JPH11250093A - Database transfer method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To smoothly transfer from an old database system to a shared database system. SOLUTION: This method transfers from an old database system to a shared database system through an intermediate database system 2. In the system 2, database S1' to S7' by target store separate data that are used only for the system and shared data which are also available to other systems. When data is updated in one of the databases by target, updated data is made a file and it is sent to a shared database DB and reflected on the database 10. In such a case, content of the database 10 undergoes deferred updating. After that, extraction processing is performed according to a rule defined in each of the databases S1' to S7' by target and the extraction results are transferred in a file to the other databases by target.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database migration method for migrating a system constituted by a plurality of individually operated databases to a shared database system.

[0002]

2. Description of the Related Art In an information society in recent years, various information processing apparatuses are used to support corporate activities.
For example, in a large company, human resource management is often performed by an individual database such as a personnel system, a salary system, or a company housing system.

In such a system, since processing is performed based on mutually related data, when data is updated by one application, the updated data is transferred to another application in the form of a file. Needs may arise. For example, suppose that an employee is transferred from one establishment to another due to promotion, and information on this transfer is first input to the personnel system. In this case, since the salary increases with the promotion, it is necessary to pass this information to the salary system, and since the place of work is changed, it becomes necessary to pass information about the company home to the company home system. Therefore, in the personnel system, a file is created by converting the data format so that the data relating to the employee's transfer can be read by the partner system. And
The other party's system reads the magnetic disk storing the file, or transfers the file via a communication network.

Here, if N systems are operating in relation to each other, the number of conversion applications of the entire system is N (N-1), and the number of systems constituting the entire system increases. Then, the number of required conversion applications increases dramatically. For this reason, there is a problem that the man-hour for developing the conversion application, the man-hour required for the conversion process, or the computer resources for storing the conversion application are consumed. In addition, when data is updated in a certain system, it takes time until the updated data is reflected in another system, so that information based on the latest data cannot be provided. .

Accordingly, a shared database system has been developed in which each system shares common data. The shared database system includes, for example, individual systems such as a personnel system, a salary system, and a company home system, an original database that stores data from each individual system in an integrated manner, and a LAN (Local Area Network) that connects each individual system and the original database. ).

In the original database, common data common to the individual systems is stored in the form of various tables, and individual data used only in the system is stored in each individual system in the form of a table. ing.

[0007]

By the way, such a shared database system is usually constructed as a distributed system in which individual systems are arranged at different locations. For this reason, it is necessary to ensure the consistency of each distributed individual system. As a method for guaranteeing consistency, a control method called two-phase commit control has been proposed. In this method, it is first asked whether each individual system can be updated or not. Done. However, in the two-phase commit control, the transaction amount is large. For example, if the processing is concentrated at a specific time such as 10:00 on Monday,
There is a problem that processing takes a long time. In addition, when a shared database system is used, it is necessary to discard the conventional individual system and to recreate each individual system, resulting in a large burden on development man-hours and development time.

The present invention has been made in view of the above circumstances, and provides a database migration method for smoothly moving from a current system to a shared database.

[0009]

According to the first aspect of the present invention, there is provided an information processing apparatus, comprising: a plurality of database apparatuses for handling mutually related common data; A first database system for transferring data between each other, each of the individual database devices storing at least individual data independent of other devices, and each of the individual database devices being connected to each other;
A database migration method that includes a shared database device that stores at least common data related to another device and migrates the shared data to a shared database system that updates the common data in real time. A plurality of purpose-specific database devices for transferring update data to the shared database device, wherein the shared database device reflects the transferred update data in its stored content and a rule determined for each purpose-specific database device. A second database system for transferring the extracted data extracted according to the above-described method to the plurality of purpose-specific database devices, respectively, is provided in an intermediate stage of shifting from the first database system to the shared database system.

[0010] In the invention according to claim 2,
When migrating from the second database system to the shared database system, the migrating measures are executed after the two-phase commit control can be performed without hindrance in all the purpose-specific database devices.

