JPH1078937A - Inter-computer job distribution system and method and recording medium recording job distribution program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To distribute the load of computers by storing the load information on the computers in an external storage for reference and designating a computer having a margin of load to decide the transfer destination of a job. SOLUTION: A self-computer load information collection means 3 collects the load information on its own computer 1 according to designation of a monitoring item, and a self-computer load information storage means 4 stores these collected load information in an external storage 5 that is shared with another computer 13 as the load monitoring information. Then an another- computer load information reference means 6 is started to acquire the load monitoring information that is stored in the storage 5 by the computer 13. All acquired load monitoring information are stored in a load information table 7, and a computer specifying means 8 is started. The means 8 specifies a computer optimum for execution of a job via a job execution computer decision rule table 9 and stores the specified computer in a computer information table 10. Thus, the load is equalized among plural computers and the job can be effectively carried out.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention monitors the load of each computer system in order to distribute the load among a plurality of computer systems sharing a storage device. , A task distribution method, and a recording medium that records a task distribution program.

[0002]

2. Description of the Related Art Japanese Unexamined Patent Application Publication No. Hei 5-101019 discloses that, in a distributed computer system of three or more computers connected by a communication line, each computer has a load of its own computer which has a limit value according to a job execution request to the computer. A technique for adjusting the load on each computer by transmitting the job requested to be executed to another computer by using the load information transmitting / receiving means if it is determined that the number exceeds the threshold. Have been.

Japanese Patent Laid-Open Publication No. 5-151177 discloses a CP of each computer system connected by a communication line.
There is disclosed a technique of monitoring the U usage rate by a load control computer system and adjusting the load on each computer by moving a job to a computer with a low CPU load.

[0004]

However, in these conventional techniques, in order to check load information of each computer in response to a job execution request or to obtain load information of each computer from a specific monitoring computer. In particular, in a situation where there are many job execution requests, there is a problem that the overhead related to the communication processing deteriorates the performance of the computer system and the performance of the load monitoring processing.

Also, a job is moved to another computer based on the load of only the own computer, or the job is moved to a lower load even if the difference between the load of the own computer and the other computer is small. This may cause an inappropriate increase in the load on the computer to which the job has been moved.

An object of the present invention is to equalize the load among a plurality of computers so that each computer can execute a task efficiently.

Another object of the present invention is to move a task to another computer by storing load information of each computer in a shared file so that each computer can monitor each other's load information. When load balancing is performed, the performance of the destination computer does not deteriorate,
Another object of the present invention is to make it possible to select an appropriate destination computer so that the load on each computer is leveled.

[0008]

According to a first aspect of the present invention, a business operation distribution system among a plurality of computers is a computer system including a plurality of computers sharing an external storage device. Own computer load information collecting means for collecting load information relating to at least one resource of the computer at regular intervals, and own computer for storing load information of the own computer collected by the own computer load information collecting means in the external storage device Load information storage means, other computer load information reference means for extracting load information of another computer group from the external storage device, load information of the own computer collected by the own computer load information collection means, and reference to the other computer load information Load information of other computers extracted by the means And a computer specifying unit that specifies a computer having a sufficient load, and when a task executable in any one of the plurality of computers is input to the own computer, the computer specifying unit specifies the computer specified by the computer specifying unit. It is characterized by comprising a destination computer determining means for determining a destination computer to which business is to be moved and a destination computer.

[0009] In a second business distribution system among a plurality of computers according to the present invention, in the business distribution system between the first plurality of computers, the own computer load information collecting means keeps the usage rate of a specific resource of the own computer constant. Collected every time, the computer specifying unit determines that the usage rate of the resource of the own computer exceeds a preset threshold, and the usage rate of the resource is lower than that of the own computer in the other computer group. When there is a low computer, the computer is identified as a computer having a sufficient load.

[0010] In a third business distributed system among a plurality of computers according to the present invention, in the business distributed system among the first plurality of computers, the own computer load information collecting means keeps the usage rate of a specific resource of the own computer constant. Collected every hour, the computer specifying means that the usage rate of the resource of the own computer exceeds a preset threshold value, and the computer usage rate exceeds a predetermined value in the other computer group. When there is a computer with a low resource usage rate, the computer is identified as a computer having a sufficient load.

