JP2011238072A - Database management device, database system, and method and program for database management - Google Patents

Abstract

PROBLEM TO BE SOLVED: To distribute data contents so that there is no duplication of contents of a plurality of databases.
In the DB management device, the processing content received from the processing content input device is transmitted to any of a plurality of DB devices, and the DB device refers to the data block identifier and the DB identifier referred to by the processing content. Is acquired from the memory management means and transmitted to the DB management device. The DB management device stores the correspondence information in the reference destination data block storage means, and a processing request is input. Then, a data block identifier and a DB identifier that match the processing are acquired from the reference destination data block storage means, and it is determined whether the data block identifier is associated with any DB identifier. In this case, the data block identifier is associated with the DB identifier having the smallest number of associated data blocks in the correspondence storage means. The processing input from the processing input device is distributed to the DB that is stored and stored with the largest number of data block identifiers.
[Selection] Figure 9

Description

  The present invention relates to a database management apparatus, a database system, a database management method, and a program, and more particularly, to a database management apparatus for distributing loads of a plurality of databases that are distributed and arranged on a network and data contents in a database are synchronized. And a database system, a database management method, and a program.

  As a technique for distributing the processing so that the load of the nodes distributed on the network is equalized, as shown in FIG. 11, it waits until there is an input from the processing input device (step 10), and the CPU and I / O Using a statistical information such as a load, a database to which an input processing request is distributed is determined (step 20), and the process is transmitted to a distribution destination database via a “process transmission unit” (step 30). Yes (see, for example, Patent Document 1).

  In addition, there is a technique for equalizing memory usage when distributing processes (see, for example, Non-Patent Document 1).

JP 2004-118622 A, “Load Balancing Device, Method, and Program”

Sameh Elnikety, et al., "Tashkent +: Memory-Aware Load Balancing and Update Filtering in Replicated Database", SIGOPS Oper. Syst. Rev., Vol. 41, No. 3, pp. 399-412, 2007.

  However, conventionally, as in the technique of the above-mentioned Patent Document 1, when an input process is distributed, a load factor such as a CPU or I / O (input / output) is often used as a material for determining the distribution. Since the processing is distributed without considering the data area (data block) on the machine to which each processing refers, the memory contents (data blocks on the memory) on a plurality of machines overlap as shown in FIG. There is a problem that the memory hit rate decreases.

  In the method of Non-Patent Document 1, memory usage is equalized at the time of processing distribution, but as described above, duplication of memory contents occurs and the memory hit rate decreases.

  As described above, conventionally, there is a problem that there is no mechanism for removing duplication of memory contents in each database.

  The present invention has been made in view of the above points. In an environment where processing described in SQL is distributed to a plurality of databases whose data contents are synchronized for load distribution, there is no duplication of contents on the databases. An object of the present invention is to provide a load distribution method, apparatus, and program for a distributed arrangement database that can be distributed.

In order to solve the above problem, the present invention (Claim 1) is a database management device that distributes processing to a plurality of databases that are distributed on a network and in which data contents in the database are synchronized.
The processing content received from the processing content input device that inputs N processing content is transmitted to any of the plurality of database devices, and the processing content and the data block identifier of the data block referred to by the processing content are transmitted from the database device, Database information collecting means for acquiring correspondence information of database identifiers and storing the information in reference destination data block storage means;
A reference data block determination unit that receives a process from a process input device that inputs a process request, and acquires a data block identifier and a database identifier that match the process based on the content of the requested process;
Correspondence storage means for holding a correspondence between a database identifier and a database identifier;
Referring to the correspondence relationship storage means, it is determined whether the data block identifier acquired by the reference data block determination means is associated with any database identifier, and if not, the data block identifier is A correspondence determining means for storing in the correspondence storage means in association with the database identifier having the smallest number of associated data blocks in the correspondence storage means;
A process distribution unit that refers to the correspondence relationship storage unit and distributes the process input from the process input device to the database associated with the largest number of data block identifiers.

The present invention (Claim 2) is a plurality of database devices distributed on a network and synchronized with data contents in a database,
A memory management means having a memory and managing a database;
Acquire N processing contents from the database management apparatus that distributes the processing to the database apparatus, and acquire from the memory management means correspondence information that associates the data block identifier referred to by the processing contents with the database identifier. Database information acquisition means for transmitting to the database management device,
Memory management means
When there is no data block identifier corresponding to the processing content in the memory, the data block identifier is acquired from the database, written to the memory, and passed to the database information acquisition unit.

