JP2001175627A - Decentralized processing computer system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a decentralized processing computer system which can properly and easily cope with the stagnation of one computer in such a case and bring out the maximum processing capability of the whole system. SOLUTION: If, for example, a decentralized process computer PC4 has stagnation while the system is in operation, a decentralized process execution means 3 of a centralized management computer PC1 selects a decentralized process computer PC3 as the best computer which processes a task HHH of the decentralized process computer PC3 instead by referring pieces of information in a management information storage part 2. Then the decentralized processing execution means 3 transmits task program data 8 of the task HHH from the management information storage part 2 to the decentralized processing computer PC3.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a so-called centralized management type distributed processing computer system.

[0002]

2. Description of the Related Art In a computer system constructed on a network, a form in which a plurality of tasks are distributed and processed by a plurality of computers is often adopted in order to improve the total throughput of the system. In such a distributed processing computer system, when a load on a computer increases and stagnation occurs in the processing of the computer, a part of the task of the computer is transferred to another computer without stagnation. This ensures smooth operation of the entire system. In this case, although which computer takes over the task of the computer in which the stagnation has occurred is set in advance or determined based on a predetermined program, it can be determined by operator intervention.

[0003]

However, in the above-mentioned prior art, the computer to be replaced is determined based on fixed information or based on the experience or intuition of the operator. It was difficult to maximize the overall processing capacity. for that reason,
Even if another computer takes over some of the tasks of the computer in which the stagnation has occurred, if the capacity of the computer that took over is insufficient or the load is too large, stagnation will also occur in the computer that takes over. In some cases, smooth operation of the entire system could not be ensured.

[0004] Further, when a task of one computer is taken over by another computer, it is necessary to define a task processing program in both computers with the same contents. However, depending on the task, this work becomes complicated and a lot of labor is required. Had to spend some time.

The present invention has been made in view of the above circumstances, and it is possible to optimally and easily cope with a case where a stagnation occurs in a certain computer, thereby maximizing the processing capability of the entire system. It is intended to provide a simple distributed processing computer system.

[0006]

According to a first aspect of the present invention, a plurality of computers are interconnected on a network, and a plurality of tasks are distributed by the plurality of computers. In the distributed processing computer system for processing, at least one of the plurality of computers monitors a management information storage unit in which management information on each task and each computer is stored, and monitors the operation status of each computer. Distributed processing executing means for referring to each information of the management information storage unit and allocating a task processed by the computer to another computer when stagnation occurs in the processing of the computer. It is characterized by the following.

According to a second aspect of the present invention, in the first aspect, the information stored in the management information storage section is:
Computer / task allocation information on tasks to be processed by each computer, computer resource management information indicating the capabilities of each computer,
Task resource management information including data on resources and processing timing required for processing of each task, and each task program data indicating the processing content of each task; the distributed processing execution means includes a computer for which the stagnation has occurred; When assigning a task to the other computer, the task program data related to the task is read from the management information storage unit and transmitted to the other computer.

According to a third aspect of the present invention, in the first or second aspect of the present invention, the management information storage section includes task priority information on the priority of each task,
The distributed processing execution means, when a stagnation occurs in any of the computers and no other computers have a margin to process the task of the computer in which the stagnation occurred, the task priority information By referring to one computer processing a relatively low-priority task, suspending or delaying the processing, or instructing the transfer to another computer, to allow processing capacity, The task of the computer in which the stagnation has occurred is assigned to the one computer.

According to a fourth aspect of the present invention, in the invention of any one of the first to third aspects, the distributed processing execution means assigns the task of the computer in which the stagnation has occurred to the other computer or the one computer. And the task processing is executed, and the computer / task assignment information in the management information storage unit is updated to the contents after the assignment.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the distributed processing computer system according to this embodiment. In this figure, a central management computer PC1 and distributed processing computers PC2 to PC4 are connected via a switching hub 1. By using this switching hub 1, the central management computer PC1 can simultaneously transmit and receive data to and from a plurality of distributed processing computer systems. In this embodiment, the tasks executed by each computer are, as shown, the central management computer PC1 is the task AAA and the task BBB, and the distributed processing computer PC2 is the task CC
C, task DDD, distributed processing computer PC3 is task EEE, task FFF, task GGG, and distributed processing computer PC4 is task HHH, task III, and task JJJ.

Among the computers, only the central management computer PC1 has the management information storage unit 2 and the distributed processing execution unit 3. The management information storage unit 2 stores computer / task allocation information 4, computer resource management information 5, task resource management information 6, task priority information 7, and task program data 8 in a file format. The distributed processing execution means 3 monitors the operation status of each of the computers PC1 to PC4, and when any one of the computers stagnates, the computer refers to the information in the management information storage unit 2 to perform processing. This process executes a process of allocating the task to another computer having sufficient processing capacity.