According to a fourth task distribution system among a plurality of computers of the present invention, in the task distribution system among the first plurality of computers, the own computer load information collecting means keeps the usage rate of a specific resource of the own computer constant. Collected every time, the computer specifying unit determines that the usage rate of the resource of the own computer exceeds a preset threshold, and the usage rate of the resource is lower than that of the own computer in the other computer group. If there is a low computer, the business processing rate, which is the percentage of the tasks executed by the own computer in the group of tasks input to the own computer, is reduced, and the load on the own computer is higher than that of the other computers. Is high, the business processing rate of the own computer is increased, and the destination computer In the case where an executable task is input to the own computer in any of the cases, the computer specifying unit refers to the changed task processing rate of the own computer, and executes a number of tasks corresponding to the task processing rate on the own computer. , And decide to move the rest to another computer.

According to a fifth task distribution system among a plurality of computers of the present invention, in the task distribution system among the first plurality of computers, the own computer load information collecting means is a CPU related to a task executed on the own computer.
The usage rate is collected at regular time intervals, and when the CPU usage rate of the own computer exceeds a preset threshold value, the computer identification unit determines the idle time of the CPU for each of the plurality of computers. The computer is characterized by multiplying by a performance value of the CPU and specifying a computer having the maximum value as a computer having a sufficient load.

According to a sixth aspect of the present invention, there is provided a business distribution system among a plurality of computers, wherein the computer specifying means is one of the other computer groups in the business distribution system among the first plurality of computers. If there is a computer with a low load on the resource, the computer is specified as a computer with a sufficient load.

In a seventh business distributed system between a plurality of computers according to the present invention, in the first business distributed system among a plurality of computers, the own computer load information collecting means uses a CPU executed on the own computer. The CPU usage rate of a high-priority task and the CPU usage rate of the own computer as a whole are sampled at regular intervals, and when the CPU specifying means exceeds the threshold value, It is also determined whether or not the CPU usage rate of a task with a high priority using the CPU exceeds the threshold. If the CPU usage rate does not exceed the threshold, other tasks except for the task with a high priority using the CPU are excluded. The destination computer determining means determines that the task is to be transferred to another computer, and the task is determined to be related to a task with a high priority using the CPU. When the PU utilization even exceeds the threshold, to move all operations to other computers, characterized in that to determine the destination computer determining means.

According to an eighth aspect of the present invention, there is provided a business distribution system among a plurality of computers, wherein the own computer load information collecting means differs in the use rate of a specific resource of the own computer. Sampling for a plurality of sections having a length, wherein the computer identification means determines that the resource usage rate of the short section of the own computer exceeds a preset threshold value of the resource usage rate of the section in the section; If there is a computer in the group of other computers that uses the resource lower than the own computer, the computer is identified as a computer suitable for executing a task with a short elapsed time, and The usage rate of the resource in a long section of the computer exceeds a preset threshold value of the usage rate of the resource in the section, and If there is a computer in the other computer group whose resource usage rate is lower than that of its own computer, the computer is identified as a computer suitable for executing a task with a long elapsed time, and the destination The computer determining means, when a task executable in any one of the plurality of computers is input to its own computer, and when the elapsed time of the task is short, the task specified by the computer identifying means has a short elapsed time. It is determined that the task is to be executed by a computer suitable for executing the task, and when the elapsed time of the task is long, the computer identifying means specifies a computer suitable for executing the task having a long elapsed time. It is characterized by deciding to execute the task.

According to a first method of distributing business among a plurality of computers of the present invention, each of a plurality of computers sharing an external storage device collects load information on at least one resource of the own computer at regular intervals. A computer load information collecting step, each computer storing its own computer load information collected in the own computer load information collecting step in the external storage device, and a respective computer load information storing step; The other computer load information reference step of retrieving the load information of another computer group from the other computer load information of the own computer collected in the own computer load information collection step and the other computer extracted in the other computer load information reference step Compilation Based on the load information of the over data group,
A computer specifying step of specifying a computer having a sufficient load; and when each computer inputs a task executable in any of the plurality of computers to its own computer, the computer specifying the computer specified in the computer specifying step corresponds to the task. And a destination computer determining step of determining a destination computer to be moved.

According to a second method of distributing business among a plurality of computers according to the present invention, in the first method of distributing business among a plurality of computers, the step of collecting load information of the own computer comprises: Collected every hour, the computer identification step determines that the usage rate of the resource of the own computer exceeds a preset threshold, and the usage rate of the resource is lower than the own computer in the other computer group. When there is a low computer, the computer is identified as a computer having a sufficient load.