The present invention (Claim 3) is a database system comprising a plurality of database devices that are distributed on a network and whose data contents in the database are synchronized, and a database management device that distributes processing to the database,
The database device
A memory management means having a memory and managing a database;
Acquire N processing contents from the database management apparatus that distributes the processing to the database apparatus, and acquire from the memory management means correspondence information that associates the data block identifier referred to by the processing contents with the database identifier. Database information acquisition means for transmitting to the database management device,
Memory management means
When there is no data block identifier corresponding to the processing content in the memory, the data block identifier is acquired from the database, written to the memory, and includes means for passing to the database information acquisition unit,
Database management device
The processing content received from the processing content input device that inputs N processing content is transmitted to any of the plurality of database devices, and the processing content and the data block identifier of the data block referred to by the processing content are transmitted from the database device, Database information collecting means for acquiring correspondence information of database identifiers and storing the information in reference destination data block storage means;
A reference data block determination unit that receives a process from a process input device that inputs a process request, and acquires a data block identifier and a database identifier that match the process based on the content of the requested process;
Correspondence storage means for holding a correspondence between a database identifier and a database identifier;
Referring to the correspondence relationship storage means, it is determined whether the data block identifier acquired by the reference data block determination means is associated with any database identifier, and if not, the data block identifier is A correspondence determining means for storing in the correspondence storage means in association with the database identifier having the smallest number of associated data blocks in the correspondence storage means;
A process distribution unit that refers to the correspondence relationship storage unit, and distributes the process input from the process input device to the database associated with the largest number of data block identifiers;
Have

The present invention (Claim 4) is a database management method in a system comprising a plurality of database devices that are distributed on a network and whose data contents in a database are synchronized, and a database management device that distributes processing to the database devices. Because
Database management device
The database information collection means transmits the processing content received from the processing content input device that inputs N processing content to any of the plurality of database devices,
The database device
The database information transmitting means acquires N processing contents, acquires correspondence information associating the data block identifier referred to by the processing contents with the database identifier from the memory management means for managing the database, and To the management device,
The database information collection means of the database management device
Acquire correspondence information from the database device, store it in the reference data block storage means,
The reference block determining means of the database management device is
Accepting the processing from the processing input device that inputs the processing request, based on the content of the requested processing, obtains the data block identifier and database identifier that match the processing from the reference destination data block storage means,
The correspondence determining means of the database management device is
Whether the data block identifier is associated with any one of the database identifiers is determined by referring to the correspondence storage means that holds the correspondence between the database identifier and the database identifier. The block identifier is stored in the correspondence storage unit in association with the database identifier having the smallest number of associated data blocks in the correspondence storage unit,
The processing distribution means of the database management device is
With reference to the correspondence relationship storage means, the processing input from the processing input device is assigned to the database with the largest number of data block identifiers.

The present invention (Claim 5) is the database apparatus according to Claim 4,
When acquiring N pieces of processing contents and acquiring the correspondence information associating the data block identifier and the database identifier referred to by the processing contents from the memory management means for managing the database,
If there is no data block identifier corresponding to the processing content in the memory in the memory management means, the data block identifier is acquired from the database, written to the memory, and transmitted to the database management device.

  The present invention (Claim 6) is a program for causing a computer to function as each means constituting the database management apparatus according to Claim 1.

  The present invention (Claim 7) is a program for causing a computer to function as each means constituting the database management apparatus according to Claim 2.

  As described above, according to the present invention, based on the processing content input from the processing content input device, data block information to be referred to when a certain process is executed on the database is collected, and the reference data block storage The information about which data block on the database is referred to by the process input from the process input device is stored in the means from the information in the reference data block storage means {process, database, data block} By using the information associated with, it can be distributed so as not to overlap between databases.

  In addition, since there is no duplication of the memory contents of each database, the corresponding area becomes a free area, so the database data blocks that have not been stored in the memory until now are stored in the memory. The memory hit rate is improved. Thereby, when compared with the prior art, the average processing completion time of the input processing is shortened, and the memory can be effectively used.

