock quantity D2 are stored in association with each other. The store name D1 is information for specifying the data D1D. The inventory quantity D2 is information for specifying the data D2D.

  Each of the data C1D of the address information C1 is classified into either data C1D1 of user address information or data C1D2 of store address information.

  The user address information data C1D1 is also associated with the data A1D and represents the address of the user specified by the data A1D.

  The store address information data C1D2 is also associated with the data D1D and represents the store address specified by the data D1D.

  The data D2D of the stock quantity D2 is also associated with the data B1D.

  The data A1D, data A2D, data B1D, data B2D, data C1D, data C2D, data D1D, and data D2D are examples of a plurality of data in the copy source DB 101.

  The table and data stored in the replication source DB 101 are not limited to the table and data shown in FIG.

  Returning to FIG. 1, the replication destination DBs 21 and 22 are examples of the second database, respectively.

  The replication destination DB 21 sets the target data for each of n (n is an integer less than or equal to 2 and less than m) items of data (hereinafter referred to as “target data”) of the m items of data in the replication source DB 101. Corresponding data that is either a copy (copy) of the data or processed data obtained by performing predetermined processing (predetermined processing) on the target data is stored.

  The replication destination DB 22 stores corresponding data for each of target data of p items (p is an integer of 2 or more and less than m) of m items of data in the replication source DB 101.

  In the following, the replication destination DBs 21 and 22 store the replication of the target data as corresponding data.

  FIG. 3 is a diagram illustrating an example of data in the replication destination DB 21.

  The replication destination DB 21 stores a table TE.

  The table TE includes, as items, ID information E1 that is a copy of the ID information A1 in the table TA and address information E2 that is a copy of the address information C1 in the table TC. In table TE, for each data E1D of ID information E1, data E1D of ID information E1 and data E2D of address information E2 are stored in association with each other.

  The data E1D is a copy (corresponding data) of the data A1D in the table TA. Data E2D is a copy (corresponding data) of data C1D1 of user address information in table TC.

  The data E1D is associated with the data E2D of the address information E2 that represents the address of the user specified by the data E1D.

  Regarding the replication destination DB 21, among the data in the replication source DB 101, data A1D and user address information data C1D1 are the target data.

  In the replication source DB 101, the user address information data C1D1 as target data is associated with data A1D as related target data. In the replication source DB 101, data A1D that is target data is associated with data C1D1 of user address information that is related target data.

  FIG. 4 is a diagram illustrating an example of data in the replication destination DB 22.

  The replication destination DB 22 stores tables TF and TG.

  The data structure in the table TF is the same as the data structure in the table TE.

  The table TF includes as items, a product name G1 that is a duplicate of the product name B1 in the table TB, and a stock quantity G2 that is a duplicate of the stock quantity D2 in the table TD. In the table TG, for each data G1D of the product name G1, data G1D of the product name G1 and data G2D of the stock quantity G2 are stored in association with each other.

  The data G1D is a copy (corresponding data) of the data B1D in the table TB. Data G2D is a copy (corresponding data) of data D2D in table TD.

  Data G1D is associated with data G2D of inventory quantity G2 representing the inventory quantity of the product name specified by data G1D.

  Regarding the replication destination DB 22, among the data in the replication source DB 101, data A1D, user address information data C1D1, data B1D, and data D2D are the target data.

  In the replication source DB 101, the user address information data C1D1 as target data is associated with data A1D as related target data. In the replication source DB 101, data A1D that is target data is associated with data C1D1 of user address information that is related target data.

  In the replication source DB 101, data D2D that is target data is associated with data B1D that is related target data. In the replication source DB 101, the data B1D that is the target data is associated with the data D2D that is the related target data.

  Note that the number of replication destination DBs is not limited to two but may be plural.

  Returning to FIG. 1, the update unit 102 updates (for example, registers, changes, or deletes) data in the replication source DB 101 in accordance with an update request from the user terminal 4.

  The storage unit 103 is an example of a storage unit.

  The storage unit 103 stores management information 103a and processing server load information 103b.

  The management information 103a indicates the processing contents for reflecting the data update in the replication source DB 101 to the replication destination DB 21 or 22.

  FIG. 5 is a diagram showing an example of the management information 103a.

  In FIG. 5, the management information 103a includes management information No (number) 103a1, logical data name 103a2, copy destination information 103a3, logical data 103a4, distributed value 103a5, and processing server information 103a6. In the management information 103a, the management information No 103a1, the logical data name 103a2, the copy destination information 103a3, the logical data 103a4, the distributed value 103a5, and the processing server information 103a6 are associated with each other.