Further, in the invention according to claim 3,
A first database system that includes a plurality of database devices that handle mutually related common data and transfers the common data among the plurality of database devices from each other; each individual device that stores at least individual data independent of other devices; A database migration method for migrating to a shared database system that includes a database device and a shared database device that is connected to each of the individual database devices and stores at least common data related to other devices, and that updates the common data in real time. The first database system, the shared database device is provided, and a part of the plurality of database devices having mutually overlapping ranges of common data is changed to the individual database device corresponding to real-time update. A first step of transferring the other database apparatus to a second database system changed to a purpose-specific database apparatus that performs a delayed update by performing file transfer with the shared database apparatus; and And a second step of migrating from a database system to the shared database system.

Further, in the invention according to claim 4,
A first database system that includes a plurality of database devices that handle mutually related common data and transfers the common data among the plurality of database devices from each other; each individual device that stores at least individual data independent of other devices; A database migration method for migrating to a shared database system that includes a database device and a shared database device that is connected to each of the individual database devices and stores at least common data related to other devices, and that updates the common data in real time. Changing the first database system from the first database system to a purpose-specific database device that includes the shared database device and transfers updated data to the shared database device when data is updated; A second database system for reflecting the transferred update data in its stored contents, and then transferring extracted data extracted according to rules defined for each purpose-specific database device to each of the plurality of purpose-specific database devices. A first step of shifting to the individual database corresponding to a real-time update of a part of the plurality of purpose-specific database devices having a common data range overlapping each other from the second database system. The method is characterized by comprising a second step of shifting to a third database system changed to an apparatus, and a third step of shifting from the third database system to the shared database system.

Further, in the invention according to claim 5,
The second step is characterized in that a transition measure is executed after the two-phase commit control can be performed without hindrance in the purpose-specific database device to be changed.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a database migration method according to an embodiment of the present invention will be described with reference to the drawings.

A. Assumptions As described above, when implementing a distributed shared database system, it is necessary to consider the processing time involved in two-phase commit control. In this case, the amount of update data and the data related to the update data differ depending on the type of the individual system that performs the update, so that the processing time required for the two-phase commit control differs for each individual system.
On the other hand, if the communication speed and the processing speed of the original database are high, the processing time of the two-phase commit control is short, and even if the update processing is concentrated, the processing can be performed in a short time, the shared database system It is possible to shift to.

Here, the migration method includes a first mode in which all the individual systems are migrated to the shared database system, and a method in which some individual systems (new individual systems) are migrated to the shared database system and other individual systems are migrated to the shared database system. The system (old individual system) can be considered as a second mode in which data is accumulated in the individual system, only the changed part is filed, and exchanged with the original database.

By the way, since the data of each individual system is related to each other, the second aspect has the following problems and has few advantages. That is, in the old individual system, the update data is filed and transferred to the original database, so that the new individual system cannot see the data from the old individual system in real time. Also, the old individual system receives data updated in the new individual system by file transfer, and therefore cannot see the data of the new individual system in real time. For this reason,
It is desirable to migrate to a shared database system in an environment in which almost individual systems can update the original database in real time.

However, if the sharing of common data is postponed until such an environment is realized, in the meantime, as described above, it is necessary to perform file transfer between the individual systems, and When the specification changes in any of the systems, it is necessary to change all the related individual systems, and there arises a problem that the processing is inflexible.

In consideration of the future potential and flexibility of the system, the present inventors temporarily change the data from the old database system, which mutually transfers files between the individual systems, to the shared database. Was migrated to an intermediate database system that performs delayed updates, and the intermediate database system was migrated to a distributed shared database system.

B. First Embodiment Hereinafter, a migration method according to a first embodiment will be described with reference to the drawings. 1. Migration from old database system to intermediate database system 1-1: Old database system First, the configuration of the old database system will be described. FIG.
FIG. 1 is a conceptual diagram showing the entire configuration of an old database system. The old database system 1 is composed of databases S1 to S7 as shown in the figure. The personnel system S1 manages personnel changes, attendance, evaluation, family composition, etc. of the employee, and the salary system S2 manages the employee's employee. The salary management, the pension system manages the payment of pensions to retired employees, the company housing system manages the occupancy status of company housing, the retirement allowance system S
5 manages the retirement allowance of each employee, the logistics system S6 manages the physical distribution, and the accounting system S7 manages the accounting of the entire company. These databases S1 to S7 are configured as purpose-specific databases, and one database stores individual data used only for the database and common data used for another database in a mixed manner. Was.

In this case, since the individual data is used only for the database concerned, even if the change occurs, it does not affect other databases. However, if the change occurs to the common data, it affects other databases. The generated common data had to be filed and passed to other related databases.

