JPS61117644A - Double inquiry preventing system of inquiry terminal in cluster system - Google Patents

Abstract

PURPOSE:To reduce required resource and processing time of a host computer and use resources and the throughput for another purpose by providing each cluster with a table which manages inquiry states of inquiry terminals connected to the cluster. CONSTITUTION:A cluster 2 is provided with management tables which manage inquiry states of inquiry terminals connected to this cluster, and double inquiry is managed. When an inquiry text comes from a terminal A3, the cluster 2 registers the inquiry execution in the table for the terminal A3 and sends the inquiry text to a host computer 1. The computer 1 performs the processing for this inquiry and transmits a response text. When receiving the response text from the computer 1, the cluster 2 transmits it to the inquiry terminal A3; and when receiving the last of the response text, the cluster 2 registers the idle state in the management table for the terminal A3. Thus, double inquiry is prevented to reduce required resources and processing time of the computer 1, and resources and the throughput can be used for another purpose.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for preventing double inquiries from inquiry terminals in a cluster system. A data processing system consisting of a large number of terminals, a large number of clusters (terminal control units W) that control the terminals and edit the terminal data into images that can be understood by the central host computer, and a central host computer is called a cluster system. It is called.

FIG. 2 is a system configuration block diagram of the cluster system. In the figure, 1 indicates a central host computer, 2 indicates a cluster, and 3 indicates a terminal.0For example, in a bank online system, there are approximately 200 clusters that control approximately 50 terminals, and a central host computer that controls approximately 50 terminals. In other words, it is a cluster system with approximately 10,000 terminals.

Like the various terminal devices in the bank online system, all processing inquiries are made to the central host computer nIa,
The terminal that executes the inquiry based on the result is called an inquiry terminal. It is strictly necessary to prevent one of these inquiry terminals from issuing another quantity adjustment while making an inquiry to the central host computer, as this could cause malfunction. The present invention relates to a management method for preventing double queries.

[Prior Art] Conventionally, in order to prevent double queries in this type of system, a central host computer has management information for all query terminals and executes the management.

[Problems to be Solved by the Invention] In this way, it is impossible for a central host computer to manage double queries for all terminals, as in the example of the bank online system described above. When the number of terminals reaches 1,000, it has become impossible to ignore the host computer's resources (main storage, external storage, etc.) and processing capacity.

[Means for solving the problem] The above problem can be solved by providing each cluster with a table for managing the query status of the query terminals connected to the cluster, and performing management to prevent double queries. This problem is solved by the double inquiry prevention method of the inquiry terminal in the cluster system of the present invention.

[Operation] That is, each cluster has information and a management table for the inquiry terminals connected to it, and each time an inquiry message is received from a terminal, "inquiry in progress" is recorded in the table of that terminal, and the host computer Each cluster manages double inquiries by recording "idle" each time a response message ends.

[Example] The gist of the present invention will be specifically explained below with reference to an example shown in FIG.

FIG. 1 is a processing sequence diagram in one embodiment of the present invention.

Cluster 2 has information on each inquiry terminal connected to this cluster and a management table to manage double inquiries.

■When an inquiry message is sent from terminals A and 3, cluster 2 registers the message "inquiry in progress" in the table of terminals A and 3, and sends the inquiry message to the host computer.

■The host computer 1 processes the query,
Send a response message.

- In cluster 2, upon receiving a response message from the host computer, terminal A inquires about it.

3, and when the end of the response message is received, the "idle state" is registered in the management table of terminal A.

As explained above, when a host computer manages double queries for an extremely large number of query terminals, its resources and processing power cannot be ignored, but when managing this on an individual cluster, it This can be done with the resources and processing power of

[Effects of the Invention] As explained above, the present invention has the effect of reducing the required resources and processing time in a host computer, and making it possible to use free resources and processing power for other purposes. .

[Brief explanation of drawings]

FIG. 1 is a processing sequence diagram in one embodiment of the present invention, and FIG. 2 is a system configuration block diagram of a cluster system. In the drawing, 1 is a host computer, 2 is a cluster, and 3 is an inquiry terminal.

Claims (1)

[Claims] In a cluster system having a large number of inquiry terminals, each cluster is equipped with a table for managing the inquiry status of the inquiry terminals connected to the cluster, and is configured to perform management to prevent double inquiries. Double inquiry prevention method for inquiry terminals in cluster systems.