1 is a system configuration diagram according to an embodiment of the present invention. It is a flowchart of a series of processes in one embodiment of the present invention. It is a flowchart of the process of the database information initialization part in one embodiment of this invention. It is a flowchart of the process by the side of the database at the time of the database information initialization in one embodiment of this invention. It is a flowchart of the process of the reference data block determination part of the process of step 300 in one embodiment of this invention. It is a flowchart of the process of the reference data block of a process and database correspondence determination part of one Embodiment of this invention. It is a figure which shows the process which allocates the data block which is not matched in one embodiment of this invention to a database. It is a figure which shows the distribution process of the database in one embodiment of this invention. It is a sequence chart of a series of processes in one embodiment of the present invention. It is an example of the process distribution in one embodiment of this invention. It is a flowchart of the distribution process of a prior art. It is an example of the distribution of the conventional database.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 shows a system configuration according to an embodiment of the present invention.

  The system shown in the figure includes a processing input device 10, a result output device 20, a processing content input device 30, a machine A100, a machine B200, and a machine C300.

  The machine A100 is a device that operates as a database management device, receives processing content (processing request) executed from the processing input device 10, and distributes this processing to the machines B200 and C300 that operate as the database device. The processing result acquired from C300 is output to the result output device 20. Since the machine B 200 and the machine C 300 synchronize data blocks via a data synchronizer described later, the process received by the machine A 100 can be executed by either machine. However, the processing speed differs depending on the data block existing on the memory managed by the memory management unit of the machine B200 or machine C300. The machine A100 recognizes the data block referred to by the processing content and distributes the processing, thereby improving the memory hit rate of the machines B200 and C300.

  The machine A100 includes a processing reference data block determination unit 110, a processing reference data block and database correspondence determination unit 120, a processing reference data block storage unit 130, a correspondence relationship storage unit 140, a database information initialization unit 150, and a database. An information transmitting / receiving unit 160, a processing distribution unit 170, and a processing result receiving unit 180 are configured.

  The database information initialization unit 150 transmits N processing contents input from the processing content input device 30 from the database information transmission / reception unit 160 to the machine B200 or C300, and the processing is performed from these machines (databases). The relationship of the data block to be referred to (processing content, data block identifier, database identifier) is acquired and recorded in the reference data block storage unit 130 of the processing. From the processing content input device 30, all processing content to be handled in the system (processing content such as “search for xxx”, “search for yyy”) is the database processing language SQL (SELECT… FROM…). Input using.

  The process reference data block storage unit 130 stores the data block identifier used for each process by the database information initialization unit 150 and the identifier of the database storing the data block. The process reference data block determination unit 110 stores the data block identifier. Referenced.

  The database information transmission / reception unit 160 transmits the processing content input from the processing content input device 30 and acquired through the database information initialization unit 150 to the machine B200 or C300, and performs the processing from the machine B200 or machine C300. The corresponding data block identifier and database identifier are acquired and passed to the database information initialization unit 150.

  When the process reference data block determination unit 110 (hereinafter referred to as “reference destination determination unit”) acquires the process contents from the process input device 10, the process reference data block determination unit 110 refers to the process reference data block storage unit 130 to input the process. The data block referred to is searched, and the processing content provided with the data block identifier to be referred to is transferred to the correspondence determining unit 120 between the reference data block of the processing and the database. Note that a processing request (for example, “search for xxx”) designated by the user is input from the processing input device 10.

  The reference data block for processing and the correspondence determining unit 120 (hereinafter referred to as “corresponding relationship determining unit”) of the database acquire the processing content provided with the data block identifier to be referred to from the reference destination determining unit 110 to obtain the data After determining the correspondence between the block and the database and recording the correspondence in the correspondence storage unit 140, the database having the most data blocks to which the requested process refers is determined, and the processing content to which the information related to the processing reference information is given is determined. The data is transferred to the processing distribution unit 170.

  The processing distribution unit 170 acquires the processing content to which the data block identifier referred to from the correspondence determining unit 120 is attached, and based on the information attached to the received processing content and statistical information such as CPU and I / O loads. Then, the database (machine B200 or machine C300) to which the process is distributed is determined, and the process is transmitted to the database.

  The correspondence storage unit 140 stores the correspondence between the data block and the database.