The computer / task assignment information 4 indicates the correspondence between computers and tasks executed by the computers, and information on the correspondence is described in a table as shown in FIG. The computer resource management information 5 is
It shows data about the main resources of each computer. For example, a table as shown in FIG.
"Relative CPU performance (CPU speed)" and "Memory capacity"
Etc. are described. Task resource management information 6
4 shows data on resources necessary for processing a task and other necessary items. For example, FIG.
In the table as shown in the table, "required memory capacity", "required CPU time", "execution timing", "permissible processing time" and the like are described. Also, the task priority information 7
Is a numerical value of the degree to which the task should be processed with priority, and as shown in FIG. 4, is described together with the table in which the task resource management information is described (in the example of FIG. 4, the numerical value is Smaller things have higher priority.) The task program data 8 is data on software indicating a task processing procedure, and is defined with the same contents as the task program data originally possessed by the computers PC1 to PC4.

Next, the operation of the distributed processing computer system of FIG.
In the following, an example will be described in which the task HHH is to be processed by another computer because the stagnation has occurred in the task HHH. When the distributed processing execution means 3 that has monitored the operation status of each computer detects that a stagnation has occurred in the distributed processing computer PC4, the distributed processing computer 3 has computer / task allocation information 4, computer resource management information 5, Task resource management information 6,
And the task priority information 7.

According to the computer resource management information shown in FIG. 3, among the computers other than PC4, the centralized management computer PC1 has the highest relative CPU performance, followed by the distributed processing computer PC3 and the distributed processing computer PC3. The order is PC2. Therefore, from the viewpoint of considering the relative CPU performance, the task HHH may be assigned to the central management computer PC1. However, referring to the task priority information 7, as shown in FIG. 4, the priorities of the task AAA and the task BBB processed by the central management computer PC1 are "20" and "10", respectively. Task HH
Since the priority of H is higher than “40”, it is impossible to assign the task HHH to the central management computer PC1.

Therefore, the task HHH processing is assigned to either the PC2 or the PC3.
According to the computer resource management information shown in FIG.
2 has a memory capacity of 512 MB and a required memory capacity of 1 MB.
When processing the 28 MB task HHH, there is almost no room. On the other hand, the memory capacity of the computer PC3 is 1024 MB, and even if the task HHH is assigned, the memory capacity still has room. Therefore, the distributed processing execution means 3 is currently executing the distributed processing computer P
Distributed processing computer PC for task HHH being processed by C4
3 is assigned. Then, the distributed processing execution means 3 sends the task HH to the distributed processing computer PC4.
H, and instruct the distributed processing computer PC to stop processing.
Instruct the task 3 to execute the task HHH, read the task program data 8 relating to the task HHH from the management information storage unit 2, and transmit this to the distributed processing computer PC3.

As a result, the distributed processing computer PC4 only needs to execute the processing of the task III and the task JJJ. Therefore, the occurrence of the stagnation of the distributed processing computer PC4 is eliminated, and the task HHH is allocated to the distributed processing computer PC having a sufficient resource.
3 is processed. Thereafter, the distributed processing execution means 3
5, the contents of the computer / task allocation information 4 stored in the management information storage unit 2 are changed from the distributed processing computer PC4 to the task HHH, as shown in FIG.
3 (the contents of the table in FIG. 4 are also updated based on this). Then, when the allocation of tasks is changed thereafter, the distributed processing execution unit 3 refers to the updated information.

The task HHH allocation processing by the distributed processing execution means 3 compares the computer resource management information of each computer with the task resource management information required for the processing of the task HHH, and further considers the priority of each task. After that, the distributed processing computer PC3, which is the most suitable computer, performs the shoulder processing. Therefore, the stagnation state of the distributed processing computer PC4 can be eliminated, and the smooth operation of the entire system can be ensured. Further, the distributed processing execution means 3 only needs to transmit the task program data relating to the task to the computer that has been assigned to the processing of a certain task each time. It is no longer necessary to define task program data to be processed with the same contents.

Next, the task HHH of the distributed processing computer PC4 where the stagnation has occurred as described above is transferred to the distributed processing computer PC3.
Next, a case will be described in which the stagnation occurs in the distributed processing computer PC2 while the processing is performed in the same manner as described above.

The distributed processing execution means 3 transmits the task CCC of the distributed processing computer PC2 in which the stagnation has occurred to the central management computer P2.
Attempt to replace C1 or the distributed processing computer PC3. However, the arithmetic processing capabilities of both the central management computer PC1 and the distributed processing computer PC3 are insufficient for executing the processing of the task CCC. Therefore, the distributed processing execution means 3 sends the task BBB to the central management computer PC1.
Dispatch task AAA with lower priority than (CP
The data of the task AAA occupying the RAM of U is temporarily saved to the hard disk), and the process of the task CCC is executed. This task A
By the dispatch of the AA, the central management computer PC1
Arithmetic processing capacity necessary for executing the processing of the task CCC can be secured.