  Management information No 103a1 indicates the number of the management information 103a.

  The logical data name 103a2 represents the name of the logical data 103a4 handled by the management information 103a.

  The logical data 103a4 is an example of target data specifying information.

  The logical data 103a4 is data that defines a combination of data (target data) required by the replication destination DB 21 or 22 among data in a plurality of tables of the replication source DB 101 using data items.

  The data (target data) required by the replication destination DB 21 or 22 is also data stored in the replication destination DB 21 or 22.

  In the present embodiment, the following logical data A and B are used.

Logical data A: ID information A1 of table TA + address information C1 of table TC
Logical data B: product name B1 of table TB + stock quantity D2 of table TD
The logical data A specifies the data A1D of the ID information A1 (item) of the table TA and the data C1D of the address information C1 (item) of the table TC associated with the data A1D.

  The logical data B specifies the data B1D of the product name B1 (item) of the table TB and the data D2D of the inventory quantity D2 (item) of the table TD associated with the data B1D.

  The logical data is not limited to the logical data A and B, and can be changed as appropriate. For example, the following logical data C may be used.

Logical data C: address information C1 of table TC + stock information D2 of table TD
The logical data C specifies the data C1D of the address information C1 (item) of the table TC and the data D2D of the inventory information D2 (item) of the table TD associated with the data C1D.

  Further, when the update in the replication source DB 101 is reflected in the replication destination DB 21 or 22, the data updated in the replication source DB 101 is processed (for example, the address is processed only in the prefecture) or converted (for example, weight data). If there is a need to perform a predetermined process for converting the weight value from kg to g), the processing information indicating the processing content of the processing or change (hereinafter collectively referred to as processing) is performed on the target data. Added to logical data.

  In FIG. 5, the data required by the replication destination DB 21 is defined by the logical data A, and the data required by the replication destination DB 22 is defined by the logical data A and B.

  In the above, the combination of the target data specified by the logical data 103a4 is a combination of two target data. However, the combination of the target data specified by the logical data 103a4 is a combination of three or more target data. Also good.

  The replication destination information 103a3 represents a replication destination DB.

  The distributed value 103a5 defines the number of processing servers that perform processing for reflecting the update of the data specified by the logical data 103a4 in one replication destination DB.

  The processing server information 103a6 designates a processing server that performs an update process for reflecting an update of data specified by the logical data 103a4 to the replication destination DB represented by the replication destination information 103a3.

  The combination of the distributed value 103a5 and the processing server information 103a6 is an example of server designation information.

  Returning to FIG. 1, the processing server load information 103 b in the storage unit 103 represents the availability of each of the processing servers 31 to 3 i.

  The update data storage unit 104 is an example of a storage unit.

  The server load confirmation unit 105 is an example of a confirmation unit.

  The server load confirmation unit 105 receives the operating levels of the processing servers 31 to 3i from the processing servers 31 to 3i. The server load confirmation unit 105 creates the processing server load information 103b based on the operating levels of the processing servers 31 to 3i, and stores the processing server load information 103b in the storage unit 103.

  The control unit 106 is an example of a control unit and a setting unit.

  For example, the control unit 106 creates management information 103 a based on an input from the administrator terminal 5, and stores the management information 103 a in the storage unit 103.

  In addition, when the data in the replication source DB 101 is updated (for example, registered, changed, or deleted), the control unit 106 refers to the management information 103a to determine whether the updated data is target data. Check. When the updated data is the target data, the control unit 106 creates update data information including the data update result, and stores the update data information in the update data storage unit 104.

  FIG. 6 is a diagram showing an example of the update data information 104a.

  In FIG. 6, the update data information 104a includes, as items, a process number (number) 104a1, a logical data name 104a2, a copy destination information 104a3, a copy data 104a4, a distributed value 104a5, and a process flag (F) 104a6. And processing server information 104a7.

  Process No. 104a1 indicates the number of the update data information 104a.

  The logical data name 104a2 represents the name of the logical data 103a4 (hereinafter referred to as “corresponding logical data”) including the updated target data item among the logical data 103a4 in the storage unit 103.

  The copy destination information 104a3 represents the copy destination information 103a3 associated with the corresponding logical data in the management information 103a.

  The duplicated data 104a4 represents a combination including the updated target data among the combinations of target data specified by the corresponding logical data. The duplicate data 104a4 is an example of update processing data.

  The variance value 104a5 represents the variance value 103a5 associated with the corresponding logical data in the management information 103a.