According to a third method of distributing business among a plurality of computers according to the present invention, in the first method of distributing business among a plurality of computers, the step of collecting the load information of the own computer comprises: Collected every hour, the computer identification step is performed when the usage rate of the resource of the own computer exceeds a preset threshold, and the other computer group is set to a value that is not less than a predetermined value than the own computer. When there is a computer with a low resource usage rate, the computer is identified as a computer having a sufficient load.

According to a fourth method of distributing business among a plurality of computers of the present invention, in the first method of distributing business among a plurality of computers, the step of collecting load information of the own computer comprises: Collected every hour, the computer identification step determines that the usage rate of the resource of the own computer exceeds a preset threshold, and the usage rate of the resource is lower than the own computer in the other computer group. If there is a low computer, the business processing rate, which is the percentage of the tasks executed by the own computer in the group of tasks input to the own computer, is reduced, and the load on the own computer is higher than that of the other computers. Is higher, the business processing rate of the own computer is increased, and the destination computer determination step is performed by the plurality of computers. In the case where a task executable in any of the above is input to the own computer, the computer specifying step refers to the changed task processing rate of the own computer, and executes a number of tasks corresponding to the task processing rate on the own computer. And deciding to move the rest to another computer.

According to a fifth method of distributing business among a plurality of computers according to the present invention, in the first method of distributing business among a plurality of computers, the step of collecting load information of the own computer includes the step of: Collecting, and when the computer identification step includes determining that the CPU usage rate of the own computer exceeds a predetermined threshold value,
It is characterized by multiplying the idle time of the PU by the performance value of the CPU and specifying a computer having the maximum value as a computer having a sufficient load.

According to a sixth method of distributing business among a plurality of computers according to the present invention, in the first method of distributing business among a plurality of computers, the computer specifying step is performed in the other computer group more than one of the own computer. If there is a computer with a low load on the resource, the computer is specified as a computer with a sufficient load.

According to a seventh method of distributing business among a plurality of computers of the present invention, in the first method of distributing business among a plurality of computers, the own computer load information collecting step uses a CPU executed on the own computer. The CPU usage rate of a high-priority task and the CPU usage rate of the entire computer are collected at regular intervals, and the computer identification step is performed when the CPU usage rate of the entire computer exceeds a threshold value. It is also determined whether or not the CPU usage rate of a task with a high priority using the CPU exceeds the threshold. If the CPU usage rate does not exceed the threshold, other tasks except for the task with a high priority using the CPU are excluded. Move the task to another computer in the destination computer decision step, and move to a task with a higher priority using the CPU. That when the CPU use rate exceeds the threshold value, to move all operations to other computers, characterized in that to determine the destination computer determining step.

According to an eighth method of distributing business among a plurality of computers according to the present invention, in the first method of distributing business among a plurality of computers, the step of collecting load information of the own computer differs in a use rate of a specific resource of the own computer. Sampling for a plurality of sections having a length, wherein the computer identifying step determines that the usage rate of the resource in the short section of the own computer exceeds a preset threshold value of the usage rate of the resource in the section; If there is a computer in the group of other computers that uses the resource lower than the own computer, the computer is identified as a computer suitable for executing a task with a short elapsed time, and The usage rate of the resource in a long section of the computer exceeds a preset threshold value of the usage rate of the resource in the section, and If there is a computer in the other computer group whose resource usage rate is lower than that of its own computer, the computer is identified as a computer suitable for executing a task with a long elapsed time, and the destination In the computer determining step, when a task executable in any one of the plurality of computers is input to the own computer, and when the elapsed time of the task is short, the task specified in the computer identifying step is short. It is determined that the task is to be executed on a computer suitable for executing, and if the elapsed time of the task is long, the computer specified in the computer specifying step is a computer suitable for executing the task having a long elapsed time. It is characterized by deciding to execute the task.

A first recording medium according to the present invention is a self-computer load information collecting process for collecting load information relating to at least one resource of the own computer at regular intervals for each of a plurality of computers sharing an external storage device. An own computer load information storing process of storing the load information of the own computer collected in the own computer load information collecting process in the external storage device, and a load of another computer extracting load information of another computer group from the external storage device The information reference process, based on the load information of the own computer collected in the own computer load information collection process and the load information of another computer group extracted in the other computer load information reference process, a computer with a sufficient load Computer specifying processing to specify, and the plurality When a task executable in any of the computers is input to the own computer, a program for performing a destination computer determination process of determining the computer identified in the computer identification process as a destination computer to which the job is to be moved is described. It is recorded.

