JP2001331468A - Linked job flow management system among plural computers - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a linked job flow management system among plural computers for enabling scheduling (processing order allocation) among different platforms. SOLUTION: This system is provided with the plural computers (agents 21-2N) to be managed for asynchronously generating various events and a management computer (manager 1) for managing the schedule (processing order allocation) of the plural computers to be managed and the events. The management computer is provided with a unitary management means (message acquisition means 1-1, condition judgment means 1-2 and command execution means 1-3, etc.), for unitarily managing the events and advances to the next processing according to the management result of the unitary management means.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cooperative business flow management system between a plurality of computers, and more particularly, relates to a system in which events generated asynchronously among a plurality of computers are associated with a workflow, monitored, and automatic operation is performed. The present invention relates to a system for managing a cooperative workflow between a plurality of computers.

[0002]

2. Description of the Related Art In a computer system in which a plurality of computers are connected by a network (for example, a vertical distributed processing system shown in FIG. 5), a certain event is generated in response to an asynchronous event such as the occurrence of a message or the arrival of time. There is a management method in which monitoring is performed in association with an operation mode (hereinafter, abbreviated as “workflow”) in which another job is sequentially started after a job is completed, and command issuance of a next process is automatically controlled.

[0003]

However, the conventional management method is effective only for workflow management for one computer (for example, the distributed computer 101 in FIG. 5). Therefore, the conventional management method cannot be directly applied to the other distributed computers 102 (1 MN, for example, 102, 103,...) Shown in FIG. could not.

Therefore, an object of the present invention is to solve the problem of Windows
To provide a cooperative business flow management system between a plurality of computers capable of scheduling (processing order allocation) between different platforms such as personal computers such as (registered trademark) machines and UNIX (registered trademark) machines, workstations, and POS terminals. It is.

[0005]

According to the present invention, there is provided a computer system in which a plurality of computers are connected by a network to manage an event occurring between the plurality of computers. A cooperative workflow management system between a plurality of computers for performing cooperative operations between the plurality of computers, wherein a plurality of managed computers that asynchronously generate various events, and a schedule of the plurality of managed computers ( And a management computer for managing the event. The management computer includes a centralized management unit for centrally managing the event, and performs the following in accordance with the management result of the centralized management unit. The process is characterized by proceeding to the process of (1). If you do this,
By centrally managing the schedules of the plurality of managed computers by the management computer, scheduling (processing order allocation) in cooperation with the plurality of managed computers becomes possible.

[0006]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a system for managing a cooperative workflow between a plurality of computers according to the present invention. In FIG. 1, a cooperative workflow management system TR between a plurality of computers according to the present embodiment includes:
Manager 1, which is a "managed computer", agents 21, 22, 22, which are "a plurality of managed computers"
2N. Here, a manager refers to a computer that centrally manages the progress of an associated workflow (cooperation flow) and controls automatic operation, and an agent refers to a computer that is scheduled (managed operation) by the manager. .

[0007] The manager 1 centrally manages messages generated by each agent. The manager 1 includes: a message acquisition unit 1-1 for acquiring a message generated from an agent; a condition determination unit 1-2 for determining a relationship between an asynchronously occurring event such as the occurrence of a message and the arrival of time; Command execution means 1-3 for notifying the issuance of a command to a command, and a definition information file 1-4 for describing the relationship between the event and the workflow.
A condition determination table 1-5 for storing whether or not a defined condition is satisfied; a system clock 1-6 for notifying the arrival of the time of the system;
And a time acquisition unit 1-7 for notifying the time information to the second unit. The message acquisition unit 1-1, the condition determination unit 1-2, the command execution unit 1-3, the definition information file 1-4, the condition determination table 1-5, the system clock 1-6, and the time acquisition unit 1-7 The objective management means is constituted.

[0008] The agent 2N issues a command in response to a notification from the manager 1 and executes the command.
-1 and a resident job 2N-2, such as virus monitoring, input history storage, alarm display, etc., stored in the main memory, and a message generated by the agent, and the manager 1
And a message acquisition unit 2N-3 for transmitting the message to the user.
The message acquiring unit 2N-3 of the agent 2N acquires a message generated in the resident job on the agent and transmits the message to the message acquiring unit 1-1 of the manager 1. The message acquisition means 1-1 of the manager 1 collects messages generated in each agent, and determines the condition.
Send to 1-2. Manager 1 command execution means 1-3
Issues a command to the system via the command execution means 2N-1 of the agent 2N (instruction flow indicated by a dotted arrow in FIG. 1).

The definition information file 1-4 on the manager 1
Is a file that stores the context of each step of the workflow, the relation with the messages that occur asynchronously, and the definition of the command to be issued. A condition definition for issuing one command can have a plurality of conditions. The condition definition is defined by combining events that occur asynchronously, such as the end of a process, the arrival of a message generated at each host, and the arrival of time. The definition information file is referred to by the condition determination means 1-2, and the definition in the definition information file 1-4 is reflected in the condition determination table 1-5.
The condition determination table 1-5 is a table that stores a condition definition for each command to be issued and whether or not a condition for issuing the command is satisfied.