  The processing flag 104a6 is used to represent the processing status of the update data information 104a. At the stage when the update data information 104a is created, “unprocessed” is defined in the processing flag 104a6.

  The processing server information 104a7 represents the processing server information 103a6 associated with the corresponding logical data in the management information 103a.

  Returning to FIG. 1, the control unit 106 confirms the processing server load information 103 b in the storage unit 103, and sets the processing server information 103 a 6 in the management information 103 a based on the availability of the processing servers 31 to 3 i. Or change it.

  The processing servers 31 to 3i refer to the logical data 103a4 in the storage unit 103, and when one of the target data specified by the logical data 103a4 in the storage unit 103 is updated in the replication source DB 101, the replication is performed. Of the target data in the original DB 101, the update data information 104a including the updated target data (particularly, the replicated data 104a4) is used to reflect the update of the target data in the replica DB 101 to the replica DB 21 or 22. Update process is executed. When two or more processing servers 31 to 3i are used according to the distributed value 103a5, the update processing is executed by distributed processing.

  Each of the processing servers 31 to 3i includes a storage unit 301, a processing unit 302, and a load status confirmation unit 303.

  The processing unit 302 includes, for example, a CPU (Central Processing Unit) and a memory.

  The processing unit 302 acquires the management information 103a from the storage unit 103 in the management system 1, and stores the management information 103a in the storage unit 301 in the own processing server.

  The processing unit 302 refers to the management information 103a in the storage unit 301 and acquires information (update data information 104a) to be processed from the update data storage unit 104 in the management system 1.

  The processing unit 302 reflects the update of the target data in the replication source DB 101 on the replication destination DB 21 or 22 using the update data information 104a acquired from the update data storage unit 103.

  In addition, when the processing information showing the predetermined processing content about the replication data is added to the management information 103a, the processing unit 302 performs predetermined processing on the replication data based on the processing information, Data in the copy destination DB 21 or 22 is updated using the processed copy data.

  The load status confirmation unit 303 detects the operation level of the own processing server (for example, the usage rate of the CPU and memory included in the processing unit 302 in the own processing server), and determines the operation level of the own processing server as the server load confirmation unit 104. Send to.

  Next, the operation will be described.

(1) Setting of Management Information 103a Regarding the setting of the management information 103a, for example, the management information 103a shown in FIG. 5 is input from the administrator terminal 5, and the control unit 106 stores the input management information 103a in the storage unit 103. To remember.

  Further, when the control unit 106 receives the logical data name 103a2, the logical data 103a4, the copy destination information 103a3, and the distribution number 103a5 for each logical data from the administrator terminal 5, the logical data name 103a2 and the logical data 103a4 are copied. Based on the destination information 103a3 and the distributed value 103a5, the management information 103a having the processing server information 103a6 column blank may be created first, and then the processing servers 31 to 3i may be assigned to the processing server information 103a6 column. .

  For example, logical data name 103a2 representing logical data A, logical data 103a4 representing logical data A, duplication destination information 103a3 representing duplication destination DB 21 as a duplication destination of target data specified by logical data A, logical data The combination of the copy destination information 103a3 representing the copy destination DB 22 as the copy destination of the target data specified by A and the target data specified by the logical data A based on the data of the ID information A1 is divided into two. When the distributed value 103a5 (for example, a value in which ID001 to ID500 are set to the number 1/2 and ID501 to ID999 are set to the number 2/2) is input, the control unit 106 includes the following four pieces of management information 103a (see FIG. 5).

  Management information No. 1 management information 103a, the logical data 103a4 represents the logical data A, the copy destination information 103a3 represents the copy destination DB 21, the distributed value 103a5 represents a value having ID001 to ID500 as a number 1/2, and processing server information 103a6 represents the processing server 31.

  Management information No. 2 management information 103a, logical data 103a4 represents logical data A, duplication destination information 103a3 represents duplication destination DB22, distributed value 103a5 represents a value having ID001 to ID500 as a number 1/2, and processing server information 103a6 represents the processing server 32.

  Management information No. 3 management information 103a, the logical data 103a4 represents the logical data A, the replication destination information 103a3 represents the replication destination DB 21, the distributed value 103a5 represents a value with IDs 501 to 999 number 2/2, and processing server information 103a6 represents the processing server 33.

  Management information No. 4 management information 103a, logical data 103a4 represents logical data A, replication destination information 103a3 represents replication destination DB22, distributed value 103a5 represents a value with ID501 to ID999 as number 2/2, and processing server information 103a6 represents the processing server 34.