1-2: Intermediate Database System Next, the intermediate database system will be described. FIG. 2 is a conceptual diagram showing the entire configuration of the intermediate database system.

In the figure, the intermediate database system 2
Are purpose-specific databases S1 'to S7' arranged at different locations, a shared database DB for storing standardized data, and each purpose-specific database S1 '
To S7 'and the local area network LAN connecting the shared database DB.

The purpose-specific databases S1 'to S7' correspond to the above-described databases S1 to S7, and include a conversion application for converting standard data into a record format used for the purpose-specific databases S1 'to S7'. Except for this point, it is the same as the databases S1 to S7 of the old database system 1. When the conversion application receives the transfer file from the shared database DB, the conversion application converts the data stored therein into each purpose-specific database S.
Data rearrangement processing is performed so as to match the record format of 1 'to S7'. In this case, since the conversion application operates only with the shared database DB, a very small program (about several kilobytes) is sufficient.

Incidentally, the databases S1 to S7 constituting the old database system 1 are usually constructed according to business needs, so that a certain system is UN
IX, other systems operate under different operating systems, such as Windows,
Generally, the record formats handled by them are not unified. Therefore, in order to store data in a unified format in the shared database DB, it is necessary to reconstruct the databases S1 to S7 according to unified rules. However, it takes enormous man-hours to recreate each database from the beginning. And costs. However, by using the above-described conversion application, it is possible to migrate from the old database system 1 to the intermediate database system 2 with almost no change in the databases S1 to S7.

The purpose-specific databases S1 'to S7' are closely related to each other. For example, personnel system S
The employee's attendance data stored in 1 'is also required in the salary system S2' to reflect the employee's salary. Also, if an employee wants a company house,
Since it is necessary to arrange a company house in consideration of the family structure, family data is required in the company house system S4 '. Therefore, if the data stored in each of the purpose-specific databases S1 'to S7' is updated, the other purpose-specific databases S1 'to S7' are updated.
By transferring the update data common to S7 'in a file format, the shared database DB temporarily aggregates the update data, and the files used for the update are transferred from the shared database DB to the purpose-specific databases S1' to S7 '. Has become. In this case, the purpose-specific database S1 '
The file transfer from .about.S7 'to the shared database DB is performed at regular intervals, such as every day. For this reason, the data is updated not in real time but in a delayed manner.

Next, the main part of the shared database DB is:
It comprises an original database 10 that stores common data, and a database management system DBMS that performs reflection extraction processing on the original database 10. by the way,
The data type (for example, the number of characters of one data) of each purpose-specific database S1 'to S7' depends on the application. For this reason, the original database 10 stores standard data obtained by standardizing the transfer files from the individual databases S1 to S7. When transferring data to the purpose-specific databases S1 'to S7', a process of extracting required standard data from the original database 10 is performed. For this reason, the original database 1
The data structures of 0 are grouped so that the extraction process can be executed efficiently. Specifically, it is composed of various tables such as employee, educational background or affiliation,
You can search by specifying the table name.

The purpose-specific databases S1 'to S7'
The data required for each is different. For this reason, the database management system DBMS stores the database S for each purpose.
The extraction process is executed according to the extraction conditions (rules) defined for each of 1 'to S7'. This extraction process
This is performed by extracting key information, extracting a table using the key information, and combining the extracted tables. The key information is extracted by providing a processing module for each function such as extraction timing, availability of extraction timing, and specification of related data, and is performed by giving an extraction parameter to the processing module. In this example, an employee code is used as key information, and a corresponding table is extracted from the employee code and is combined. In this case, the database management system DBMS stores what processing modules and extraction parameters are used for each purpose-specific database S1 'to S7', and performs extraction processing by referring to these. The extracted standard data is filed and transferred to the purpose-specific databases S1 'to S7'. Thus, when there is a change in the common data in a certain purpose database, the change result can be reflected in another purpose database.