  The processing result receiving unit 180 outputs the execution result received from the machine B 200 or the machine C 300 to the result output device 20.

  Next, the machine B200 and the machine C300 will be described.

  The machine B200 and the machine C300 execute the process received from the process distribution unit 170 of the machine A100. Data blocks necessary for processing execution are acquired using respective memory management units 260 and 360 described later, and are transmitted to the machine A100. At this time, the data blocks that are not on the memories 261 and 361 of the memory management units 260 and 360 are acquired from the data storage units 270 and 370 by the memory management units 260 and 360, and replaced with the data blocks on the memories 261 and 361. I do. In addition, a data block identifier referred to by the processing content transmitted from the database information initialization unit 150 of the machine A100 is also searched. In addition, the replacement of data blocks between the memories 261 and 361 and the data storage units 270 and 370 stores data blocks with high reference frequency on the memories 261 and 361. The reference frequency can be managed by counting each data block in the memory management units 260 and 360.

  The machine B 200 includes a database information transmission / reception unit 210, a database processing transmission / reception unit 220, a database information acquisition unit 240, a database processing unit 250, a memory management unit 260, a data storage unit 270, and a data synchronization unit 280. Since the machine C300 has the same configuration as the machine B200, the constituent elements are not described.

  The database information transmission / reception units 210 and 310 output the processing content received from the database information transmission / reception unit 160 of the machine A 100 to the database information acquisition unit 240.

  The database information acquisition units 240 and 340 search the data block identifier via the database processing units 250 and 350 and the memory management units 260 and 360 based on the input processing content, and the processing content and the reference data block An identifier correspondence table is created in a buffer (not shown) in the database information acquisition units 240 and 340.

  The database processing transmission / reception units 220 and 320 output the processing acquired from the processing distribution unit 170 of the machine A100 to the database processing units 250 and 350, and the processing results acquired from the database processing units 250 and 350 are the processing results of the machine A100. Transmit to the receiver 180.

  The database processing units 250 and 350 acquire necessary information from the memory management units 260 and 360, generate processing results, and pass them to the database information acquisition units 240 and 340 or the database processing transmission / reception units 220 and 320.

  The memory management units 260 and 360 retrieve the data block requested from the database processing unit 250 from the data block identifier, and if it is on the memory, return it to the database processing units 250 and 350. Search from the data storage units 270 and 370, replace the data block on the memory, and return the data.

  When the data synchronization units 280 and 380 are updated in the data storage units 270 and 370 by the memory management units 260 and 360, the data synchronization units 280 and 380 similarly perform the update processing in the paired data storage units 370 and 270, Synchronize data contents.

  A series of operations in the above configuration will be described below.

  FIG. 2 is a flowchart of overall processing in one embodiment of the present invention.

  In the following processing, step 100 is preprocessing for storing the correspondence between processing contents and data blocks in the reference data block storage unit 130 of processing, and steps 200 to 600 are processing for distributing requested processing. In the preprocessing, the processing content input device 30 connected to the database information initialization unit 150 is used, and in the distribution processing, the processing input device 10 connected to the reference destination determination unit 110 is used.

  Step 100) The database initialization unit 150 of the machine A 100 refers to which data block on the database (Machine B or Machine C in the present embodiment) the N process contents input from the process content input device 30 refer to. The memory management unit 260 of the machine B200 and the machine C300 is searched for, and the result (correspondence between the processing content and the reference data block) is stored in the reference data block storage unit 130 for processing. Detailed processing will be described in detail with reference to FIG.

  Step 200) Wait until the next processing request is input from the processing input device 10.

  Step 300) When a process request is input from the process input device 10, the reference destination determining unit 110 refers to the reference data block storage unit 130 of the process and searches for a data block referred to by the input process. A data block identifier to be referred to is assigned to the process and passed to the correspondence determining unit 120. Detailed processing will be described in detail with reference to FIG.

  Step 400) The correspondence relationship determination unit 120 determines the correspondence relationship between the data block identifier and the database based on the processing content to which the data block identifier to be referenced is assigned. Detailed processing will be described in detail with reference to FIG.

  Step 500) In the process distribution unit 170, the database to which the input process refers most frequently and the other statistical information (CPU and I / O) are used to finally determine the database to which the process is distributed.