  The control unit 106 sets the processing server information 103a6 in the management information 103a with a blank, and when any of the processing servers 31 to 3i acquires the management information 103a, the processing server information indicating the processing server 103a6 may be set.

  Further, the control unit 106 may set the processing server information 103a6 in the management information 103a with a blank, and may set the processing server information 103a6 representing the processing server in the order of the low availability of the processing servers 31 to 3i. Good.

  The control unit 106 stores the created management information 103 a in the storage unit 103.

(2) Pre-processing by the processing server In the processing servers 31 to 3 i, the processing unit 302 reads all management information 103 a in the storage unit 103 of the management system 1 and stores all management information 103 a in the storage unit 301. To do.

  Subsequently, the processing unit 302 refers to the management information 103a in the storage unit 301 and starts up a process for executing a series of update processing. The series of update processes is a process of acquiring update data information to be processed by itself and updating the data in the replication source DB 21 or 22 using the update data information.

  For example, when the management information 103a illustrated in FIG. 5 is stored in the storage unit 103, the processing servers 31 to 34 each start up a process as follows and take charge of the update process.

  The processing server 31 has a management information No. 1 and the management information No. 1 are processed. And a process for processing 5.

  The processing server 32 manages the management information No. 2 and the management information No. 2 are processed. Each process for processing 6 is started.

  The processing server 33 has management information No. 3 and the management information No. 3 are processed. And a process for processing 7.

  The processing server 34 has the management information No. A process for processing 4 is started.

  7 shows that the processing server 31 has the management information No. 1 and No. Each process for processing the server 5 is started up, and the processing server 32 manages the management information No. 5. 2 and No. Each process for processing No. 6 is started, and the processing server 33 manages the management information no. 3 and no. Each process for processing the management information No. 7 is started up. 4 is a diagram for explaining a situation where a process for processing 4 is started up. FIG.

(3) Update Process When the update unit 102 receives an update request for registration, change, or deletion of data from the user terminal 4 to the replication source DB 101, the update unit 102 updates the replication source DB 101 according to the update request.

  When data update occurs in the replication source DB 101 or periodically, the control unit 106 checks which table data has been updated, refers to the management information 103a in the storage unit 103, and updates the updated data. Is the data of the table defined by the logical data 103a4.

  When the updated data is the data (target data) of the table defined in the logical data 103a4, the control unit 106 updates the updated data and the logical data 103a4 from the replication source DB 101. A combination of the data and the data paired with the data (hereinafter also referred to as “related object data”) is extracted as the duplicate data 104a4.

  Subsequently, the control unit 106 refers to the management information 103a including the logical data 103a4 used when extracting the duplicate data 104a4, creates the update data information 104a using the duplicate data 104a4, and the update data information 104a. Is additionally stored in the update data storage unit 104 (see FIG. 6).

  For example, in a situation where the distributed value of the logical data A is a value representing ID information: ID001 to ID500 representing the number 1/2 and ID information: ID500 to ID900 representing the number 2/2, the user of the ID information “ID001” When the address information data is updated to “Tokyo □□ ...” and the user of ID information “ID567” updates the address information data to “Chiba □□ ...”, these Since the update corresponds to the update of the address information data C1D of the table TC defined in the logical data A, the control unit 106 determines that the updated data is the target data, and adds the address information to the address information. The following four pieces of update data information are created together with ID information that is corresponding target data.

  In the first update data information, the process No. 104a1 is “1”, the logical data name 104a2 is “logical data A”, the replication destination information 104a3 is “replication destination DB21”, and the replication data 104a4 is “ID”. Information: ID001 + Address information: Tokyo □□ ... ”, variance value 104a5 is“ 1/2 (value where ID001 to ID500 are number 1/2) ”, and processing flag 104a6 is“ unprocessed ”. The processing server information 104a7 is update data information that is “processing server 31”.

  In the second update data information, the process No. 104a1 is “2”, the logical data name 104a2 is “logical data A”, the replication destination information 104a3 is “replication destination DB22”, and the replication data 104a4 is “ID”. Information: ID001 + Address information: Tokyo □□ ... ”, variance value 104a5 is“ 1/2 (value where ID001 to ID500 are number 1/2) ”, and processing flag 104a6 is“ unprocessed ”. And the processing server information 104a7 is update data information indicating “processing server 32”.

  In the third update data information, process No. 104a1 is “3”, logical data name 104a2 is “logical data A”, replication destination information 104a3 is “replication destination DB21”, and replication data 104a4 is “ID”. Information: ID567 + address information: Chiba □□..., Variance value 104a5 is “1/2” (value with ID501 to ID999 being 1/2), and processing flag 104a6 is “unprocessed”. And the processing server information 104a7 is update data information indicating “processing server 33”.

