JPH08339323A - Distributed data management system for database management system - Google Patents

Abstract

PURPOSE: To provide a distributed data management system in a data base management system which attains the optimum arrangement of distributed data by shifting the distributed data having their access frequency higher than a prescribed level in a distributed data unit. CONSTITUTION: A server 8 of each of systems forming a distribution processing system is provided with a distributed data store part 12 which stores the distributed data and a counter part 14 which integrates the access frequency of each of distributed data stored in the part 12 for every computer system. Furthermore, a statistic information management part 18 to record the access frequency counted by the part 14 into a statistic information management table 16, a comparison part 22 which compares the counted access frequency with the highest access frequency that is defined at a highest access frequency definition part 20, and a rearrangement execution part 24 which automatically shifts the distributed data having their access frequency higher than the highest access frequency to the part 12 of another computer system, i.e., the access destination of the relevant distributed data are provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed data management system in a database management system, and more particularly to a distributed data management system for arranging distributed data at an optimum position in a distributed database in a distributed processing system.

[0002]

2. Description of the Related Art Conventionally, in a system constructed by distributing a plurality of database servers on a network, the data to be stored in each database server (hereinafter referred to as "distributed data") is optimized for data access. Efficient management and operation of the entire system such as speed, communication cost, etc. are achieved. Japanese Unexamined Patent Publication No. 2-122361 discloses a management method for transferring the entire database between database servers according to the access frequency from a workstation. This management method is LAN
In a distributed processing system in which computer systems including a plurality of clients that perform data processing on a (local area network) and a database server that comprehensively manages data common to each client, are accessible to each other, other computers A database frequently accessed by the system was transferred to the computer system.

Also, when there is an application from a department or person who manages the distributed data, or when there is a request from a user who frequently uses the distributed data, an administrator responds to the request. Data was moved by manual operation such as.

[0004]

However, in the past, when transferring distributed data, the entire database had to be transferred between database servers. Therefore, there is a problem in that the amount of transferred data becomes large and the resource needs to be used so much.

Further, when paying attention to each distributed data, there is a problem that the distributed data is not always arranged in the optimum database.

[0006] In addition, when moving the distributed data, a human must be always involved.

The present invention has been made to solve the above problems, and an object of the present invention is to move distributed data whose access frequency is a certain value or more to the distributed data unit, thereby It is to provide a distributed data management method in a database management system for optimizing arrangement.

[0008]

In order to achieve the above object, the present invention provides a distributed processing system in which a computer system including a server and a client is connected to form distributed data in each server. In a database management system that centrally manages a database to be stored, the number of accesses from any computer system is recorded for each distributed data, and the database is stored according to the result of comparison between the number of accesses and a preset maximum number of accesses. It is characterized in that the rearrangement is performed in units of distributed data.

Further, the server includes an access number accumulating means for accumulating the access number for each distributed data for each computer system, and a statistical information managing means for recording the access number accumulated by the access number accumulating means in a statistical information management table. And an access count maximum value storage unit that stores the maximum access count value for the distributed data, an access count comparison unit that compares the access count with the maximum access count value, and distributed data whose access count exceeds the maximum access count value. Relocation means for automatically moving to a database of another computer system that is the access destination.

Further, each of the servers accumulates the access count for each distributed data for each computer system, and the statistical information management for recording the access count accumulated by the access count integrating means in a statistical information management table. Means, an access count maximum value storage means for storing the maximum access count value for the distributed data, an access count comparison means for comparing the access count and the access count maximum value, and the access count for each distributed data is the maximum access count. And a rearrangement confirmation means for confirming the movement of the distributed data when the value is exceeded, and the optimized arrangement of the distributed data is performed according to the instruction.

Further, the rearrangement confirming means moves the target distributed data when an optimized arrangement instruction is received.

Further, the relocation confirmation means resets the number of times of access to the target distributed data and does not perform the optimized allocation of the distributed data when receiving the instruction not to perform the optimized allocation.

Further, the relocation confirming means does not perform the optimized placement of the distributed data by subtracting the number of times of access to the target distributed data when receiving the instruction not to perform the optimized placement.

Further, the access number maximum value storage means stores the access number maximum value for each of the servers.

Further, the access number maximum value storage means stores the access number maximum value for each of the statistical information management tables.

Further, the maximum access number storage means stores the maximum access number for each distributed data managed for each computer system.

Further, it is characterized in that all entries in which the number of accesses is recorded are prepared in advance in the statistical information management table.

