JPH0448352A - File decentralization system - Google Patents

Abstract

PURPOSE:To attain the effective application of the file resources by providing a file control part to update a control file and controlling collectively the data files decentralized to each subprocessor through a main processor for reference and updating. CONSTITUTION:A control file 2 is produced in a file production state, and the presence information on each data record is edited into the control data record and registered in the file 2. A sub-CPU 10a produces a reference request for the data record and the relevant data record is not included in its own decentralized file 15a. In such a case, a data control part 12 informs the reference conditions and asks a main CPU 1 for reference. Thus a file control part 4 refers to the file 2 to transfer the designated data record to a sub-CPU 10b of the referred side and then to a sub-CPU 10a of the referring side. The sub- CPU 10a changes the received data record and sends it to the CPU 1, and the part 4 registers the presence information as a control data record. As a result, no duplex file is produced and therefore the file resources can effectively be applied.

Description

[Detailed Description of the Invention] [Summary] The present invention relates to a file distribution method, in which data files are distributed under the management of multiple subprocessors, and the distributed files are collectively managed by a main processor, with the aim of effectively utilizing file resources. A file distribution method that includes a data management unit provided in each sub-processor that references and updates the distributed file via the main processor, and a data management unit provided in the main processor that stores location information of each data record in the distributed file. A management file for registering the data record, a reference request from each subprocessor to relay the relevant data record, an update request to relay the relevant data record, and register the changed data record in the corresponding distributed file, and the management file. The system is configured such that data files are distributed under each sub-processor, and the main processor collectively manages the distributed files for reference and updating.

[Industrial application field]

The present invention relates to a file distribution method for distributing mask files provided in a main processor to distributed sub-processors.

[Conventional technology]

FIG. 6 is an explanatory diagram of the prior art.

Figure 6 shows the main processor of the center (hereinafter referred to as main CP).
This is an example of a distributed system in which there is a mask file 21 in U) 20, and sub-CPUs 30 provided in each office etc. perform predetermined business processing while referring to this master file 21.

In order to refer to this master file 21, each sub CPU
30 is provided with a local file 31. For example, at the start of business, the mask file 21 is copied to the local file 31, and this is referred to during processing, and when registering and updating data records, new data records are copied. Alternatively, a changed data record 23 is created and transferred to the main CPU 20.

In the main CPU 20, the transferred data record 23
is stored in the transaction file 22 and reflected in the mask file 21 after the business is completed.

[Problem to be solved by the invention]

The method of directly referencing the mask file 21 requires communication time and causes congestion on the main CPU 20. For this reason, as explained above, a method is used in which the mask file 21 is copied to the local file 31 and referenced, but the file is duplicated and there is a problem that file resources are not used effectively. .

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to provide a file distribution method that effectively utilizes file resources.

[Means to solve the problem]

The principle components will be extracted and explained from FIG. 1 - the configuration diagram of the embodiment.

Reference numeral 1 denotes a main processor (main CPU), which includes a management file 2 and a file management section 4.

2 is a management file for each sub-processor (sub-cpu
) The location information of each data record of the distributed files 15a and 15b placed in the management data record is edited and registered in the management data record.

4 is a file management unit that relays designated data records between sub-020s based on reference requests and update requests;
In the case of an update request, the corresponding distributed file is updated based on the notified change data record, and the management file 2 is updated.

10a and 10b are sub CPUs, each of which includes distributed files 15a and 15b and a data management section 12.

Reference numeral 12 denotes a data management unit, which requests the main CPLI i to receive and receive the specified data record when referencing other distributed files, and when updating, sends an update request to the main CPLI and receives and receives the specified data record. The changed data record is sent to the main CPU 1, and when an update instruction is issued from the main CPU 1, the transferred data record is registered at the designated address of its own distributed file.

[For production]

The distributed files 15a and 15b are distributed by each sub CPU 10a,
10b, each file is created depending on the job, or the created files are divided and placed.

The management file 2 is generated at the time of file creation, and the location information of each data record is edited and registered in the management data record.

