JPH06119226A - Distributed data base management system - Google Patents

Abstract

PURPOSE:To dispense with considering an updating timing of a data file of each work station, and to secure the simultaneity of data. CONSTITUTION:The system is provided with a data file having two areas of an area used for writing data and an area used for reading out data, and in the case the data is varied in the own station work station (a), a data communication part 5 broadcasts its data to other work stations (b)-(d), and the work station for transmitting the data, and the work station for receiving the data issue a data write request to the respective data managing parts 6 by the timing of transmitting completion, and by the timing of receiving completion, respectively, and the data managing part 6 executes a data write processing, issues a resources obtaining request of an area changeover switch, so that the data area is switched at the time point when the resources can be obtained, and also, in the case the data is used in each work station, the data in the area rewritten in the previous time is used.

Description

Detailed Description of the Invention

[0001]

[Industrial application] Steel plants, power generation systems, F
The present invention relates to a controller system provided with an external storage device used in an A system or the like, and particularly to a distributed database management system.

[0002]

2. Description of the Related Art In a conventional distributed database system, as shown in FIG.
Then, for example, when data changes due to input from the keyboard or changes in the state of the device, the system control unit 2a first writes the data to the data file 3a on its own hard disk (hereinafter referred to as HD), and then the network. The data is sequentially transmitted to each of the above workstations, and each workstation executes the writing process to each HD at the timing of receiving the data.

[0003]

However, in this method, when the writing process of the received data is executed,
It is difficult to manage the data update timing when each workstation is using the data (the data is being accessed by some program 4). Further, since the generated data are not transmitted at the same time, there is a drawback that the simultaneity of the data cannot be guaranteed. On the other hand, JP-A-63
Japanese Patent No. 300336 discloses an example of file duplication, but is not applied to a distributed database. Japanese Unexamined Patent Publication No. 62-49460 discloses an example in which broadcast is applied to a distributed database, but it does not guarantee data simultaneity. Therefore,
SUMMARY OF THE INVENTION It is an object of the present invention to provide a distributed database system capable of guaranteeing data concurrency without having to consider the update timing of the data file of each workstation.

[0004]

In a distributed database system in which a plurality of workstations are connected via a network, each workstation has two areas, an area used for writing data and an area used for reading data. A data file having an area, a data management section having means for switching between the two areas each time data is written, and a data communication section for communicating with other workstations and broadcasting when data changes The data communication unit broadcasts the data to other workstations when the data is changed in the local workstation, and the workstation that has transmitted the data receives the data at the transmission completion timing. At the time of reception completion at the station,
When a data write request is issued to each data management unit, the data management unit receives the write request, executes the data write process, issues a resource acquisition request for the area changeover switch, and acquires the resource. It is characterized in that the data area is switched by and the data in the area rewritten last time is used when the data is used in each workstation.

[0005]

By broadcasting the generated data and issuing a data write request to the data management unit at the timing of the completion of transmission / reception, there is almost no time lag in data writing in each workstation, so that simultaneous data transmission is possible. Data can be guaranteed, and the HD has two areas for each data file, and the switching of the uses of these two areas is managed by the data management unit, so that the data update is performed without considering the data update timing. You can

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below based on an example shown in the drawings. FIG. 1 is a block diagram showing the outline of the present invention, and FIG. 2 is a flow for explaining the operation of the present invention. When data changes in the workstation a, the data communication unit 5a broadcasts the data to all the workstations b, c, d on the network (). Next, at the workstation a that has become the data transmitting station, at the timing of completion of transmission, at the workstations b, c, d that become the data receiving stations, at the timing of completing data reception.
A data write request is issued from the data communication units 5a to 5d to the data management units 6a to 6d (a to d).
Then, both the transmitting station and the receiving station write data to the hard disk 7 (a to d), but in the present invention, two areas of the disk are provided, and they are automatically switched to each other for each writing. It That is, in FIG. 1, the workstation a is switched to HD by the switch 8a1.
The area is written in the area (1). In this case, the control unit 10a, via the data access unit 9a,
Although the data in the area (2) is selected by the switch 8a2, this area (2) is the area selected in the previous data writing. Similarly, workstation b
And c are areas (2), workstation d is area (1)
And the control unit 10 of each workstation is
Data in an area different from that area is selected. As a result, each workstation commonly uses the data before the data is written, and the simultaneity of the data is ensured. Next, the processing flow of the data management unit will be described. When the event waiting state of 13 is exited, it is checked at 14 whether or not there is a data write request from the data communication unit. If there is no request, it waits for 13 events. If there is a request, the data storage area address of the data file corresponding to the received data in 15 (becomes the address of the area that was not updated last time)
Is calculated and data is written to the HD at 16 at that address (a to d). That is, the switches shown in the block diagram of FIG. 1 are realized by setting addresses on software. Next, in order to switch between the data storage area (1) and the data access area (2), the area changeover switch (8a
Issue a resource acquisition request of 1, 8a2). At 18, it is determined whether the resource has been acquired. At this time, the data access unit 9a is already acquiring the resource, and if the resource cannot be acquired, it waits at time 19 and loops 18 and 19 until the resource can be acquired (during this loop, the same process is performed again). When the data of the data file is generated, the resource acquisition is canceled, and the process of 15 is started again). When the resource is acquired, the switches 8a1 and 8a2 are switched at 20. By this processing, the data written in 16 is replaced with the data access area, the previous data access area is replaced with the data write area (for the next writing), and the next time the data is updated, the data updated this time is replaced. Will be referenced. At 21, the resources are released and the event waits again.

[0007]

As described above, according to the present invention, there is a great effect that it is possible to provide a distributed database system which can guarantee data simultaneity without considering the update timing of a data file.

[Brief description of drawings]

FIG. 1 is a schematic diagram of a distributed database system of the present invention.

[Fig. 2] Operation flow of data writing process of the data management unit

FIG. 3 is a schematic diagram of a conventional distributed database system.

[Explanation of symbols]

Reference numeral 5 is a data communication unit, 6 is a data management unit, 7 is a hard disk, 8 is a switch, 9 is a data access unit, 10 is a control unit, and a, b, c and d are workstations.

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] August 18, 1993

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be amended] Title of invention

[Correction method] Change

[Correction content]

Title of the invention Distributed database management system

Claims (1)

[Claims] 1. In a distributed database system in which a plurality of workstations are connected via a network, each workstation has data having two areas, an area used for writing data and an area used for reading data. The data communication unit includes a file, a data management unit having means for switching between the two areas each time data is written, and a data communication unit that communicates with another workstation and broadcasts when data changes. When the data changes in its own workstation, it broadcasts the data to other workstations, and the workstation that sent the data indicates the completion of the transmission and the workstation that received the data indicates the completion of the reception. At the timing, the data management section A data write request is issued, and the data management unit executes a data write process in response to the write request, issues a resource acquisition request for the area changeover switch, and switches the data area when resources are acquired. In addition, when using data at each workstation, the distributed database management method is characterized in that the data in the area rewritten last time is used.