  Step 600) Based on the information given to the received processing contents, the processing is distributed to the corresponding database (machine B or machine C).

  Next, the database information initialization process (step 100) will be described.

  FIG. 3 is a flowchart of the process of the database information initialization unit in one embodiment of the present invention.

  Step 110) The database information initialization unit 150 uses the database information transmission / reception unit 160 to transmit the N processing contents (described in SQL) to be handled by the system input from the processing content input device 30, using the machine B 200 and the machine C 300. Send to.

  Step 120) Wait until data (correspondence table between processing contents and data blocks to be referenced) is received from the database information transmitting / receiving unit 210 of the machine B200 and the database information transmitting / receiving unit 310 of the machine C300.

  Step 130) The data received from the machine B200 and the machine C300 (correspondence table between the processing content and the data block to be referred to) is stored in the reference data block storage unit 130 for processing.

  Next, the operation on the machine B 200 and machine C 300 side when the database information is initialized (step 100) in FIG. 2 will be described.

  FIG. 4 is a flowchart of processing on the database side at the time of database initialization according to the embodiment of the present invention.

  Step 121) The database information acquisition units 240 and 340 of the respective machines B200 and C300 stand by until data (N processing contents) is received from the database information transmission / reception units 210 and 310.

  Step 122) The counter i is initialized to 1.

  Step 123) The database information acquisition units 240 and 340 inquire the database processing units 250 and 350 about the i-th processing content to acquire the data block identifier to be referred to when the processing is executed.

  Step 124) The database information acquisition units 240 and 340 store the i-th processing content and the data block identifier referred to by the processing in a transmission buffer (not shown).

  Step 125) The database information acquisition units 240 and 340 determine whether or not the counter i exceeds N. If it exceeds, the process proceeds to step 126. If not, the process proceeds to step 127.

  Step 126) Add 1 to the counter i and proceed to Step 123.

  Step 127) The information of the identifier of the data block to be referred to is added to the processing content, the process proceeds to Step 121, and the reception of the process from the next machine A100 is awaited.

  Next, the operation of the process (step 300) of the reference destination determination unit 110 in FIG. 2 will be described in detail.

  FIG. 5 is a flowchart of the process of the reference data block determination unit in the process according to the embodiment of the present invention.

  Step 310) The reference destination determination unit 110 performs disassembly processing on the input processing request. For example, it is converted into a form that can be easily processed by the machine.

  Step 320) Based on the process request, the reference data block storage unit 130 is searched for a match with the content of the input process request, and a data block identifier to be referenced is acquired.

  Step 330) The reference destination determining unit 110 checks whether or not the acquired data block identifier to be referenced exists. When it exists, it transfers to step 340, and when it does not exist, the said process is complete | finished.

  Step 340) The information of the identifier of the data block to be referred to is added to the processing content and passed to the correspondence determining unit 120.

  Next, the operation of the processing (step 400) of the correspondence determining unit 120 in FIG. 2 will be described in detail.

  FIG. 6 is a flowchart of the process of the reference data block and the database correspondence determining unit in the process according to the embodiment of the present invention.

  Step 410) The correspondence determining unit 120 checks whether the data block referred to by the process is already associated with the database (Machine B or Machine C). If the data block is associated, the process proceeds to Step 430. If not, the process proceeds to step 420.

  Step 420) The correspondence determining unit 120 assigns the identifier of the reference data block that has not been associated to the database having the smallest number of associated data blocks. At this time, as shown in FIG. 7, the correspondence determining unit 120 refers to the correspondence storing unit 140, and the total number (7) of databases B (machine B) associated is the database C (machine C). Since it is less than the total number of associations (8), the identifier of the reference data block is assigned to machine B.

Step 430) Based on the data block identifier to be referred to, it is determined to distribute the process to the database in which the most data blocks to which the process refers are associated.
In the example of FIG. 8, since many data block identifiers “2” and “8” referred to by the machine B 200 are included, the data block identifiers “2” and “8” are distributed to the machine B 200.

  FIG. 9 is a sequence chart of a series of processes in one embodiment of the present invention.

  When the machine A100 receives the processing request from the processing content input device 30 (step 501), the processing content of the processing request is randomly transmitted to the DB. Here, it is assumed that the data is transmitted to the machine B 200 (step 502).