A second recording medium according to the present invention, in the first recording medium, wherein the own computer load information collecting process is
The usage rate of a specific resource of the own computer is collected at regular time intervals, and the computer specifying process has determined that the usage rate of the resource of the own computer exceeds a preset threshold value,
When there is a computer having a lower usage rate of the resource than the own computer in the other computer group,
The computer is specified as a computer having a sufficient load.

A third recording medium according to the present invention, in the first recording medium, wherein said own computer load information collection processing is performed by:
The usage rate of a specific resource of the own computer is collected at regular time intervals, and the computer specifying process has determined that the usage rate of the resource of the own computer exceeds a preset threshold value,
If there is a computer in the group of other computers that has a lower usage rate of the resource than a value separately determined than the own computer, the computer is identified as a computer with a sufficient load. And

According to a fourth recording medium of the present invention, in the first recording medium, the own computer load information collection processing is performed by:
The usage rate of a specific resource of the own computer is collected at regular time intervals, and the computer specifying process has determined that the usage rate of the resource of the own computer exceeds a preset threshold value,
When there is a computer having a lower usage rate of the resource than the own computer in the other computer group,
In the group of tasks input to the own computer, the task processing rate, which is the ratio of tasks executed on the own computer, is reduced, and when the own computer has a higher load than the other computer group, the The business processing rate is increased, and when the business operation executable in any of the plurality of computers is input to the own computer, the business processing rate of the own computer changed by the computer identification processing is increased. For reference, the number of tasks corresponding to the task processing rate is executed on the own computer, and it is determined that the rest is to be transferred to another computer.

According to a fifth recording medium of the present invention, in the first recording medium, the own computer load information collection processing is performed by:
Collects the CPU usage rate of its own computer at regular intervals,
The computer identification processing is performed by the CPU of the own computer.
If usage exceeds a preset threshold,
For each of the plurality of computers, a CPU idle time is multiplied by a CPU performance value, and a computer having the maximum value is identified as a computer having a sufficient load.

In a sixth recording medium according to the present invention, in the first recording medium, there is a computer in which the computer specific processing has a lower load on any resource than the own computer in the other computer group. If you do
The computer is specified as a computer having a sufficient load.

A seventh recording medium according to the present invention, in the first recording medium, wherein the own computer load information collecting process is
CPU usage rate of high-priority business using CPU executed on own computer and CP of entire own computer
U usage rate is collected at regular time intervals, and when the CPU usage rate of the entire computer exceeds the threshold value, the computer identification processing determines that It is also determined whether the threshold value has exceeded the threshold value.
The transfer destination computer determination process decides to move other tasks except for tasks having higher priority using U to another computer, and the CPU usage rate for tasks having higher priority using CPU is also determined as a threshold value. If the number exceeds the limit, the destination computer determination process determines that all tasks are to be transferred to another computer.

An eighth recording medium according to the present invention, in the first recording medium, wherein the own computer load information collecting process is
The usage rate of a specific resource of the own computer is collected for a plurality of sections having different lengths, and the computer specifying process determines that the usage rate of the resource in a short section of the own computer is the usage rate of the resource in the section set in advance. If the computer has a lower utilization rate of the resource than the own computer in the other computer group in the section, the computer executes a task with a shorter elapsed time. And the resource usage rate of the long section of the own computer exceeds a preset threshold value of the resource usage rate of the section, and the other computer group in the section. If there is a computer whose resource usage rate is lower than that of the own computer in the computer, the computer is suitable for executing a task with a long elapsed time. Computer, and the destination computer determination process is executed in a case where a task executable in any one of the plurality of computers is input to the own computer. It is determined that the task is executed on a computer suitable for executing the task whose elapsed time is short, and if the elapsed time of the task is long, the task whose elapsed time specified in the computer identifying process is long is determined. It is characterized in that it is determined that the task is executed by a computer suitable for execution.

[0032]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, a first embodiment of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 1, the first embodiment of the present invention comprises a self-computer 1 and another computer 13 which operate under program control, and an external storage device 5 shared by them. In the present embodiment, a case will be described in which two computers are used. However, the present invention can be similarly applied to a case in which three or more computers are used.