Further, it is characterized in that an entry is generated in the statistical information management table when the access number accumulating means accumulates the access number for the first time.

Further, an index area for registering a key used for searching an entry recorded in the statistical information management table is provided.

The entries recorded in the statistical information management table are made up of all combinations of data identifiers for identifying distributed data and server identifiers for identifying the servers, of which only data identifiers are used as a key for indexing. It is characterized in that it is registered in the area.

The entries recorded in the statistical information management table are composed of all combinations of data identifiers for identifying distributed data and server identifiers for identifying the servers, and of these, only the data identifier is used as a key for indexing. It is characterized in that it is registered in the area.

Further, the access number comparison means is characterized in that the corresponding distributed data is sequentially searched from the head of all entries recorded in the statistical information management table.

[0023]

In the distributed data management method in the database management system according to the present invention having the above-mentioned configuration, when each server having a database accesses the distributed data stored in its own database, it is distributed to each computer system. The number of accesses for each data is recorded in the statistical information management table. The recorded access count and the maximum access count are compared, and in the rearrangement unit according to the present invention, the distributed data whose access count exceeds the maximum access count is stored in the database of another computer system as the access destination. Move it automatically.

In the rearrangement confirming means according to another invention, the distributed data is moved to a database of another computer system which is an access destination after confirming whether the distributed data can be moved.

When the distributed data is not moved, the number of accesses corresponding to the distributed data is reset,
By not counting up, whether or not the distributed data can be moved can be confirmed again at the time of subsequent access.

Further, the maximum value of access times is
It can be defined for each statistical information management table or for each distributed data.

The entries recorded in the statistical information management table are composed of all combinations of data identifiers and server identifiers. These entries are prepared in advance or when each entry is accessed. Can be generated.

The entry can be searched sequentially from the beginning or by registering a key in the index area.

[0029]

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

Example 1. FIG. 1 is a schematic configuration diagram of a distributed processing system adopting a first embodiment of a distributed data management system according to the present invention. The distributed processing system in this embodiment is realized by connecting the computer systems 2a, 2b, 2c with communication lines 4a, 4b. Of course, it is not limited to the three computer system as shown in FIG.

FIG. 2 is a diagram showing the basic configuration of each computer system in this embodiment. The computer system 2 in the present embodiment is a system having a client / server type database management system, and a database server computer (hereinafter referred to as “server”) 8 and a plurality of client computers (hereinafter referred to as “client”) 8 on the LAN 6. 10 are connected and built. Each computer system 2a, 2
The database management system in b and 2c can communicate via the communication lines 4a and 4b, and also functions as the entire distributed processing system. The server 8 includes a distributed data storage unit 12 as a database for storing distributed data,
A counter unit 14 as an access number accumulating unit for accumulating the access number for each distributed data stored in the distributed data storage unit 12 for each computer system, and the access number counted by the counter unit 14 are recorded in the statistical information management table 16. A statistical information management unit 18 as a statistical information management unit, an access number maximum value definition unit 20 as an access number maximum value storage unit that stores a maximum access number value for defined distributed data, and an access number and an access number maximum. A comparison unit 22 as an access count comparison unit for comparing the value with a value, and the distributed data whose access count exceeds the maximum access count are automatically moved to the distributed data storage unit 12 of another computer system that is the access destination. Relocation execution unit 24 as relocation means
A distributed data transfer unit 26 that mutually transfers distributed data with a server on another computer system; and a server configuration information definition unit 28 that defines configuration information of the servers included in the distributed processing system. Have. And FIG.
Each computer system 2a, 2b, 2
c is the LAN 6a, 6b, via the communication lines 4a, 4b,
6c are combined to construct a distributed processing system for a wide area network. Further, as described above, the database management system in each computer system 2a, 2b, 2c can mutually transfer the distributed data stored in each database. In the servers 8a, 8b, 8c shown in FIG. 1, only the main parts of the configuration shown in FIG. 2 are shown.

Since the server 8 is composed of a general computer, the distributed data storage unit 1 described above is used.
2. The statistical information management table 16, the maximum access count definition unit 20, and the server configuration information definition unit 28 are generated on a memory or in a disk device, and the counter unit 14, the statistical information management unit 18, the comparison unit 22, and the relocation execution are performed. The unit 24 and the distributed data transfer unit 26 are configured by an application loaded on the CPU and the memory and executed by the CPU.

