JP2748407B2 - Database processing system - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a database processing system for storing information, and more particularly to a database processing system suitable for speeding up a search process.

[Conventional technology]

Conventionally, as a method in which a plurality of tasks search in a database system and another task updates, for example,
As disclosed in Japanese Patent Application Laid-Open No. 61-88349, there is known a method of distributing the update processing load by giving priority to an update task and performing processing in units of tracks. According to this method, it is possible to reduce the waiting time of the search task.

[Problems to be solved by the invention]

The prior art described above does not consider the point of speeding up the search performance, and has a problem that the search task cannot be processed while the update task is processing.

The present invention has been made in view of the above circumstances, and an object of the present invention is to solve the above-described problems in the conventional technology, eliminate the conflict between database resources due to search processing and update processing, and speed up search processing. It is to provide a database processing system aiming at.

[Means for solving the problem]

An object of the present invention is to provide a database system which is separated into a search system database and an update system database or update data history information. Alternatively, processing is performed using updated data history information and a search database, and in response to a search request for a database that does not require up-to-dateness, the search request is processed using the search database. Is achieved by a database processing system.

[Action]

In the database processing system according to the present invention,
A search request that does not require recentness of data is executed by the search database system that does not perform any data resource sharing control processing and exclusive control processing, so that the performance of the search processing is improved as compared with a general database system. be able to.

〔Example〕

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

FIG. 1 is a diagram showing a configuration example of a database (hereinafter, referred to as “DB”) system according to an embodiment of the present invention.
In FIG. 1, a DB system (1) 40 and a DB system (2) 50 are independent DB systems, and the DB system (1) 40 transmits an update (including reference) request to an update DB.
(A) 60 is accessed and processed, and the update information is stored in the update data history information 70. The DB system (2) 50 accesses and processes only the reference request using the search DB (B) 80. The search DB (B) 80 is the update DB (A) 60
DB that was copied at a certain point in time.

The update DB (A) 60 may be omitted. In this case, access and processing are performed using both the update data history information 70 and the search DB (B) 80.

The user specifies the access language shown in FIG. 2 according to the processing purpose. The language analysis unit 10 analyzes the designated DB access language and passes it to the DB system management unit 20. The DB system management unit 20 determines a DB system to be processed based on the analysis information, and performs a target DB process by requesting the DB system to execute the DB system.

The analysis information is a request that does not require recency, that is,
It is determined whether the request requires high speed, and when high speed is required, the system load management unit 30 is called. System load management unit
30 manages the load status of each DB system and determines a system with a low system load.

If the analysis information is in an update language, a DB update process is performed by requesting the DB system (1) 40 having the update DB (A) 60 to execute the DB.

If the analysis information is a request to reflect the update data, a request to replace the update data history information 70 with a new storage area is issued to the DB system (1) 40, and the update data history information 70 is turned off. Let go. A DB update request is input to the DB system (2) 50 with the update data history information 70 as an input, and the search system DB (B) 80 is updated.

FIG. 3 is a table for managing the state of the entire system. The system name 31 and the DB name 32 are used to determine a DB system to be processed. Copy source DB name 33 is an update
This is a DB that is a duplicate of the DB, and indicates the name of the update DB at the copy source, and is used for searching for a DB with a low system load. The number of transactions executed in DB units (X) 34 indicates the number of DB access languages being processed for the DB, and the allowable number of transactions in the system (Y) 35 is the maximum DB access that can be processed concurrently in the DB system. Represents the number of language processes.
From these, the value of the load factor (Z) 36 is obtained by the following equation.

Z = X / Y The load factor (Z) 36 is used to determine a DB system with a low system load. The update data reflection flag 37
It is used to prohibit search processing when a search request is made while updating the search system DB (B) 80. Update DB omission flag 38
Is a flag expressing the presence or absence of the update DB.

FIG. 4 is a flowchart showing the operation of the DB system management unit 20 of the present embodiment configured as described above. Hereinafter, an operation example will be described in detail in accordance with this.

First, in step 200, the DB access language is "A"
If it is, the corresponding
The number of execution transactions in the database unit of the DB is increased by 1 (step 201), and in step 202, the update DB
(A) If 60 is not omitted, a search process is instructed to the update DB system (1) 40 (step 203).
If the B (A) 60 is omitted, a search process is commanded to the update data history information 70 (step 204).