  In the machine B200, the data block referred to by the process received from the data storage unit 270 is acquired (step 503), the correspondence between the data block and the processing content is accumulated in the buffer (step 504), and all the processes related to the process are stored. A data block is acquired, a correspondence relationship of the data block with the processing content is generated, an identifier of the DB is given, and the data block is transmitted to the machine A (step 505).

  The machine A100 assigns a data block identifier to each process and stores the correspondence between the process and the data block acquired from the machine B200 and the data of the identifier (database identifier) of the machine A in the “process reference data block storage unit 130”. (Step 506).

  The same applies to steps 507 to 512.

  When the machine A100 receives the processing request from the processing input device 10 (step 513), the reference destination determination unit 10 refers to the “processing reference data block storage unit 130” and acquires a data block identifier that matches the processing content. (Step 514).

  A processing destination, a data block identifier, and a database identifier are paired (step 515), and a DB (Machine B or Machine C) in which many data blocks referred to by the process input from the process input apparatus 10 are associated is assigned to the distribution destination candidate. (Step 516). The statistical information is referred to from the distribution destination candidates, the final distribution destination database is determined, and the process is transmitted (step 517).

  As shown in steps 501 to 512 of FIG. 9 described above, data block information to be referred to when a certain process is executed on the database (machine B or machine C) based on the information input from the process input device 10 Are collected by the database information acquisition unit 240 or 340, and processing for recording the reference information of the processing in the “processing reference data block storage unit 130” of the machine A100 is performed.

  Then, as to which data block on the database (machine B or machine C) the processing content input from the processing input device refers to, as shown in step 514 of FIG. By searching from the reference data block storage unit 130 "and using the information, the processing is distributed so that the data blocks to be referenced do not overlap among a plurality of databases.

  As described above, as shown in FIG. 9, there is no duplication of memory contents of each database. In the example of FIG. 10, the data block referred to by the process received by the machine A100 is acquired by referring to the “data block storage unit 130 referred to by the process” according to the process content, and is transmitted to the machine B200 and the machine C300. Sorting. This eliminates the duplication of the memory contents of the data storage units 270 and 370 as shown in FIG. 12, and the database data that has not been stored in the memory until now is freed up from the area used for the duplication of the memory. Blocks are stored in memory. As a result, the memory hit rate is improved. By adopting such a method, the hit probability of the memory is improved as compared with a method of evenly and randomly distributing (round robin). In the experiment, when the processing was divided into four database devices, it was confirmed that the memory hit rate was improved by about 10% compared to round robin.

  The processing operations of the machine A100 that operates as the database management apparatus, the machine B200 that operates as the database apparatus, and the machine C300 are constructed as programs, and are installed and executed on a computer that is used as the database management apparatus and database apparatus. Alternatively, it can be distributed via a network.

  In addition, the constructed program can be stored in a portable storage medium such as a hard disk, a flexible disk, or a CD-ROM, and can be installed or distributed in a computer.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

10 Process Input Device 20 Result Output Device 30 Process Content Input Device 100 Machine A (Database Management Device)
110 processing reference data block determining unit 120 processing reference data block and database correspondence determining unit 130 processing reference data block storage unit 140 correspondence storing unit 150 database information initializing unit 160 database information transmitting / receiving unit 170 processing distributing unit 180 Processing result receiving unit 200 Machine B (database device)
210, 310 Database information transmission / reception unit 220, 320 Database processing transmission / reception unit 240, 340 Database information acquisition unit 250, 350 Database processing unit 260, 360 Memory management unit 261, 361 Memory 270, 370 Data storage unit 280, 380 Data synchronization unit 300 Machine C (database device)

Claims (7)