FIG. 3 is a diagram showing an example of setting contents of the maximum access number value defining unit 20 provided in each computer system in this embodiment. As shown in this figure, the maximum access count values "MAX20A" and "MAX20" are assigned to the maximum access count definition parts 20a, 20b and 20c.
"B" and "MAX20C" are defined for each server one by one. This definition can be set by the system designer or system administrator at first depending on his experience, and can be changed by referring to the usage status thereafter. Is.

FIG. 4 is a diagram showing an example of setting contents of the server configuration information defining unit 28 provided in each computer system in this embodiment. The server configuration information is composed of a "server identifier" for identifying each server, and the server identifiers of the servers 8a, 8b, 8c in each computer system 2a, 2b, 2c are "SVR8".
FIG. 4 shows A "," SVR8B ", and" SVR8C ".

FIG. 5 is a diagram showing an example of setting contents of the distributed data storage unit 12 provided in each computer system in this embodiment. The "data identifier" is a unique number or the like given to each distributed data for each server, which is internally assigned and held by the data management system in each computer system 2a, 2b, 2c, and cannot be seen by the database user. . In the present embodiment, since it is assumed that there are salespeople at sales offices with branch offices nationwide, "salesman", "man number" and "address" are set for each data identifier in correspondence with each data identifier. ing.

FIG. 6 is a diagram showing the structure of the statistical information management table used in this embodiment. The statistical information management table 16 is internally created when the system is activated. The statistical information management table 16 includes the above-mentioned “data identifier”, “server identifier”, and “access count” for recording the access count for each entry specified by the combining condition of the data identifier and the server identifier.
In the “access count”, a value obtained by accumulating 1 in the counter unit 14 is recorded for each access such as search or update of the distributed data. Further, an index area 30 for registering an index key used for searching the entry generated by this join condition is provided. By setting the index key, the comparison unit 22 can search and access data at high speed even when the total entry amount of the statistical information management table 16 is large.

FIG. 7 is a diagram showing an example of setting contents of the statistical information management table provided in each computer system in this embodiment. In this embodiment, all combinations of data identifiers and server identifiers in each computer system 2a, 2b, 2c, that is, (data identifier x
An entry for (server identifier) is registered in advance. At this time, the number of accesses is initialized to "0".

A feature of this embodiment is that the distributed data whose access count exceeds the maximum access count is automatically moved to the database of another computer system which is the access destination. In this way, since it is possible to transfer in distributed data units based on statistical information such as the number of accesses, it is possible to provide an optimal distributed data arrangement and provide a distributed processing system with the most efficient and lowest communication cost. can do.

FIG. 8 is a flow chart showing the processing of the distributed data management system of this embodiment. The operation of this embodiment will be described below with reference to this figure.

In the configuration shown in FIG. 1, for example, when Mr. E who was using the computer system 2b is transferred to a sales office that uses the computer system 2a,
Mr. E uses one of the clients 10a, but Mr. E is supposed to access the distributed data of the computer system 2b from the client 10a-1.

In this case, Mr. E is the client 10a-1.
Then, for example, when the data regarding the user is accessed by designating the "man number", the database management system of the server 8a first searches for the distributed data storage unit 12a of the server 8a. In this embodiment, since Mr. E's data does not exist in the distributed data storage unit 12a, the server 8b, which is another computer system, is requested to retrieve the target data. In this way, the server 8 first checks the distributed data belonging to itself, and if the corresponding distributed data does not exist, requests the server registered in the server configuration information definition unit 28 to search the data in the order of registration. To do.

In FIG. 8, the server 8b that has received the request
Searches the distributed data storage unit 12b managed by itself. Here, if the corresponding distributed data exists, the access count corresponding to the connection condition between the data identifier of the corresponding distributed data and the server identifier of the server requesting the data access is incremented by 1 (step 101).
In the case of this example, since the distributed data of the data identifier ID012 regarding Mr. E who is the search target from the server 8a having the server identifier "SVR8A" exists in the server 8b,
The access count of the entry having the data identifier “ID012” and the server identifier “SVR8A” included in the statistical information management table 16b is incremented by one.

Next, the server 8 that manages the requested data
b checks whether or not the value of the counted access count exceeds the maximum access count defined in the maximum access count definition section 20b (step 102).
If not, the corresponding distributed data is sent to the server 8a which issued the data search request, and the processing is terminated.