  In the fourth update data information, the process No. 104a1 is “4”, the logical data name 104a2 is “logical data A”, the replication destination information 104a3 is “replication destination DB22”, and the replication data 104a4 is “ID”. Information: ID567 + address information: Chiba □□..., Variance value 104a5 is “1/2” (value with ID501 to ID999 being 1/2), and processing flag 104a6 is “unprocessed”. And the processing server information 104a7 is update data information indicating “processing server 34”.

  If the data (target data) whose variance value is only 1/2 is updated, the process No. 1 and process no. 2 is created.

  At this time, if processing information is added to logical data (for example, logical data A), the control unit 106 also adds the processing information to the update data information.

  Next, the update operation performed by the processing servers 31 to 3i will be described.

  In each of the processing servers 31 to 3i, when the processing unit 302 notifies the update notification to the effect that the management system 1 (for example, the control unit 106) has updated the replication source DB 101 by broadcast, for example, by broadcast. When the update notification is received or when the process currently being processed ends, the update data storage unit 104 is connected.

  Subsequently, the processing unit 302 sends the own processing server from the update data storage unit 104 based on the processing server information 103a6 of the management information 103a in the storage unit 301 or based on the processing server information 104a7 of the update data information 104a. The update data information 104a of the process No to be processed is acquired.

  In the management system 1, the control unit 106 includes the processing flag 104 a 6 in the update data information 104 a for the update data information 104 a acquired by the processing server among the update data information 104 a in the update data storage unit 104. Change the status of to "Processing".

  For example, the processing unit 302 in the processing server 33 changes the update data information to be updated by its own device based on the processing server information 103a6 of the management information 103a in the storage unit 301 (logical data name, distributed value, etc.). : Logical data A & distributed value: 2/2 update data information or logical data name: logical data B & distributed value: 3/3 update data information) or update data in the update data storage unit 104 The information 104a (FIG. 6) is confirmed, and based on the processing server information 104a7, the update data information to be updated is identified by the process No. (update data information of process No. 3) in the update data storage unit 104 Among the update data information specified in the update data information 104a, the process No. 3 update data information 104 a is acquired as update data information to be processed by the processing server 33.

  Subsequently, the processing unit 302 in the processing server 33 performs processing No. The update data acquisition notification indicating that the third update data information 104 a has been acquired by the processing server 33 is transmitted to the control unit 106 in the management system 1.

  The control unit 106 performs processing No. 3 is received by the processing server 33, the process No. specified in the update data acquisition notification is received. The information represented by the processing flag 104a6 in the update data information 104a of No. 3 is changed from “unprocessed” to “processing”.

  When the processing unit 302 acquires the update data information 104a, and processing information is added to the update data information 104a, the processing unit 302 performs a predetermined process according to the processing information on the duplicate data 104a4 in the update data information 104a. . Note that the processing unit 302 may store processing information for each item of target data in advance, and perform predetermined processing according to the processing information of the item of target data on the target data in the replicated data 104a4. .

  Subsequently, the processing unit 302 transmits the copy data or the processed copy data to the copy destination DB specified by the copy destination information 103a3 in the management information 103a corresponding to the acquired update data information 104a, and Of the corresponding data in the replication destination DB, the corresponding data associated with the corresponding data for the related target data in the replicated data 104a4 is updated data (corresponding data) in the replicated data 104a4 or predetermined processing Is updated to the updated data.

  For example, when the corresponding data for the target data before the update is a copy of the target data before the update (when processing information is not added to the update data information 104a), the processing unit 302 stores the data in the replication destination DB. The corresponding data associated with the corresponding data for the related target data is updated to a copy of the updated target data.

  When the correspondence data for the target data before update is processed data obtained by performing a predetermined process on the target data before update (when processing information is added to the update data information 104a), the processing unit 302 updates the corresponding data associated with the corresponding data for the related target data in the replication destination DB to the processed data obtained by performing a predetermined process on the updated target data.

  Subsequently, the processing unit 302 transmits, to the control unit 106 in the management system 1, a process completion report indicating that the update process has been completed for the update data information 104a of process No. for which the update process has been completed.

  Upon receiving the processing completion report, the control unit 106 changes the state of the processing flag 104a6 in the update data information 104a that has received the processing completion report to “processed”.

  FIG. 8 is a diagram illustrating processing of update data information by the processing server.