Next, the operation of this embodiment will be described with reference to FIGS. 2 is a conceptual flowchart of the entire embodiment, FIG. 3 is a flowchart of a specific example mainly on the manager side, and FIG. 4 is a flowchart of a specific example on the agent side.

As shown in FIG. 2, in the agent 2N (see FIG. 1), the resident job 2N-2 is changed to the message 2N-4.
Is output. Message acquisition means 2N of agent 2N
-3 acquires the message 2N-4 and notifies the condition determination means 1-2 via the message storage means 1-1 of the manager 1 (step S1). The condition determination unit 1-2 refers to the condition determination table 1-5 and establishes a condition that is satisfied with the arrival of the message 1-8 (step S2). It is determined whether or not all the conditions of the command to be issued have been satisfied because the conditions have been satisfied (step S3). If the command is not executed, the process waits for the next event (step S3: NO).

When all the conditions are satisfied by the message output from the agent 2N or the arrival of the time 1-9, the condition determining means 1-2 informs the command executing means 1-3 (instruction 1-10). Instruction from manager (Instruction 1-
1) Command execution means 2N-1 of agent 2N receiving
Executes a command to the system on the agent 2N (instruction 2N-6). At the same time, the manager 1 refers to the condition determination table 1-5 by issuing the command and determines whether or not the condition for issuing the next command is satisfied.
In this way, the scheduling of the event is repeated by repeatedly waiting for an asynchronously occurring event and issuing a command.

Here, a specific flow (cooperation) between the manager 1 and the agents 21 to 2N will be described with reference to FIGS. Manager 1 starts at: Time (6:00) (Fig. 3
Step S11), a command is issued and the agent 2
When the first resident job 21 is started (step S1 in FIG. 3)
2) When the agent 21 receives this command and starts the resident job 21-2 (step S13 in FIG. 4),
The agent 21 generates an activation message for the resident job 21-2 (step S14 in FIG. 4). The manager 1 receives the activation message from the agent 21 and ends (step S15 in FIG. 3). At this time, Manager 1
Waits for an end message from the agent 21, and the start conditions for issuing the next command are time (9:00) and the end of the preceding process (that is, the resident job 21-2 of the agent 21). End message and end message of resident job 22-2 of agent 22)
(Step S21 in FIG. 3).

On the other hand, the manager 1 sets the start condition: time (6:
At 05) (step S16 in FIG. 3), when a command is issued to activate the resident job 22-2 of the agent 22 (step S17 in FIG. 3), the agent 22 receives this command and starts the resident job 22-2. And (FIG. 4
Step S18), the agent 22
-2 is generated (step S1 in FIG. 4).
9). The manager 1 receives the activation message from the agent 22 and ends (step S20 in FIG. 3). The start conditions for the manager 1 to issue the next command are the time (9:00) and the end of the preceding process (that is, the end message of the resident job 21-2 of the agent 21 and the resident job 22-2 of the agent 22). End message)
Step S21 in FIG. 3).

When the agent 21 generates an end message (step S22 in FIG. 4), the manager 1 receives the end message and starts the resident job 23 of the agent 23 (step S23 in FIG. 4). The resident job 23 starts in the agent 23, and generates an activation message for the resident job 23 (step S24 in FIG. 4). This process (steps S22 to S22)
A series of processing between the manager 1 and the agent 21 is completed by 24), and a processing relationship newly occurs between the manager 1 and the agent 23. That is, Manager 1
Performs a workflow (cooperation work flow) between the agent 21 and the agent 23.

The asynchronous message 30 on the right side of FIG.
Generates an end message of the agent 21 asynchronously (start condition), the manager 1 issues a command to start the resident job 23 on the agent 23 according to the start condition, and the end condition is that the resident job 23 is started from the agent 23. Message.

[0017]

As described above, according to the present invention, the following effects can be obtained. According to the present invention, the schedule (processing order allocation) in cooperation with each agent can be performed by centrally managing the schedule of each agent (managed computer) by the manager (management computer). Also, the agent is an effective method for any agent that can execute message output and command issuance regardless of the platform type.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of an embodiment of the present invention.

FIG. 2 is a flowchart of the embodiment.

FIG. 3 is a flowchart showing a specific process of a manager in the embodiment.

FIG. 4 is a flowchart showing a specific process of an agent in the embodiment.

FIG. 5 is a system configuration diagram of a conventional vertical distributed processing system.

[Explanation of symbols]

 1 Manager 21-2N Agent

Claims (2)

[Claims] 1. A computer system comprising a plurality of computers connected via a network to configure a computer system, managing events occurring between the plurality of computers, and performing cooperative operations between the plurality of computers. An interworking task flow management system including: a plurality of managed computers that asynchronously generate various events; a management computer that manages schedules (processing order allocation) of the plurality of managed computers and the events. A plurality of computers, wherein the management computer includes a centralized management unit that centrally manages the event, and proceeds to next processing according to a management result of the centralized management unit. Cooperation workflow management system between. 2. The system according to claim 1, wherein the various events are a time and a resident job message generated by the managed computer.