If the maximum number of access times is exceeded, the server identifier of the server requesting the search is compared with the server identifier of the distributed data to be searched (step 103). If they match, it means that the server has access to its own distributed data, so the access counts of all entries relating to that server are reset (step 104). In this example, the server 8a that has made the search request is different from the server 8b in which the distributed data to be searched exists. Therefore, if they do not match as in this example, the statistical information management table 16b for the data identifier "ID012" All entries are automatically deleted (step 105). That is, in this example,
(Data identifier, server identifier) = (ID012, SV
R8A), (ID012, SVR8B), (ID01
2, SVR8C) will be deleted. FIG. 9B shows a state in which these entries are deleted from the statistical information management table 16b.

Subsequently, the distributed data is moved to the server having the highest access frequency to the distributed data, and an entry relating to the distributed data is added to the statistical information management table 16 of the server (step 106). That is, the distributed data regarding Mr. E registered in the distributed data storage unit 12b is moved to the distributed data storage unit 12a. This is processed in the same manner as when the distributed data is newly created, and the data identifier is a number uniquely assigned to each server 8. Therefore, in this example, the data identifier for Mr. E is "ID004". . The state of movement of this distributed data is shown in FIGS. In addition, the distributed data storage unit 1 of the server 8a
On the basis of the data identifier “ID004” newly registered in 2a, (statistic information management table 16a) (data identifier, server identifier) = (ID004, SVR8A),
(ID004, SVR8B), (ID004, SVR8
The entry C) is registered as shown in FIG.

Thereafter, in response to the data access request from Mr. E, the data is returned to the client 10a-1 to end the processing.

As described above, according to the present embodiment, the number of times data is accessed is recorded as statistical information, and the data is automatically moved according to the access status. It can be optimized automatically. Therefore, as shown in the above example, with the movement of the person who is entered as the distributed data, the data is moved without the intervention of the system administrator or the like.

Example 2. In the first embodiment, as shown in FIG. 8, when the maximum access count is exceeded, if the server identifier of the server requesting the search and the server identifier of the distributed data to be searched do not match, All the entries in the statistical information management table 16 relating to the data identifier are automatically deleted and the distributed data is automatically moved. The present embodiment is characterized in that this process can be performed manually. FIG. 10 is a diagram showing the basic configuration of each computer system in this embodiment. Instead of the maximum access number value defining unit 20 in the first embodiment shown in FIG. A relocation confirmation unit 32 is provided as relocation confirmation means for confirming the movement of the distributed data when the access count exceeds the maximum access count. Other configurations are the first
It is similar to the embodiment.

FIG. 11 is a flow chart showing the processing of the distributed data management system of this embodiment. The operation of this embodiment will be described with reference to this figure. Steps 101 to 106 in FIG. Since it is similar to the embodiment, the description will be omitted as appropriate.

Similarly to the example shown in the first embodiment, Mr. E uses the client 10a-1 to the computer system 2b.
If you decide to access the distributed data of, the number of access times of the distributed data exceeds the maximum number of access times,
In addition, when the server 8a that has made the search request is different from the server 8b in which the distributed data exists as in this example, the relocation confirmation unit 32 manages whether the corresponding distributed data may be moved to a server with a high access frequency. The person in charge (step 107). This inquiry is made by displaying it on a terminal device (not shown). The relocation confirmation unit 32
When the administrator receives an instruction to move the distributed data to perform the optimized allocation, the processes of steps 105 and 106 are performed to move the target distributed data. When the administrator gives an instruction not to perform the optimized placement, the process of step 104 is performed to suspend the optimized placement.

As described above, since it is possible to prevent the movement of the distributed data automatically, it is effective when it is desired to suspend the movement for a temporary reason. Then, the corresponding distributed data need not be moved until the number of times of access exceeds the maximum value again.

The automatic optimization arrangement shown in the first embodiment and the manual optimization arrangement shown in this embodiment may be switched to be usable.

Example 3. In the second embodiment, when the instruction not to perform the optimized placement is received, the optimized placement is suspended by resetting the access count of the corresponding entry included in the statistical information management table 16. The rearrangement confirming unit 32 in the present embodiment is characterized in that, when receiving the instruction not to perform the optimized arrangement, it reduces the access count of the corresponding entry by one.

FIG. 12 is a flow chart showing the processing of the distributed data management system of this embodiment, but the processing of step 108 is carried out when an instruction not to make an optimized arrangement is received. This is different from the processing shown in FIG.