  Further, in the processing servers 31 to 3 i, the processing unit 302 uses the load status confirmation unit 303 to confirm the operating level of the own server. It should be noted that the timing for confirming the operation level (for example, load) may be periodically or when the process of updating data in the replication destination DB is completed.

  The processing unit 302 confirms the operation level of the own server. If the operation level of the own server is below a certain threshold, the processing unit 302 starts an additional process, and then processes the update data information 104a in the update data storage unit 104. Check the flag 104a6.

  Subsequently, the processing unit 302 acquires one piece of update data information 104a to be processed by the own processing server from the processing data information 104a including the processing flag 104a6 representing “unprocessed”.

  Subsequently, the processing unit 302 transmits to the control unit 106 in the management system 1 a server change notification indicating that the update data information 104a has been acquired by the own processing server.

  Upon receipt of the assigned server change notification, the control unit 106 changes the processing server indicated by the processing server information 104a7 in the update data information 104a designated by the assigned server change notification to the processing server that is the transmission source of the assigned server change notification. Then, the processing flag 104a6 in the processing server information 104a is changed to “processing”.

  The processing unit 302 counts the update data information 104a in which the processing flag 104a6 is “unprocessed” for each processing server designated by the processing server information 104a7, and the unprocessed processing server having the largest count value is processed. Update data information 104a to be processed may be selected from the item (update data information 104a).

  Hereinafter, an example will be described.

  FIG. 9 is a diagram illustrating an example of the operating rate (operating degree) of the processing servers 31 to 33 and the update data information in the update data storage unit 104.

  In FIG. 9, it is assumed that the operating rate of the processing server 31 is 70%, the operating rate of the processing server 32 is 60%, and the operating rate of the processing server 33 is 20%, and the threshold is 30%.

  In this case, the processing unit 302 in the processing server 33 confirms the update data information 104a (FIG. 9) in the update data storage unit 104, and the processing server having the largest number of “unprocessed” update data information (hereinafter “processing server”). 31 ”), which is unprocessed update data information. 1 update data information 104 a is acquired as update data information to be processed by the processing server 33.

  Subsequently, the processing unit 302 in the processing server 33 performs processing No. The server-in-charge change notification indicating that the update data information 104a of 1 is acquired by the processing server 33 is transmitted to the control unit 106 in the management system 1.

  The control unit 106 performs processing No. 1 is received by the processing server 33, the process No. specified in the notification of server change notification is received. The processing server indicated by the processing server information 104a7 in the update data information 104a is changed to the processing server 33 that is the transmission source of the responsible server change notification. The processing flag 104a6 in the one update data information 104a is changed to “processing”.

  FIG. 10 is a diagram illustrating a proxy process of update data information by the processing server.

  Note that the processing unit 302 in the processing server 33 may acquire any of the unprocessed update data information 104a of the processing server having the highest operating rate as update data information to be processed by the processing server 33. In that case, in the processing servers 31 to 3i, the load status confirmation unit 303 informs the server load confirmation unit 105 in the management system 1 of the availability of the own processing server periodically or when the update data information 104a is acquired. Etc. The server load confirmation unit 105 stores the operating level of each processing server as processing server load information 103b.

  Next, the effect of this embodiment will be described.

  In the present embodiment, the storage unit 103 stores logical data (target data specifying information) for specifying target data.

  The processing servers 31 to 3 i refer to the logical data in the storage unit 103, and when one of the target data specified by the logical data is updated in the replication source DB 101, the processing data of the target data in the replication source DB 101 is updated. Of these, update processing that reflects the update of the target data in the replication source DB 101 in the replication destination DB 21 or 22 is executed using the replicated data (update processing data) including the updated target data.

  For this reason, only the update processing data necessary for updating the corresponding data in the replication destination DB 21 or 22 among the data in the replication source DB 101 is provided to one or more replication source DBs without a large time lag.

  Of the data in the replication source DB 101, only the update processing data necessary for updating the target data in the replication destination DB 21 or 22 is used, for example, when one part of the table in the replication source DB 101 is updated. Therefore, it is not necessary to replicate all the tables in the replication source DB 101. In addition, by processing the data specified by the logical data, the amount of data (information amount) to be replicated is narrowed, so the bandwidth for data communication is secured and the number of processes on the processing server is reduced. It becomes possible to improve efficiency.

  Therefore, unnecessary data update processing can be suppressed, and the load of update processing can be suppressed from being concentrated on one processing server.

  In this embodiment, in the replication destination DB 101, each target data is associated with one or more other target data (related target data). In the replication destination DBs 21 and 22, each corresponding data is the corresponding data. Is stored in association with related correspondence data, which is correspondence data for target data (related target data) associated with the target data that is the source of. The replicated data includes updated target data and related target data that is target data associated with the updated target data.