The own computer 1 (and the other computer 13) are provided with an interval timer 2 for giving an opportunity to execute the program at predetermined time intervals, and an own computer load which is a program which measures the load of the own computer 1 and uses it as load information. An information collection program (hereinafter referred to as own computer load information collecting means) 3 and a self computer load information storage program (hereinafter referred to as own computer load information) which is a program for storing load information of the own computer 1 in the external storage device 5 Storage computer) 4, another computer load information reference program (hereinafter referred to as other computer load information reference means) 6, which is a program for extracting load information of the other computer 13 from the external storage device 5, and own computer 1. And the load information between the A load information table 7 for storing in the data 1, a computer specifying program (hereinafter referred to as a computer specifying means) 8 which is a program for specifying a computer having a sufficient load, and a computer for executing a task. A task execution computer determination rule table 9 for storing rules for determining and load monitoring items necessary for the determination, and a program for storing rules for determining a computer to execute a task in the job execution computer determination rule table 9. A decision rule registration program (hereinafter referred to as decision rule registration means) 12
A computer information table 10 for storing identification information of computers having a sufficient load and business distribution information;
A destination computer determination program (hereinafter, referred to as destination computer determination means) 11, which is a program for determining a destination computer to which a task before schedule is moved with reference to the computer information table 10.

Next, the operation of the first embodiment of the present invention will be described in detail with reference to FIGS.

FIG. 2 is a flowchart showing the flow of the processing of the decision rule registration means 12.

Referring to FIG. 1 and FIG. 2, the own computer 1 uses a determination rule registration unit 12 to determine whether or not to monitor the load and how to monitor the load on the computer. Based on the result, how to distribute the business is stored in the business execution computer determination rule table 9 (step 2).
01).

FIG. 6 is a diagram showing an example of information in the task execution computer decision rule table 9 created by the decision rule registration means 12.

Referring to FIG. 6, a load monitoring execution flag 61 indicating whether or not to perform load monitoring is "ON" indicating that monitoring is to be performed ("OFF" when monitoring is not performed, and the default value is "OFF"). The load monitoring time interval 62, which is the time interval for performing the monitoring operation when performing the load monitoring, is "1000 milliseconds", and indicates whether the CPU usage is monitored or not.
“ON” indicating that the monitoring flag 63 performs monitoring (“OFF” when monitoring is not performed and the default value is “OFF”);
As a criterion for performing the distribution, the CPU load threshold 64 indicating the threshold of the CPU usage rate is “80%”. In the present embodiment, the monitoring item of the load is the CPU usage rate.
Appropriate depending on the contents of the distributed business and the system resources most required for executing the business, such as the BKST usage, the I / O load on the DISK device, the task usage of various service programs, and the network line usage. Can be set.

FIG. 3 is a flowchart showing the flow of processing of the own computer load information collecting means 3 shown in FIG.

Referring to FIG. 1, FIG. 3 and FIG. 6, after registering the decision rule in the task execution computer decision rule table 9, the own computer 1 starts the own computer load information collecting means 3, and at the time of the first start, With reference to the task execution computer determination rule table 9, a load monitoring execution flag 61, which is information on whether or not to execute task distribution, is acquired (step 301).

The self-computer load information collecting means 3 determines that there is no designation to perform business distribution (the load monitoring execution flag 61).
Is "OFF"), the process is terminated as it is, and the job distribution is specified (the load monitoring execution flag 61 is "ON").
Continues the process for collecting the load information of the own computer 1 (step 302).

Then, the own computer load information collecting means 3 operates the load monitoring function (step 302).
Is YES), obtains the load monitoring time interval 62 for performing the operation of monitoring the load of the own computer 1 at regular time intervals from the business execution computer determination rule table 9 (step 303), and specifies the time interval. The interval timer 2 is set (step 304).

When the interval timer 2 is set, if it is a short period of about 1 second to several seconds, there is no problem even if it is set independently in each computer, because there is not much difference in the load information collection time. If it is specified that monitoring is to be performed with a timer of several tens of seconds to one minute or more, a timer is set at a specific time, all computers collect load information at once, and all load information is updated. It is desirable to perform synchronization control as referred to after the synchronization.

After setting the interval timer 2, when the interval timer 2 notifies the own computer load information collecting means 3 that the designated time has elapsed, the own computer load information collecting means 3 executes the task execution. A load monitoring item defining which load item is to be monitored is acquired from the computer determination rule table 9 (step 305).