Next, the distributed processing execution means 3 reads the task program data relating to the task CCC from the management information storage unit 2 and outputs this to a task processing execution unit (not shown) in the central management computer PC1. As a result, the processing of the task BBB and the task CCC can be performed by the central management computer PC.
1 will be executed. In this case, the dispatched task AAA can be transferred to the distributed processing computer PC3 for processing. Further, when the processing of either the task AAA or the task BBB is completed in the central management computer PC1, , Again centralized management computer P
The processing may be performed in C1.

Thereafter, the distributed processing execution means 3 stores the contents of the computer / task allocation information 4 stored in the management information storage unit 2 as shown in FIG. When the content is updated to the content after the change to the management computer PC1 and the task assignment is changed thereafter, the updated information is referred to.

As described above, in the distributed processing computer system according to the present invention, while the task HHH of the computer PC4 in which the stagnation has occurred is taken over by the computer PC3, the stagnation occurs in the computer PC2. Regarding the arithmetic processing capability of the computer PC3 for executing the task CCC, if none of the computers PC1 and PC3 is insufficient, the task AAA of the lower priority of the tasks of the computer PC1 is dispatched. Alternatively, the processing is performed by the computer PC3. That is, according to the distributed processing computer system according to the present invention, flexible processing corresponding to the actual operation status of the system is performed, and the processing capacity of the entire system is maximized under limited resources. be able to.

In the above-described embodiment, an example is described in which the number of centralized management computers provided with the management information storage unit and the distributed processing execution means is one. However, when the number of computers in the system is large, Two or more management computers may be installed.

[0024]

As described above, according to the present invention, at least one of a plurality of computers includes a management information storage unit and distributed processing execution means, and the distributed processing execution means The operation status of the computer is monitored, and if a stagnation occurs in one of the computers, a task that the computer has processed is referred to by referring to each information in the management information storage unit. Since the assignment is made to the computers, it is possible to optimally and easily cope with a stagnation of a certain computer, and to realize a distributed processing computer system capable of maximizing the processing capacity of the entire system. be able to.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a distributed processing computer system according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a data configuration example of computer / task assignment information 4 in FIG. 1;

FIG. 3 is an explanatory diagram showing a data configuration example of computer resource management information 5 in FIG. 1;

FIG. 4 is an explanatory diagram showing a data configuration example of task resource management information 6 and task priority information 7 in FIG. 1;

FIG. 5 is an explanatory diagram showing a data configuration example of computer / task assignment information 4 when a task of a computer in which stagnation has occurred is assigned to another computer.

FIG. 6 is an explanatory diagram showing a data configuration example of computer / task assignment information 4 when a task of a computer in which stagnation has occurred is assigned to another computer.

[Explanation of symbols]

 PC1 Centralized management computer PC2 to PC4 Distributed processing computer 1 Switching hub 2 Management information storage unit 3 Distributed processing execution means 4 Computer / task allocation information 5 Computer resource management information 6 Task resource management information 7 Task priority information 8 Task program data

Claims (4)

[Claims] 1. A distributed processing computer system in which a plurality of computers are mutually connected on a network and a plurality of tasks are distributed and processed by the plurality of computers, wherein at least one of the plurality of computers includes: A management information storage unit in which management information on each task and each computer is stored; and an operation status of each computer is monitored. If a stagnation occurs in the processing of any of the computers, each of the management information storage units Distributed processing execution means for referring to information and allocating a task being processed by the computer to another computer. 2. The information stored in the management information storage unit includes:
Computer / task allocation information on tasks to be processed by each computer, computer resource management information indicating the capabilities of each computer,
Task resource management information including data on resources and processing timing required for processing of each task, and each task program data indicating the processing content of each task; the distributed processing execution means includes a computer for which the stagnation has occurred; 2. The distributed processing according to claim 1, wherein when assigning a task to the other computer, the task program data related to the task is read from the management information storage unit and transmitted to the other computer. 3. Computer system. 3. The management information storage section includes task priority information on the priority of each task, and the distributed processing execution means determines that a stagnation occurs in one of the computers and the other If the computer also does not have enough time to process the task of the computer in which the stagnation has occurred, the computer refers to the task priority information, and the computer that is processing the task with a relatively low priority is sent to one computer. On the other hand, by giving a command to suspend or delay the processing, or to shift to another computer, a margin is provided in the processing capacity, and the task of the computer in which the stagnation has occurred is assigned to this one computer. 3. The distributed processing computer system according to claim 1, wherein: 4. The distributed processing execution means allocates the task of the computer in which the stagnation has occurred to the other computer or the one computer to execute the task processing, and executes the computer / task of the management information storage unit. 4. The distributed processing computer system according to claim 1, wherein the allocation information is updated to the content after the allocation.