That is, when it is determined in step 107 that the corresponding distributed data is not moved, that is, the optimized arrangement is not performed, the access count of the corresponding data is decremented by 1 (step 108).
Since the access count is incremented by 1 in step 101, the access count does not change in this series of processes.

Therefore, according to the present embodiment, when it is not possible to determine whether or not the distributed data can be moved, it is possible to hold the availability of the movement of the distributed data for one access and see the situation.

Example 4. In each of the above embodiments, the maximum access number definition section 20 is prepared for each server, but in the present embodiment, when a plurality of statistical information management tables are provided, the maximum access number definition section is provided. Is provided for each statistical information management table.
FIG. 13 is a diagram showing a statistical information management table in this embodiment. As shown in this figure, the maximum access count definition unit and the statistical information management table are combined, but in other words, the maximum access count column is provided in the statistical information management table and the maximum access count is set in that column. The value will be set. The present embodiment is characterized in that the maximum access number is set for each statistical information management table, so that the maximum access number is set to the same value.

In this embodiment, the same processing as in the first or second embodiment is carried out except that the maximum access count value corresponding to the distributed data is referenced when the maximum access count value is referenced. The description is omitted because it is performed.

Therefore, according to the present embodiment, for example, when the accounting information table other than sales is provided as the statistical information management table, the maximum access number can be defined for each department. .

Example 5. In the first to third embodiments, the access count maximum value defining unit 20 and the statistical information management table 16 are provided separately, but the access count maximum value defining unit in the present embodiment is stored in the statistical information management table. The feature is that the maximum number of access times can be defined for each registered data. Therefore, it has a structure in which the maximum number-of-accesses definition part and the statistical information management table are combined, and has the same structure as the fourth embodiment. The maximum number of access times defined in each statistical information management table is all the same value in the fourth embodiment, whereas the present embodiment is characterized in that a different maximum access number value is defined for each data. That is, the maximum access count can be stored for each distributed data managed for each computer system, or for each data specified by (data identifier × server identifier) in the first embodiment. As a result, the maximum access count can be defined for each data.

Example 6. In each of the above-described embodiments, all the entries included in the statistical information management table 16 are prepared in advance with the access count set to the initial value “0”.
The present embodiment is characterized in that an entry is generated in the statistical information management table 16 and the number of accesses is recorded when the counter unit 14 first integrates the number of accesses. That is, as shown in FIG. 14, before the data access, the statistical information management tables 16a, 16b, 16c.
No data has been entered in.

FIG. 15 is a flow chart showing the process of the distributed data management system of this embodiment. The process of this embodiment will be described with reference to this flow chart, which corresponds to the distributed data to be accessed first. It is checked whether or not the data is entered in the statistical information management table 16 (step 109). If there is no entry, a new entry is created at that time and the access count is set to 1 (step 110). Subsequent processing is the same as the first
The description is omitted because it is similar to the embodiment.

In this way, the time required for preprocessing when newly creating distributed data is shortened. Particularly in the present embodiment, when the accessed data is concentrated and specified, the extra entries are the statistical information management table 1
6 is not registered, the storage capacity can be reduced. In addition, since the total number of entries is small as compared with the case where all entries are prepared in advance, the search speed can be improved.

Example 7. In each of the above-described embodiments, the index area 30 is provided in the statistical information management table 16, and the index key used for searching the entry generated by the combination condition of the data identifier and the server identifier is registered. In the first embodiment, this index key is generated under the join condition of the data identifier and the server identifier, but if the number of distributed data is large and the number of servers in the distributed processing system is small, only the data identifier is used as the index key. May be registered in the index area 30 as

Example 8. In the seventh embodiment, the number of distributed data is large and the number of servers in the distributed processing system is small. On the contrary, when the number of servers in the distributed processing system is large and the number of distributed data is small, Only the server identifier may be registered in the index area 30 as an index key.

Example 9. In each of the above embodiments, the index area 30 is provided in the statistical information management table 16 to increase the speed when the total number of entries is large. However, the number of servers handled in the system and the number of distributed data are small, and as a result, the total number of entries is small. In this case, the index area may not be used. In this case, the comparison unit 22 sequentially searches the entries corresponding to the distributed data from the beginning of all the entries recorded in the statistical information management table 16.

As described above, in this embodiment, since the index area 30 is not provided and the search by the index key is not performed, the storage area can be reduced by the index area, and the entry in the statistical information management table 16 can be reduced. Addition becomes easier.