The own computer load information collecting means 3 sends the own computer 1 according to the specified load monitoring item.
The corresponding load information is collected (step 306), and stored as load monitoring information in the external storage device 5 by the own computer load information storage means 4 (step 307).

The external storage device 5 is a storage device shared by a plurality of computers, and is a storage device such as a memory, an electronic disk, and a magnetic disk. When the processing of step 307 is performed in each computer, the load monitoring information of each computer is stored in the external storage device 5, and the information is updated at each load monitoring interval time. Therefore, the latest load monitoring information of each computer can be obtained by referring to the external storage device 5 from each computer.

Next, the own computer load information collecting means 3
Starts the other computer load information reference means 6, acquires the load monitoring information of the other computer 13 stored by the other computer 13 from the external storage device 5,
(Step 308). After the other computer load information reference means 6 reads the load monitoring information of the other computer 13, the own computer load information collecting means 3
Stores the load monitoring information of all the computers in the load information table 7 (step 309), activates the computer specifying means 8, and notifies that the information of the load information table 7 has been updated (step 310).

FIG. 7 is a diagram showing an example of load monitoring information stored in the external storage device 5 by the own computer load information storage means 4 and load monitoring information obtained from the external storage device 5 by the other computer load information reference means 6. It is.

Referring to FIG. 7, the load monitoring information 75 is the load monitoring information of the own computer 1, and the load monitoring information 75
Reference numeral 6 denotes load monitoring information of the other computer 13. The content of each load monitoring information is, for example, the load monitoring information 75 is such that the computer number 71, which is the identification information of the computer, is “1”, and the information collection time 7, which is the time when the load information is collected
2 indicates “17:30:18 12 ms”, the load monitoring time interval 73 indicates “1000 ms”, and the CPU usage rate within the load monitoring time interval is “92%”. . In the load monitoring information 76, load monitoring data is stored in the same format as the load monitoring information 75.
This indicates that there are two computers that can distribute the work.

FIG. 8 is a diagram showing an example of the load monitoring information of all the computers stored in the load information table 7 by the own computer load information collecting means 3.

Referring to FIG. 8, for example, load monitoring information 8
3 indicates that the computer number 81, which is the identification information of the computer, is “1”, and the CPU usage rate 82, which is the usage status of the CPU of the computer, is “92%”. The load monitoring information 84 stores the load monitoring data in the same format as the load monitoring information 83, and indicates that there are two computers that can distribute the work.

FIG. 4 is a flowchart showing the flow of the processing of the computer specifying means 8 shown in FIG.

Referring to FIG. 1 and FIG. 4, the computer identification means 8 sends the load information table 7
Is notified that the information has been updated, a determination rule for determining an optimum computer for executing the business is acquired from the business execution computer determination rule table 9 (step 401), and the load information table 7 The load monitoring information of all the stored computers is acquired (step 402).

Next, from the obtained decision rule and the load monitoring information of all the computers, the most suitable computer for executing the business is determined (step 403), and the identification number of the determined computer and whether or not to distribute the business are determined. Is stored in the computer information table 10 (step 404). After updating the computer information table 10, the computer specifying unit 8 acquires a load monitoring time interval for performing an operation of monitoring the load of the own computer at regular time intervals from the task execution computer determination rule table 9 (step 405). The interval timer 2 is set by designating the obtained load monitoring time interval (step 406). As a result, the own computer load information collecting means 3 is continuously activated from the interval timer 2 at regular intervals.

FIG. 9 is a diagram showing an example of the information on the distribution of the tasks stored in the computer information table 10 by the computer specifying means 8.

Referring to FIG. 9, for example, a load monitoring flag 91 indicating whether or not the load monitoring function is operating on the own computer 1 indicates "O" indicating that the load is being monitored.
N "(" OFF "when the load is not monitored and the default value is" OFF "), indicating whether to distribute the business among the computers or to execute the business on the own computer without performing the distribution. The distribution execution flag 92 indicates “ON” indicating that the distribution is performed (“OFF” when the distribution of the business is not performed, and the default value is “OFF”), and the business which is the identification information of the computer most suitable for executing the business. Execution computer number 93
Indicates “2” which is the identification number of the other computer 13.

Hereinafter, the processing for determining the contents of the computer information table 10 from the load information table 7 and the task execution computer determination rule table 9 will be described in detail with reference to the examples shown in FIGS. 6, 8 and 9.