  The processing servers 31 to 3i perform an update process in which the corresponding data associated with the corresponding data for the related target data in the replication destination DB 21 or 22 is updated to the corresponding data for the updated target data.

  For this reason, it becomes possible to identify the corresponding data to be updated in the replication destination DB 21 or 22 using the related target data.

  In this embodiment, the storage unit 103 designates, for each target data, an execution processing server that is a processing server that executes an update process for reflecting the update of the target data in the replication source DB 101 to the replication source DB 21 or 22. Processing server information (server designation information and distributed value) is stored.

  Each of the processing servers 31 to 3i is associated with the updated specific target data and the updated specific target data when the specific target data designated as the execution processing server in the processing server information is updated. The update processing is executed using the related target data obtained.

  For this reason, it becomes possible to designate an execution processing server for the target data.

  In this embodiment, the control unit 106 refers to the logical data in the storage unit 103 and updates the target data in the replication source DB 101 and updates the target data in the replication source DB 101. The related target data associated with the target data is associated with each other and stored in the update data storage unit 104.

  Each of the processing servers 31 to 3i updates the specified target data in the update data storage unit 104 when the specified target data is updated in a situation where the own server is designated as an execution processing server in the processing server information. And the related target data associated with the updated specific target data are used to execute the update process.

  For this reason, the control part 106 prepares the data used for an update process, and it becomes possible for each of the process servers 31-3i to perform an update process using the prepared data.

  In the present embodiment, each of the processing servers 31 to 3i has target data corresponding to processing server information for designating another processing server as an execution processing server when the operating level of the own server is lower than a predetermined operating level. The update process using the update result and the related target data associated with the update result of the target data is executed on behalf of the other processing server.

  For this reason, when there is a difference in load between the processing servers, the processing server with less load processes the unprocessed replicated data, and thus the distributed processing in the processing servers 31 to 3i can be efficiently performed. .

  The other processing server has the highest operating rate among the processing servers 31 to 3i having the largest number of unexecuted update processes assigned to the processing server or the processing servers 31 to 3i. A processing server is desirable. In this case, distributed processing in the processing servers 31 to 3i can be performed efficiently.

  Note that the management server 1 (for example, the control unit 106) may change the processing server information based on the operation level of each processing server without determining the update processing to be performed by the processing server.

  For example, the control unit 106 refers to the processing server load information 103b in the storage unit 103, and in the processing servers 31 to 3i, a processing server (hereinafter referred to as a “low operating processing server”) whose operating rate is lower than a predetermined threshold. In order to assign the update data information assigned to the processing server having a large number of unprocessed update data information or having a high operation level to the low operation processing server among the update data information, The processing server specified by the processing server information in the update data information or the processing server specified by the processing server information in the management information used when creating the processing server information is changed to a low operation processing server.

  In this case, it is possible to reduce the load on the processing server as compared with the case where the processing server performs a process for determining the update process to be performed.

  In the above-described embodiment, the values of ID001 to ID500 are set to 1/2 and ID501 to ID999 are set to 2/2 as the variance values. However, the variance values can be changed as appropriate.

  For example, when the product name data is divided into three, the data from the product name starting with the “a” line to the data starting with the “sa” line is numbered 1/3, and the product name starts with the data starting with the “ta” line. The value starting with the data starting with the line "2/3" and the product name starting with the "ma" line to the data starting with the "wa" line with the number 3/3 may be used as the distributed value. .

  Further, the method of distributing the update process to a plurality of servers can be appropriately changed, such as a method using a distributed hash table.

  In this embodiment, one processing server is assigned as one execution processing server for one management information. However, for redundancy, a plurality of processing servers are assigned as one execution processing server for one management information. May be.

  Further, the management system 1 shown in FIG. 1 excluding the replication source DB 101 may be realized by a computer. In this case, the computer reads and executes a program recorded on a recording medium such as a CD-ROM (Compact Disk Read Only Memory) readable by the computer, whereby the update unit 102, the storage unit 103, and the update data storage Functions as a unit 104, a server load confirmation unit 105, and a control unit 106. The recording medium is not limited to the CD-ROM and can be changed as appropriate.

  Each processing server may be realized by a computer. In this case, the computer functions as the storage unit 301, the processing unit 302, and the load status confirmation unit 303 by reading and executing a program recorded on a computer-readable recording medium.

  In the embodiment described above, the illustrated configuration is merely an example, and the present invention is not limited to the configuration.