Although the features of each of the above-described embodiments have been described, the respective functions can be appropriately combined and used according to the system environment and the like.

[0069]

According to the invention described in claims 1 and 2, it is possible to optimize the arrangement of the distributed data by moving the distributed data based on the statistical information obtained by recording the number of accesses. Becomes Particularly, according to the invention described in claim 2, it is possible to automatically perform the above processing. As a result, when a person moves due to a personnel change or a move, the distributed data about the person also moves to the server of the computer system that is the move destination, so that no manpower or time is required to move the distributed data. In addition, it is possible to prevent an increase in communication cost due to a person moving.

On the other hand, according to the invention of claim 3,
It is possible to perform the above processing after the administrator or the like confirms it. That is, according to the inventions of claims 3 to 6,
It is possible to prevent the movement of distributed data automatically, so it is effective when you want to suspend the movement for temporary reasons, or when you want to confirm it with an administrator before moving it. Is. In particular, according to the invention described in claim 6, since it is possible to suspend only once, it is possible to inquire again whether the distributed data can be moved or not at the next access.

According to the invention described in claim 7, for each server,
According to the invention described in claim 8, when there are a plurality of statistical information management recording means, each is registered for each statistical information management recording means, and according to the invention described in claim 9, each registered in the statistical information management recording means. It is possible to define the maximum access count for each data. If the maximum number of access times is defined for each server, a small recording capacity is required. Further, if the maximum number of times of access is defined for each data, the data amount increases, but it can be uniquely defined for each data.
Therefore, an effective system can be provided by appropriately selecting any one according to the resources, operation, scale, etc. of the distributed processing system.

According to the tenth aspect of the present invention, the area for all entries can be reserved in advance.
It is possible to reserve the area without fail. In addition, since no new entry is generated during use, it is possible to improve the speed during use.

According to the eleventh aspect of the present invention, it is possible to reduce the time required for preprocessing when newly creating distributed data. Further, when the data to be accessed is concentrated and specified, the extra entry is not registered in the statistical information management table, so that the storage capacity can be reduced. In addition, since the total number of entries is small compared to the case where all the entries are prepared in advance, the search speed can be improved.

According to the twelfth aspect of the present invention, by providing an index area and setting an index key, it is possible to search and access data at high speed when the total amount of entries in the statistical information management table is large. Become. However, claim 13
When the number of distributed data handled by the entire distributed processing system is large and the number of servers in the distributed processing system is small as in the described invention, only the data identifier can be registered in the index area as an index key. On the contrary, when the number of servers in the distributed processing system is large and the number of distributed data is small as in the invention described in claim 14, only the server identifier can be registered in the index area as an index key.

According to the fifteenth aspect of the present invention, since the index area is not provided and the search is not performed by the index key, the storage area can be reduced by the index area, and the entry to the statistical information management table can be added more. It can be easily performed.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of a distributed processing system adopting a first embodiment of a distributed data management system according to the present invention.

FIG. 2 is a diagram showing a basic configuration of each computer system in the first embodiment.

FIG. 3 is a diagram showing an example of setting contents of a maximum access number value definition unit provided in each computer system in the first embodiment.

FIG. 4 is a diagram showing an example of setting contents of a server configuration information definition unit provided in each computer system in the first embodiment.

FIG. 5 is a diagram showing an example of setting contents of a distributed data storage unit provided in each computer system in the first embodiment.

FIG. 6 is a diagram showing a configuration of a statistical information management table used in the first embodiment.

FIG. 7 is a diagram showing an example of setting contents of a statistical information management table provided in each computer system in the first embodiment.

FIG. 8 is a flowchart showing the processing of the distributed data management system in the first embodiment.

FIG. 9 is a diagram used for explaining movement of distributed data in the first embodiment.

FIG. 10 is a schematic configuration diagram of a distributed processing system adopting a second embodiment of the distributed data management system according to the present invention.

FIG. 11 is a flowchart showing processing of a distributed data management system in the second embodiment.

FIG. 12 is a flowchart showing a process in a third embodiment of the distributed data management system according to the present invention.

FIG. 13 is a diagram showing a statistical information management table in a fourth embodiment of the distributed data management system according to the present invention.

FIG. 14 is a diagram showing a statistical information management table in a sixth embodiment of the distributed data management system according to the present invention.

FIG. 15 is a flowchart showing the processing in the sixth embodiment.