As described above, the intermediate database system 2
Performs the delayed update, but individually manages the individual data and the common data in the original database 10, so that the file transfer between the individual databases S1 to S7 is mutually performed like the old database system 1. No more hassle. Further, at least when the purpose-specific databases S1 'to S7' receive files transferred from the shared database DB, the data stored in all the purpose-specific databases S1 'to S7' are unified. Therefore, the purpose-specific databases S1 'to S7'
It is possible to guarantee the uniformity of the processing result. Further, when the specification is changed in any of the purpose-specific databases S1 'to S7', it is sufficient to simply review the relationship between the individual database and the shared database DB. The man-hour and time required to create a program can be greatly increased.

2. Transition from intermediate database system to shared database system From intermediate database system 2 to shared database system, as described above, the time required for two-phase commit control is reduced, and there is no problem even if the concentration of reflection processing is considered. Perform at such times.

FIG. 3 is a conceptual diagram showing the entire configuration of the shared database system. In the figure, each individual database S1 ″ to S7 ″ of the shared database system 3 is
In storing the individual data used only in the database, the purpose-specific database S1 ′ of the intermediate database system 2 that stores not only the individual data but also the common data.
~ S7 '. For example, if the current salary of the employee managed by the salary system S4 ″ is not required by another system, the data on the current salary corresponds to the individual data.

Here, the individual data and the common data are composed of various tables, and are related by their items. For example, data indicating the current salary amount of the employee described above is given in a salary table shown in FIG.
This is stored as individual data in the payroll system S4 ″, while the affiliation table (FIG.
(B)) is the original data base 10 as common data.
Shall be stored in In this case, the employee code is an item common to both tables, and thereby, both tables can be associated with each other.

Next, each of the individual databases S1 '' to S7 ''
Has a determination mechanism for determining whether the updated data is individual data or common data, and the reflection processing is performed using the determination mechanism. First, when update data is input to each of the individual databases S1 ″ to S7 ″, the determination mechanism determines whether the update data is individual data or common data.

When the update data is individual data,
Data is updated in each of the individual databases S1 "to S7" itself. On the other hand, when the update data is common data, the update data is transmitted to the shared database DB via the LAN.
Send to In the shared database DB, the update data is reflected on the storage contents of the original database 10 using the database management system DBMS.

As described above, the shared database system 3
Then, since the individual data and the common data are stored in a distributed manner, the individual databases S1 ″ to S7 ″ can immediately perform the reflection process by transmitting only the common data to the shared database DB. Therefore, according to this system, each purpose-specific database S1 'is converted from the original database 10 like the intermediate database system 2.
It is not necessary to transfer a file to S7 ', the amount of communication related to data update can be reduced, and data can be updated in real time.

Each of the individual databases S1 '' to S7 ''
In the extraction process, a search is performed based on the key information corresponding to the input command. For example, in the salary system S4 '', when an employee code is input as a search key and a command for outputting the current salary amount and affiliation of the employee is given, the determination mechanism of the salary system S4 '' detects the command and Recognize that it is necessary to perform extraction processing from individual data and common data. Then, the current salary amount is acquired by referring to the salary table shown in FIG. 4A using the employee code as key information, and an instruction is sent to the shared database DB to output the affiliation corresponding to the employee code. . When the shared database DB receives the instruction, the database management system DBMS searches the affiliation table shown in FIG. 4B based on the employee code, extracts the affiliation corresponding to the employee code, and sends it to the payroll system S4. Return to ''. After this, the payroll system S
4 '' outputs the current salary amount and affiliation as an extraction result.
As a result, the user can obtain a necessary extraction result without worrying about which location on the network stores the data to be searched.

As described above, according to the database migration method according to the first embodiment, the old database system 1 is migrated to the intermediate database system 2 as the first stage, and the shared database system is migrated from the intermediate database system 2 as the second stage. 3, the transition can be made smoothly according to the development speed of the communication network and the computer. That is, in the first stage, the migration can be performed without making significant changes to the purpose-specific databases S1 'to S7' of the old database system 1, so that the man-hours and time required for program development can be reduced, and the original database can be reduced. In 10, the common data can be managed centrally. Further, in the second stage, the speed of the communication network or the computer shifts at a timing at which the processing does not hinder even if the two-phase commit control is performed, so that the advantages of the real-time update can be sufficiently exhibited.

C. Second Embodiment In the above-described embodiment, the transition of the second stage is executed when the condition of the two-phase commit control is satisfied in all the purpose-specific databases. This is because the data of each purpose database is related to each other, so even if only some purpose databases are migrated to individual databases that perform real-time updating, if other purpose databases delay update, other This is based on the reason that the update data cannot be viewed in real time in the purpose-specific database, and that common data from other purpose-specific databases cannot be viewed in real time in some individual databases.