When a data record reference request occurs in the sub CPU (10a) and the data record does not exist in its own distributed file 15a, the data management unit 12 notifies the reference conditions (location information, etc.) and requests the main CPU 1 to reference it.

Thereby, the file management unit 4 refers to the management file 2, causes the reference destination sub CPU 10b to transfer the specified data record, and transfers it to the reference source sub CPU 10a.

Further, when a data update request occurs in the sub CPU 10a, the main CPU 1 is notified of the location information and an update request is made. As a result, the file management unit 4 of the main CPU 1 transfers the specified data record to the update destination sub CPU (10
a or 10b, herein referred to as 10b), and then to the requesting sub CPU 10a.

The sub CPU 10a that received this changes the data record and sends it to the main CPU υ1, and the file management unit 4
After registering the location information as a management data record and deleting the previous management data record,
The U 10b is notified of the changed data record and location information and is instructed to register.

Thereby, the update destination sub CPU 10b registers the transferred changed data record at the specified address of the distributed file 15b.

As described above, since the main CPU 1 only has the management file 4, duplication of data files is eliminated and effective use of file resources is achieved.
, processing time and comparison of the main CPU 1 are improved.

〔Example〕

Embodiments of the present invention will be described in detail with reference to the drawings.

Fig. 1 is a configuration diagram of an embodiment, Fig. 2 is a diagram showing an example of a management data record, Fig. 3 is a flowchart of a distributed file creation process, Fig. 4 is a flowchart of a reference process, and Fig. 5 is a diagram showing an example of a management data record.
The figure is an update processing flowchart.

In this embodiment, the sub CPUs 10a and 10b each create data files, and the distributed files 15a and 15
In FIG. 1, which shows an example of storage as b, 1 is the main processor (1) that manages the sub CPUs 10a and 10b.
The main CPU includes a management file 2 and a file management section 4.

The management file 2 includes distributed files 15a and 15b created and held by each subCPυ10a and 10b, respectively.
The location information of each data record is stored as a management data record 3.

The file management unit 4 relays designated data records based on reference requests and update requests, and in the case of update requests, updates the corresponding distributed file and updates the management file 2.

10a and 10b are sub CPUs, which respectively handle distributed files 15a and 15b, a data management section 12, and a creation section 11.
Equipped with.

The data management section 12 includes a reference section 13 and an update section 14. When the reference data record does not exist in the distributed file it owns, the reference unit 13 requests the main CPU 1 to refer to another distributed file, and receives and receives the specified data record. Further, when a change request occurs, the update unit 14 notifies the main CPU 1 of the update conditions, requests the update, receives the target data record, and sends the changed data record to the main CPU 1. When the main CPU 1 notifies the changed data record, it registers it in its own distributed file.

The creation unit 11 is activated by a distributed file processing command, causes the main CPU 1 to transmit the file definition 5, creates data records in a predetermined format, and notifies the main CPU 1 of the location information of each created data record. .

Here, the file definition body 5 defines the format of the data record and management data record 3, and each sub-C
When the PUs 10a and 10b refer to this and create data records, the distributed file 15a in the same format is created.
, 15b are generated.

In addition, the management data record 3 is composed of location information of the data record, and as an example is shown in Fig. 2, it is composed of the sub CPU number, library salt, file name, relative address within the distributed file, etc. , the record number of management file 2 is issued and registered.

In the distributed system with the above configuration, the following file processing and file management are performed.

(File creation process) See Figure 3 (1) When creating a file, issue a distributed file processing command. At this time, specify the file name as a parameter.

(2) The creation unit 11 activated by this notifies the main CPU of file creation, creates a data record based on the sent file definition 5, and generates a distributed file. The location information of each data record includes the sub CPU name and library name.

The main CPU 1 is notified of the file name, relative address within the file of each data record, etc.

(3) The file management unit 4 of the main CPU i generates a management file 2 by referring to the file definition 5 if the management file 4 has not been generated, and creates a management data record 3 for each notified data record. Each record number is issued and registered in the management file 2.

