JPS61173352A - Processing system of vertical dispersing data base - Google Patents

Abstract

PURPOSE:To transfer the data between CDBs at a high speed and to prevent the reduction of a parallel processing by making the capacity of the file storing a central data base CDB into a necessary minimum quantity and storing continuously the page. CONSTITUTION:The section between respective data base control parts 14 and 15 of a center system 11 and a local system 12 are connected by a communication channel, and a data base processing system is constituted. The CDB13a of the CDB file device 13 of a system 11 is divided into seven blocks A-G, and the necessary minimum pages are continuously stored to respective blocka A-G. The page of respective blocks A-G of the CDB13a is read by a control part 14, and transmitted to the control part 15 of the system 12. The transmitted page is stored into the PDB17 of a private data PDB file device 17 of the system 12 and the page-upping table 16a of a main memory 16. The data between CDBs 13a are transferred at the high speed, and the reduction of the parallel processing is prevented.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a vertically distributed database processing method in which a part of a database is copied and transferred from a central system of a network to each local system for processing. The present invention relates to a vertically distributed database processing method having means for reducing the amount of data transferred between a center system and local systems and the amount of data stored in local systems.

[Conventional technology]

The vertically distributed database processing method uses a database as a central database CDB (hereinafter simply referred to as CDB) for centrally managing various types of data used in various operations.
) and an arbitrary number of private databases, PDB (hereinafter simply referred to as PDB), which are extracted from CDB for specific tasks, and each is managed independently for a certain period of time. In the national network that manages power lines, power lines, etc.
The central system has a CDB that covers all regions, and the local systems of business offices in each region have PDBs that cover each region and perform database processing on a regular basis. The contents of each PDB are then reflected in the CDB.

Figure 2 illustrates the concept of a network in such a vertically distributed database processing system, where 1 to 5 are computer systems (here simply referred to as systems) that make up the network, and 1 is a particularly central computer system. The remaining 2 to 5 are local systems. And 1a is the CDB placed in the center system 1,
2a to 5a are local systems 2 to 5, respectively.
This is a PDB located in .

Both CDB and PDB are page managed.

Each PDB is a complete copy of the CDB, but each local system is managed so that only the page area cut out from the CDB by the local system can be accessed, while the corresponding area within the CDB is locked so that it cannot be updated.

For example, CDB (la) is divided into blocks A, B5C5D shown in the figure for each region of the local system. Each block contains multiple pages. Then, block A is allocated to local system 2, block B to local system 3, and so on, and the processing range of each local system is allocated. As a result, for example, local system 2 can access block A of PDB (2a).
For database processing such as query update, 1
This is done independently from other systems for a reasonably fixed period of time, such as months or 40 days, while block A of the CDB is accessed during that time. When the lights are lit for a certain period of time, the PDB is transferred to the center system to synchronize the systems.

[Problem that the invention seeks to solve]

In the conventional vertically distributed database processing method, CDs, including unnecessary pages, are sent from the center system to the local system.
I had to forward a copy of B. Therefore C
The DB and PDB were the same size, and each local system required the same storage space as the center system. In addition, when unnecessary pages are transferred to the local system, the CDB including those unnecessary pages is locked and cannot be updated from the center system or other local systems.
This was a cause of reducing the parallel processing performance of the PDB.

[Means for solving problems]

The present invention provides a page mapping table in the local system so that the storage space for storing the PDB in the local system is only for the necessary pages, and the data transfer between the local system and the center system is only for the necessary pages. , unnecessary pages are managed as dummy pages, and its configuration is a vertically distributed database processing method in which a specific business database that forms part of the page-managed central database of the center system is managed as a private database of the local system. In the local system, only the specific business database nohage is stored in the file device, and a page mapping table means is provided for comparing the pages of the central database with the storage location of the file device, and the page mapping table is In this method, by comparing unnecessary pages in the central database with a dummy, only the pages of the specific business database can be made compatible with the central database, and the database without unnecessary pages can be transferred between the center system and the local system. It is characterized by doing.

〔Example〕

The details of the present invention will be explained below according to the embodiment shown in FIG.

In FIG. 1, 11 is a center system, 12 is a local system, 13 is a CDB file device, and 13a is a C
DB (central database), 14 and 15 a database management unit, 16 a main memory, 16a a page mapping table, 17 a PDB file device, and 17a a PDB (private database).

CDB 13 a is divided into seven blocks represented by A, B, C, DSE, F, and G, and the number of pages included in each block is A: 100, B: 200, and C:
1500, l): 200, E: 1000, F: 500
, G:6500. Here, B, D, F
A total of 900 pages in the three blocks are required pages in the PDB 17a of the local system 12.

Introduction Database management section 1 of center system 11
4 takes out each page of blocks B, D, and F of the CDB 13a from the CDB file device 13 and transfers it to the database management unit 15 of the local system 12, but at the same time, the number of pages of each block A to G is transferred to the PDB.
Transfer as management information.

The database management unit 15 of the local system 12 sequentially stores the pages of the BLD and F programs transferred from the center system into the PDB 17a of the file device 17.
Continuously stored in the page mapping table 16a and simultaneously stored in the page mapping table 16a
Create. Block B is stored in the PDB file device 17.
, D, and F are provided. Main memory 1
The page mapping table 16a provided in 6 is
It shows the correspondence between the logical pages of CDH and the real pages of PDH, and shows the unnecessary page parts (block A, block A,
C5ESC) is displayed as a dummy.

As a result, a PDB is created as a compressed file with unnecessary pages removed from the CDB, and moreover, the PDB can be equally accessed using the logical pages of the CDB.

FIG. 3 shows the details of the above-mentioned page mapping table 16a. Here, block boundaries within the CDB 13a are mapped to page numbers P#1, P#101, P#300.
1060. Also, the block position in PDB 17 a is a relative block number (offset) 1.201.401 from the beginning position of PDB file device 17.
It is expressed as .

The page mapping table 16a includes the number of management pages for each block, the target file number, and the relative block number within the target file. The number of management pages is block B,
The target file number for D and F is the number of the PDB file device, but in this embodiment, since there is only one PDB file device, it is set to "1". Also, the relative block number is "1" for block B, and block D is "1" for block B.
"201" is set for block F, and "401" is set for block F. However, blocks A, C, and E.

The dummy for G can be identified by setting both the target file number and the relative block number within the target file to "θ".

〔Effect of the invention〕

According to the present invention, the space of the file device that stores the PDB is required to be the minimum amount, and since pages are physically stored consecutively in the file device, data transfer between the CDB and the CDB can be performed at high speed. be able to. Furthermore CDB
Since this method does not pass unnecessary pages to the PDB and does not lock them, it is possible to prevent deterioration in parallel processing performance.

[Brief explanation of drawings]

FIG. 1 is a block diagram of one embodiment of the present invention, FIG. 2 is a conceptual diagram of a network of a vertically distributed database processing system, and FIG. 3 is a detailed diagram of a page mapping table. In the figure, 11 is a center system, 12 is a local system, 13a is a CDB (central database), 1
4.15 is a database management section, 16a is a page mapping table, and 17a is a PDB (private database).

Claims (1)

[Claims] In a vertically distributed database processing method in which a specific business database that forms part of the page-managed central database of the center system is managed as a private database in the local system, the local system stores only the pages of the specific business database in the file device. In addition, a page mapping table means is provided for comparing the pages of the central database with the storage location of the file device, and the page mapping table means compares the unnecessary pages in the central database with a dummy, so that the specific business database is A vertically distributed database processing method characterized by making only pages compatible with the central database, and transferring the database excluding unnecessary pages between the center system and local systems.