Since the range of the common data is determined in accordance with the specifications of the purpose-specific databases S1 'to S7', for example, the data of the purpose-specific databases S1 'to S7' are as shown in FIG. It can be a relationship. In this case, the physical distribution system S6 'has common data Da with the company housing system S4', but has another purpose-specific database S
There is no common data between 1 'to S5' and S7 '. Therefore, in the company house system S4 'and the physical distribution system S6', when these are shifted to real-time updating when the conditions of the two-phase commit control are satisfied, the company house system S4 '
At 4 ', the common data Da updated in real time by the physical distribution system S6' can be viewed in real time.

Therefore, in the second embodiment, after the first stage of the first embodiment, a migration method for shifting some purpose-specific databases to real-time update, and thereafter shifting all the purpose-specific databases to real-time update. Is provided. That is, in the first stage, the old database system 1 is migrated to the first intermediate database system (the intermediate data system 2 of the first embodiment),
In the second stage, the process is shifted from the first intermediate database system 2 to the second intermediate database system 2 'including a partial real-time update.
From the intermediate database system 2 ′ to the shared database system 3. The first stage is the same as in the first embodiment, and the third stage executes the migration when all the individual databases satisfy the conditions of the two-phase commit control, as in the second stage of the first embodiment. Therefore, the description is omitted here, and the second stage (migration from the first intermediate database system 2 to the second intermediate database system 2 ') will be described.

1. Conditions for migration The conditions for executing the second stage are as follows: among the purpose-specific databases S1 'to S7' constituting the first intermediate database system 2 (see FIG. 2), the purpose-specific databases having mutually overlapping ranges of common data. As a set of S1 'to S7', migration is performed when the condition of the two-phase commit control is satisfied for each set. For example, in the example shown in FIG. 5, the physical distribution system S6 ′ and the company housing system S4 ′ are the first set, and
HR system S1 ', salary system S2', pension system S3 ', retirement allowance system S5', distribution system S6 ',
And a payroll system S7 'are a second set.

Here, the second set of personnel systems S1 'is a database serving as a backbone of the first intermediate database system 2, and has a large amount of common data, and accesses the common data from another purpose-specific database. The frequency is high and the processing associated with the two-phase commit control requires a long time. On the other hand, in the first set of distribution systems S6 ', the amount of common data is small and the processing time associated with the two-phase commit control is short. I do. In this case, when the speed of the communication network or the computer gradually increases, there is a time when the first set of transition conditions is satisfied, but the second set of transition conditions is not satisfied. At that timing, only the first set is shifted to real-time update, and the second intermediate database system 2 'is constructed.

2. Second intermediate database system Next, the configuration of the second intermediate database system 2 'will be described with reference to FIG. In this example,
It is assumed that the physical distribution system S6 'and the company home system S4' have been shifted to real-time updating.

In the second intermediate database system 2 'shown in FIG. 6, the physical distribution system S6''and the company home system S
4 ″ stores the individual data by itself, and the original database 10 stores common data of the physical distribution system S6 ″ and the company housing system S4 ″ and other purpose-specific databases S1 ′ to S4.
3 ', S5' and S7 'are stored.

In this case, the purpose-specific databases S1'-S
The reflection processing and extraction processing of 3 ′, S5 ′, and S7 ′ are the first processing.
This is performed in the format of the file transfer described in the embodiment. On the other hand, in the reflection processing of the distribution system S6 ″ and the company home system S4 ″, if the update data is individual data, the update data is stored in the individual database itself. On the other hand, if the update data is common data, the update data is sent to the shared database DB, and the storage contents of the original database 10 are updated in real time.

As described above, in the second embodiment, the first
After the transfer from the intermediate database system 2 to the second intermediate database system 2 ′, the data is transferred to the shared database system 3, so that real-time updating can be realized at an early stage.

In the above-described second embodiment,
Although the first intermediate database system 2 has been migrated to the second intermediate database system 2 ′, the old database system is directly migrated to the second intermediate database system 2 ′, and then migrated to the shared database system 3. You may.