A database management device that distributes processing on a network and distributes processing to a plurality of databases in which data contents in the database are synchronized.
The processing content received from the processing content input device for inputting N processing content is transmitted to one of the plurality of database devices, and the processing content and the data block identifier of the data block referred to by the processing content are transmitted from the database device. Database information collecting means for acquiring correspondence information of the database identifier and storing it in the reference data block storage means,
A reference data block determination unit that receives a process from a process input device that inputs a process request, and acquires a data block identifier and a database identifier that match the process based on the content of the requested process;
Correspondence storage means for holding a correspondence between a database identifier and a database identifier;
Referring to the correspondence relationship storage means, it is determined whether the data block identifier acquired by the reference data block determination means is associated with any database identifier. A correspondence determining means for storing the identifier in the correspondence storage means in association with the database identifier having the smallest number of associated data blocks in the correspondence storage means;
Referring to the correspondence relationship storage means, processing distribution means for distributing the processing input from the processing input device to a database that is associated with the largest number of data block identifiers;
A database management apparatus comprising: A plurality of database devices distributed on the network and synchronized in the database.
A memory management means having a memory and managing a database;
From the database management device that distributes processing to the database device, N processing contents are acquired, and correspondence information that associates a data block identifier and a database identifier that are referred to by the processing content is obtained from the memory management unit. Database information acquisition means for acquiring and transmitting to the database management device,
The memory management means includes
A database that includes a means for acquiring a data block identifier from the database and writing the data block identifier to the database information acquisition means when there is no data block identifier corresponding to the processing content in the memory; apparatus. A database system comprising a plurality of database devices distributed on a network and synchronized in the database, and a database management device that distributes processing to the database,
The database device includes:
A memory management means having a memory and managing a database;
From the database management device that distributes processing to the database device, N processing contents are acquired, and correspondence information that associates a data block identifier and a database identifier that are referred to by the processing content is obtained from the memory management unit. Database information acquisition means for acquiring and transmitting to the database management device,
The memory management means includes
If there is no data block identifier corresponding to the processing content in the memory, including acquiring the data block identifier from the database, writing to the memory, and passing to the database information acquisition means,
The database management device includes:
The processing content received from the processing content input device for inputting N processing content is transmitted to one of the plurality of database devices, and the processing content and the data block identifier of the data block referred to by the processing content are transmitted from the database device. Database information collecting means for acquiring correspondence information of the database identifier and storing it in the reference data block storage means,
A reference data block determination unit that receives a process from a process input device that inputs a process request, and acquires a data block identifier and a database identifier that match the process based on the content of the requested process;
Correspondence storage means for holding a correspondence between a database identifier and a database identifier;
Referring to the correspondence relationship storage means, it is determined whether the data block identifier acquired by the reference data block determination means is associated with any database identifier. A correspondence determining means for storing the identifier in the correspondence storage means in association with the database identifier having the smallest number of associated data blocks in the correspondence storage means;
Referring to the correspondence relationship storage means, processing distribution means for distributing the processing input from the processing input device to a database that is associated with the largest number of data block identifiers;
A database system comprising: A database management method in a system comprising a plurality of database devices that are distributed on a network and whose data contents in a database are synchronized, and a database management device that distributes processing to the database devices,
The database management device includes:
The database information collection means transmits the processing content received from the processing content input device that inputs N processing content to any of the plurality of database devices,
The database device includes:
The database information transmitting means acquires the N processing contents, acquires correspondence information associating the data block identifier referred to by the processing contents and the database identifier from the memory management means for managing the database, Sent to the database management device,
The database information collecting means of the database management device comprises:
Obtaining the correspondence information from the database device, storing it in a reference data block storage means,
The reference block determining means of the database management device comprises:
Accepting a process from a process input device that inputs a process request, based on the content of the requested process, obtaining a data block identifier and a database identifier that match the process from the referenced data block storage means,
The correspondence determining means of the database management device comprises:
Whether the data block identifier is associated with any database identifier is determined by referring to the correspondence storage unit that holds the correspondence relationship between the database identifier and the database identifier. The data block identifier is stored in the correspondence storage unit in association with the database identifier having the smallest number of associated data blocks in the correspondence storage unit,
The process distribution means of the database management device is:
A database management method characterized in that the process input from the process input device is assigned to a database in which the data block identifier is most frequently associated with reference to the correspondence relationship storage means. In the database device,
When acquiring N pieces of processing contents and acquiring the correspondence information associating the data block identifier and the database identifier referred to by the processing contents from the memory management means for managing the database,
The data block identifier corresponding to the processing content is not stored in the memory in the memory management unit, the data block identifier is acquired from the database, written to the memory, and transmitted to the database management device. Database management method.   The program for functioning a computer as each means which comprises the database management apparatus of Claim 1.   The program for functioning a computer as each means which comprises the database management apparatus of Claim 2.

Images