100 Database management system 1 Management system 101 Source DB
102 update unit 103 storage unit 104 update data storage unit 105 server load confirmation unit 106 control unit 21, 22 replication destination DB
31 to 3i processing server 301 storage unit 302 processing unit 303 load status confirmation unit 4 user terminal 5 administrator terminal

Claims (10)

A first database for storing a plurality of data;
For a plurality of data in the first database, a plurality of second data for storing corresponding data that is either a copy of the target data or processed data obtained by performing a predetermined process on the target data. Two databases,
Second database specifying information for specifying any one of the plurality of second databases, and target data specifying information for specifying the target data necessary for the second database specified by the second database specifying information Storage means for storing
In the first database, when at least one of the plurality of data is updated, it is determined whether one of the target data specified by the target data specifying information in the storage unit is updated, Control means for creating update processing data including the updated target data when the target data is updated;
A processing server that uses the update processing data to execute an update process that reflects the update of the target data in the first database in the corresponding data in the second database specified by the second database specifying information And a database management system including. The database management system according to claim 1,
In the first database, each target data is stored in association with related target data that is one or more other target data,
In the second database, each corresponding data is stored in association with related corresponding data, which is corresponding data for related target data associated with the target data that is the source of the corresponding data,
In the target data specifying information, the target data and the related target data are associated,
The control means, with the update of the target data in the first database, the update specified by the updated target data in the first database and target data specifying information in the storage means Update processing data including related target data, which is target data associated with the target data,
The processing server obtains the update processing data, and updates the correspondence data associated with the correspondence data for the related target data in the second database to the correspondence data for the updated target data A database management system that performs processing. The database management system according to claim 2,
When the corresponding data for the target data before the update is a copy of the target data before the update, the processing server is associated with the corresponding data for the related target data in the second database. The corresponding data is updated to a copy of the updated target data, and the corresponding data for the target data before the update is processed data obtained by performing the predetermined processing on the target data before the update In the database management system, the correspondence data associated with the correspondence data for the related target data in the second database is updated to processed data obtained by performing the predetermined processing on the updated target data. The database management system according to any one of claims 1 to 3,
The storage means includes, for each target data, an execution processing server that is a processing server that executes update processing for reflecting the update of the target data in the first database to the second database among the plurality of processing servers. Memorize server designation information for designation,
Each of the plurality of processing servers executes the update processing using the update processing data including the specific target data designated as the execution processing server based on the server designation information in the storage means. A database management system. The database management system according to claim 4, wherein
Storage means for storing the update processing data created by the control means;
When each of the plurality of processing servers stores the update processing data including the specific target data designated as the execution processing server by the server designation information, A database management system that executes the update process using the update process data including the specific target data. In the database management system according to claim 4 or 5,
When each of the plurality of processing servers has an operation level of its own server lower than a predetermined operation level, another server of the plurality of processing servers displays an update result of the target data designated as the execution processing server. A database management system that executes the update processing on behalf of the other processing server using the update processing data included. The database management system according to claim 6, wherein
The other processing server is the processing server having the largest number of unexecuted update processes assigned to the processing server among the plurality of processing servers, or the highest operating degree among the plurality of processing servers. A database management system that is a processing server. In the database management system according to claim 4 or 5,
Confirmation means for accepting the operating level of each of the plurality of processing servers;
A database management system comprising: setting means for setting or changing the server designation information based on the operating degree received by the confirmation means. The database management system according to claim 8, wherein
The setting means is assigned to the processing server among the plurality of processing servers when there is a low operating level processing server whose operating level is lower than a predetermined operating level among the plurality of processing servers. Server designation information for designating the processing server with the highest number of unexecuted update processes or the processing server with the highest operating level among the plurality of processing servers as the execution processing server, the low operating level processing server with the low operating level processing server A database management system that changes to server specification information for specifying an execution processing server. Any one of a first database that stores a plurality of data, a target data among a plurality of data in the first database, a copy of the target data, and processed data obtained by performing a predetermined process on the target data A plurality of second databases for storing corresponding data, second database specifying information for specifying any one of the plurality of second databases, and a second database specified by the second database specifying information A database management method performed by a database management system including: storage means for storing target data specifying information for specifying the target data necessary for control; a control means; and a processing server,
When the control unit updates at least one of the plurality of data in the first database, one of the target data specified by the target data specifying information in the storage unit is updated. And when the target data is updated, create update processing data including the updated target data,
An update process in which the processing server reflects the update of the target data in the first database to the corresponding data in the second database specified by the second database specifying information using the update processing data. Execute database management method.

Images