The individual databases S1 ″ to S1
S7 '' stores only individual data,
The common data required in the individual database may be stored together with the individual data. In this case, for example, by storing individual data and common data for frequently used searches, the frequency of access to the original database 10 can be reduced. It becomes possible to operate.

[0049]

As described above, according to the invention specification items of the present invention, the transition to the shared database system is performed via the database system that is delayed and updated by the file transfer, so that the transition can be performed easily and smoothly. In addition, it is possible to guarantee the consistency of common data.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing the entire configuration of an old database system.

FIG. 2 is a functional block diagram illustrating a functional configuration of a shared database according to the embodiment;

FIG. 3 is a diagram showing an example of a table generated by the personnel system according to the embodiment;

FIG. 4A shows an example of a salary table according to the embodiment, and FIG. 4B shows an example of an affiliation table according to the embodiment.

FIG. 5 is a conceptual diagram showing a relationship between data stored in respective purpose databases according to the embodiment.

FIG. 6 is a block diagram showing a configuration of a second intermediate database system according to the embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Old database system (1st database system) 2 ... 1st intermediate database system 2 '... 2nd intermediate database system 3 ... Shared database system S1-S7 ... Database (database apparatus) S1'-S7' ... Purpose Separate database (purpose-specific database device) S1 "to S7" ... individual database (individual database device) DB ... shared database (shared database device) LAN ... local area network (communication network)

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Motoichi Katsuya 1-9-1, Konan, Minato-ku, Tokyo NTT Communicationware Co., Ltd. (72) Inventor Fujio Nakano Konan, Minato-ku, Tokyo 1-9-1-1 NTT Communicationware Co., Ltd. (72) Inventor Jun Tokunaga 1-9-1 Konan, Minato-ku, Tokyo NTT Communicationware Co., Ltd.

Claims (5)

[Claims] 1. A first database system comprising a plurality of database devices for handling mutually related common data and transferring the common data among the plurality of database devices, separate individual data independent from other devices. Transfer to a shared database system that includes at least each individual database device that stores and a shared database device that is connected to each individual database device and that stores at least common data related to another device and that updates the common data in real time A database migration method, comprising: the shared database device and a plurality of purpose-specific database devices that transfer update data to the shared database device when data is updated, wherein the shared database device stores the transferred update data. Reflected in that memory And a second database system that transfers extracted data extracted according to rules defined for each purpose-specific database device to the plurality of purpose-specific database devices, respectively, from the first database system to the shared database system. A database migration method provided at an intermediate stage. 2. When migrating from the second database system to the shared database system, the migrating measure is executed after two-phase commit control can be performed without hindrance in all of the purpose-specific database devices. The database migration method according to claim 1, wherein: 3. A first database system comprising a plurality of database devices for handling mutually related common data and transferring said common data between said plurality of database devices, separate individual data independent from other devices. Transfer to a shared database system that includes at least each individual database device that stores and a shared database device that is connected to each individual database device and that stores at least common data related to another device and that updates the common data in real time A database migration method, comprising the shared database device from the first database system, and a portion of the plurality of database devices, each of which has a common data range overlapping each other, corresponding to real-time updating. Individual data A first database system which is changed to a second database system in which the other database system is changed to a purpose-specific database system which performs a delayed update by performing file transfer with the shared database system; and And a second step of migrating from the second database system to the shared database system. 4. A first database system comprising a plurality of database devices for handling mutually related common data and transferring said common data among said plurality of database devices, separate individual data independent from other devices. Transfer to a shared database system that includes at least each individual database device that stores and a shared database device that is connected to each individual database device and that stores at least common data related to another device and that updates the common data in real time A database migration method, wherein the first database system is provided with the shared database device, and the database device is changed to a purpose-specific database device that transfers updated data to the shared database device when data is updated, Common use A database device that reflects the transferred update data in its stored content, and then transfers extracted data extracted according to rules defined for each purpose-specific database device to each of the plurality of purpose-specific database devices. A first step of migrating to a database system, and from the second database system, a part of the plurality of purpose-based database devices, each of which has a common data range overlapping with each other, corresponds to a real-time update. A database migration method, comprising: a second step of migrating to a third database system changed to an individual database device; and a third step of migrating from the third database system to the shared database system. . 5. The method according to claim 4, wherein the second step executes a transition measure after the two-phase commit control can be performed without hindrance in the target database device to be changed. Database migration method described in.