(Reference Process) See Figure 4 (1) When referring to the distributed file 15b provided in another sub CPU 10b, the reference unit 13 of the reference source sub CPU (10a) sets the reference condition (location information) to the main CPU 1. Please notify and request reference.

(2) The main CPII 1 refers to the management file 2, causes the reference destination sub CPt 110b to transfer the specified data record in the distributed file 15b, receives it, and transfers (relays) it to the reference source sub CPU 10b.

(Update process) See Figure 5 All update processing is performed via the main CPU 1.

(1) The update unit 14 of the update source sub CPU 10a notifies the update condition (same as the reference condition) to the main CPU i,
Request an update.

(2) Main CPU 1 is the update source sub cpυ (10
The designated data record of the distributed file 15b (denoted as b) is transferred to the update source sub CPU 10a.

(3) In the update source sub CPU 10a, the transferred data record is changed, and this changed data record is sent to the main CPU I by the update unit 14.

(4) When the file management unit 4 of the main CPU 1 receives the changed data record, the file management unit 4 of the main CPU 1 sends it to the registered distributed file 1.
After determining the relative address of 5b, creating a management data record 3, and newly registering it in the free space of the management file 2, and deleting the management data record 3 to be updated,
Transfers the changed data record and the registration destination relative address to the update source sub CPU IC-b and instructs update processing.

(5) The update source sub CPU 10b registers the changed data record at the specified relative address of the distributed file 15b and deletes the previous data record.

As described above, by having the management file 2 in the main CPU 1, the distributed files 15a and 15b are created.

References and updates can be managed all at once.

The capacity of this management file 4 is much smaller than that of the data file, and effective use of file resources can be achieved.

Furthermore, in a system where the sub CPUs 10a and 10b share tasks, they often refer to their own distributed files 15a and 15b, respectively, so there are fewer requests to the main CPU 1, resulting in an increase in processing time and
I never come wearing short clothes.

In the embodiment, location information is specified as the reference and update condition, but a key file may be provided separately and specified using a key.

〔Effect of the invention〕

As explained above, the present invention places a distributed file in each sub-processor, the main processor collectively manages this distributed file using a management file, and the sub-processor updates data and other distributed files via the main processor. This method provides a file distribution method configured to reference files, and avoids duplication of files without reducing processing time, resulting in the effective use of file resources.

[Brief explanation of the drawing]

Fig. 1 is a configuration diagram of an embodiment, Fig. 2 is a diagram showing an example of a management data record, Fig. 3 is a flowchart of a distributed file creation process, Fig. 4 is a flowchart of a reference process, and Fig. 5 is a diagram showing an example of a management data record.
The figure is an update processing flowchart, and FIG. 6 is an explanatory diagram of the prior art. In the figure, 1.20 is the main processor, main CPU, 2
1 is a management file, 3 is a management data record, 4 is a data management section, 13 is a reference section, 14 is an update section, and 15a. 15b is a distributed file, 16a and 16b are data records, 17a and 17b are libraries, 21 is a mask file, 22 is a transaction file, 23 is a transferred data record, and 31 is a local file. Distributed file creation processing flowchart Figure 3 Diagram representing an example of management data record Figure 2 Update processing flowchart Figure 5

Claims (1)

[Claims] A data file is distributed under the management of a plurality of subprocessors (10a, 10b), and this distributed file (15a, 15
b) is centrally managed by the main processor (1), and is provided in each sub-processor (10a, 10b), and the distributed files (15a, 15a, 10b) are managed via the main processor (1).
15b) Data management unit (12) that references and updates
, a management file (2) provided in the main processor (1) that registers the location information of each data record in the distributed file, and each sub-processor (10a, 10b).
), relay the corresponding data record in response to a reference request from
A file management unit (4) is provided that relays the corresponding data record in response to an update request, registers the changed data record in the corresponding distributed file, and updates the management file (2), and manages the data under each subprocessor. A file distribution method characterized by distributing files and allowing a main processor to collectively manage the distributed files for reference and update.