[Explanation of symbols]

2, 2a, 2b, 2c computer system, 4a,
4b communication line, 6, 6a, 6b, 6c LAN, 8,
8a, 8b, 8c database server computer,
10 client computer, 12 distributed data storage unit, 14 counter unit, 16 statistical information management table, 18 statistical information management unit, 20, 20a, 20b, 2
0c Access count maximum value definition part, 22 comparison part, 24
Relocation execution unit, 26 distributed data transfer unit, 28 server configuration information definition unit, 30 index area, 32 relocation confirmation unit.

Claims (15)

[Claims] 1. A database management system for centrally managing a database for storing distributed data, which is arranged in each server in a distributed processing system formed by combining computer systems including a server and a client, wherein: In the database management system, the number of accesses from the database is recorded for each distributed data, and relocation to the database is performed in distributed data units in accordance with the result of comparison between the number of accesses and a preset maximum value of the number of accesses. Distributed data management method. 2. The distributed data management method for a database management system according to claim 1, wherein the server integrates an access count integrating means for integrating the access count for each distributed data for each computer system, and the access count integrating means. Statistical information management means for recording the number of times of access made in the statistical information management table, maximum access frequency storage means for storing the maximum access frequency for distributed data, and access frequency comparison means for comparing the access frequency with the maximum access frequency And a relocation means for automatically moving distributed data whose access count exceeds the maximum access count to a database of another computer system that is the access destination, and distributed data in a database management system. Management method. 3. The distributed data management method for a database management system according to claim 1, wherein each server comprises an access count integration means for integrating the access count for each distributed data for each computer system, and the access count integration means. A statistical information management unit that records the accumulated number of accesses in a statistical information management table, an access number maximum value storage unit that stores the maximum access number value for distributed data, and an access number comparison that compares the access number and the maximum access number. Means and relocation confirmation means for confirming the movement of the distributed data when the number of times of access to each distributed data exceeds the maximum number of times of access, and the optimized arrangement of the distributed data according to the instruction. Distributed data management system for database management system Method. 4. The distributed data management method in the database management system according to claim 3, wherein the relocation confirmation unit moves the target distributed data when an optimized placement instruction is received. Distributed data management method in database management system. 5. The distributed data management method for a database management system according to claim 3, wherein the relocation confirming unit resets the number of times of access to the target distributed data when receiving an instruction not to perform optimized allocation. A distributed data management method for a database management system, which is characterized by not performing optimized allocation of distributed data. 6. The distributed data management method for a database management system according to claim 3, wherein the relocation confirmation unit subtracts the number of accesses to the target distributed data when receiving an instruction not to perform optimized allocation. A distributed data management method for a database management system, which is characterized by not performing optimized allocation of distributed data. 7. The distributed data management method for a database management system according to claim 2, wherein the maximum access count storage means stores the maximum access count for each server. Distributed data management method. 8. The distributed data management method in the database management system according to claim 2, wherein the maximum access count storage means stores the maximum access count for each statistical information management table. Distributed data management method in database management system. 9. The distributed data management method for a database management system according to claim 2, wherein the maximum access count storage means stores the maximum access count for each distributed data managed for each computer system. A distributed data management method for database management systems. 10. The distributed data management method for a database management system according to claim 2, wherein all entries in which the number of accesses is recorded are prepared in advance in the statistical information management table. Data management method. 11. The distributed data management method for a database management system according to claim 2, wherein an entry is created in the statistical information management table when the access count integration means integrates the access count for the first time. Distributed data management method for database management system. 12. The distributed data management method for a database management system according to claim 2, wherein an index area for registering a key used for searching an entry recorded in the statistical information management table is provided. Distributed data management method for database management system. 13. The distributed data management method for a database management system according to claim 12, wherein the entries recorded in the statistical information management table are a data identifier for identifying distributed data and a server identifier for identifying the server. A distributed data management method in a database management system, which is composed of all combinations, of which only the data identifier is registered in the index area as a key. 14. The distributed data management method for a database management system according to claim 12, wherein the entries recorded in the statistical information management table are a data identifier for identifying distributed data and a server identifier for identifying the server. A distributed data management method in a database management system, which is composed of all combinations, of which only the data identifier is registered in the index area as a key. 15. The distributed data management method for a database management system according to claim 2, wherein the access count comparison means sequentially applies the corresponding distributed data from the head of all entries recorded in the statistical information management table. A distributed data management method in a database management system characterized by searching.