If the data does not exist in step 205, a search process is instructed to the search DB system (2) 50 (step 206), and after the search process is completed, the number of database unit execution transactions of the DB in the system management table is set to one.
(Step 207), and the process ends.

Next, in step 210, the DB access language is "B"
In the case of, the processing is passed to the above-mentioned system load management unit 30, and a copy DB name having the minimum load ratio is obtained from the designated DB name (step 211), and the number of database unit execution transactions of the DB in the system management table is calculated. Increment by 1 (step 212), instructing the search DB system (B) 80 to execute a search process (step 213), and after the search process is completed, the number of database unit executed transactions of the DB in the system management table is reduced by 1 (step 214). ), And terminate the process.

Finally, in step 220, if the DB access language is "C", the processing is passed to the system load management unit 30, and the specified DB name is used to retrieve the minimum load rate search DB name or update DB name. (Step 221), the number of execution transactions per database of the DB in the system management table is increased by 1 (Step 222), and a search process is instructed to the DB system (Step 223). The number of database unit transactions of the DB is reduced by 1 (step 224).
At 25, if the DB is a DB lock, execute the process again,
If the DB is not a DB lock, the process ends.

On the other hand, in step 220, the DB access language is "D"
If it is, the update DB is instructed to swap the update data history information (step 231), and the update data reflection flag of the search DB that updates the system management table is turned on (step 232). Update data history information 70
(Step 233), and after the update data history information update processing, turns off the update data reflection flag of the search DB reflecting the update of the system management table (step 234), and ends the processing. I do.

FIG. 5 is a flowchart showing the operation of the system load management unit 30 according to the present embodiment. Hereinafter, an operation example will be described in detail in accordance with this.

First, the load factor of each DB is determined from the number of transactions executed in database units and the number of allowable transactions in the system (step 300). Next, in step 301, the copy DB name is determined by referring to the system management table from the designated DB name, and in step 302, if the update data reflection flag of the DB is off from the system management table, the process proceeds to the next step. If on, search the next system management table. In step 303, if it is the minimum load factor, the process proceeds to the next step, and if it is not the minimum load factor,
Search the next system management table. If the copy DB name has been obtained in step 305, the obtained DB name is returned (step 307). On the other hand, if the copy DB name has not been obtained in step 305, the designated DB name is returned as the obtained DB name.

According to the above-described embodiment, in a database system or the like, if the latest of the data to be searched is not required, the processing overhead due to the competition of the data resources can be eliminated, and the search processing performance of the system can be improved. In addition, since user requests are distributed to the most appropriate database in consideration of the load balance of a plurality of database systems, it is effective in improving the throughput performance of the entire system.

In addition, each said Example is shown as an example,
The present invention should not be limited to these.

〔The invention's effect〕

As described above, according to the present invention, in a database system separated into a search database and an update database or update data history information, in response to a search request for a database that requires up-to-date Is processed using an update database or an update data history information and a search database, and in response to a search request for a database that does not require up-to-dateness, the search request is processed using the search database. Since the processing is performed, there is a remarkable effect that it is possible to eliminate the conflict between the database resources due to the search processing and the update processing and to realize a database processing system that achieves a high-speed search processing.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a DB system according to one embodiment of the present invention, FIG. 2 is a diagram showing an outline of a DB access language, FIG. 3 is a diagram showing a format of a system management table, and FIG. FIG. 5 is a flowchart showing the operation of the DB system management unit, and FIG. 5 is a flowchart showing the operation of the system load management unit. 10: Language analysis unit, 20: DB system management unit, 30: System load management unit, 40: DB system (1), 50: DB system (2), 60: Update DB (A), 70: Update data history information,
80: Search DB (B).

Claims (5)

(57) [Claims] In a database system which is separated into a search system database and an update system database or update data history information, in response to a search request for a database requiring up-to-date, the search request is sent to the update system database or A database processing system for processing using update data history information and a search database. 2. A search system according to claim 1, wherein said search request is processed using a search system database in response to a search request for a database that does not require up-to-dateness.
The database processing system described in the section. 3. The system according to claim 1, wherein said search request is processed using a database having a light load in a system in response to a search request of a database regardless of its up-to-dateness. Database processing system. 4. The system according to claim 3, wherein, if a database lock occurs during the processing, said database processing is stopped and said search request is processed using a database of another system. Database processing system. 5. The database processing system according to claim 1, wherein a database access language indicating whether or not the search request needs to be up-to-date is used.