Referring to FIGS. 1 and 4, the computer identification means 8 determines that the CPU monitoring flag 63 is "ON" when the job execution computer determination rule table 9 is as shown in FIG. From the CPU usage status is monitored, and since the CPU load threshold 64 is “80%”, the rules for distributing the load when the CPU usage rate exceeds 80% are obtained. Is collected. Next, when the load information table 7 is as shown in FIG. 8 in step 402, the computer identifying unit 8 determines that the CPU usage 82 of the own computer 1 is “92%” and the CPU usage 82 Is "50%".

The computer specifying means 8 first sets the load monitoring flag 91 to “ON” because the load monitoring information is present in the load information table 7, and furthermore, based on the CPU usage rate of the acquired information, the computer 1 Since the CPU usage of the computer 13 exceeds the threshold value of 80%, it is determined that the load on the own computer 1 is high and the tasks should be distributed, and the CPU usage of the other computer 13 is 50%.
Since there is more room than the computer 1, the distributed execution flag 92 is set to “ON”. Then, “2” which is the identification number of the other computer 13 is stored in the task execution computer number 93.

As described above, when the task execution computer determination rule table 9 is as shown in FIG. 6 and the load information table 7 is as shown in FIG. 8, the computer information table as shown in FIG. 10 are created.

FIG. 5 is a flowchart showing the flow of processing of the destination computer determining means 11 shown in FIG.

Referring to FIG. 1 and FIG. 5, when a task which can be executed by either the own computer 1 or the other computer 13 is input to the own computer 1, the destination computer determining means 11 Is determined to be executed. That is, information on the distribution of business is acquired from the computer information table 10 (step 501), and it is determined that the business before the schedule is moved to the computer indicated by the information (step 502).

In the following, when the contents of the computer information table 10 are as shown in FIG. 9, the processing performed by the destination computer determining means 11 to determine the destination computer that executes the task will be specifically described.

In step 501, the destination computer determining means 11 determines from the computer information table 10 that the load monitoring flag 91 is “ON” and the distributed execution flag 9
2 is “ON” and the task execution computer number 93 is “2”
Get that. Next, since the load monitoring flag is “ON” in step 502, the destination computer determining unit 11 knows that the own computer 1 is monitoring the load.
Is "ON", it is recognized that it is more efficient to distribute tasks among the computers because the load on the own computer 1 is high, and to execute the task, the task execution computer number 93 is set to "2". Therefore, it is determined that the task before the schedule is moved to another computer 13 having a low load.

As described above, the load of each computer is monitored according to the rule for determining the distribution of business, the load monitoring information is updated at regular intervals, and the business before the schedule is moved to the computer most suitable for executing the business. Is determined, distributed processing of business is performed among a plurality of computers.

When two or more other computers 13 are present, the above processing is repeated for each of them. If another computer has more CPU utilization than the own computer 1, the other computer 13 is used. Decide to move the pre-schedule business to.

Although the detailed description is omitted in the present embodiment, the work determined to be moved to the other computer 13 by the destination computer determining means 11 is
It is moved to another computer 13 using a known transmission means or a method via the external storage device 5.

Thus, the operation of the first embodiment of the present invention is completed.

According to the first embodiment of the present invention, the self computer load information collecting means 3 is started at regular time intervals, the load monitoring information stored in the external storage device 5 is updated, and the load of another computer is changed. Since the monitoring information is fetched into the own computer and the contents of the computer information table 10 are updated, the overhead required for load monitoring can be reduced as compared with the conventional method of collecting the load of another computer for each business execution request. It has the effect that business can be distributed.

Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

According to the second embodiment of the present invention, the service execution computer determination rule table 9 in the first embodiment includes the CPU usage threshold value and the own computer 1
And a margin defined by the difference between the CPU usage rates of the other computer 13 and the other computer 13. When the CPU usage rate of the own computer 1 exceeds the threshold value, When the CPU usage rate is lower than the CPU usage rate of the own computer 1 and the difference between the CPU usage rates of the other computer 13 and the own computer 1 is equal to or larger than the margin in the business execution computer determination rule table 9, the other computer 13 It is configured to decide to move the business. In the present embodiment, the case of monitoring the CPU usage will be described in detail. However, similarly, the threshold of the usage and a certain margin are set for the other load monitoring items for control. May be performed.

Next, the operation of the second embodiment of the present invention will be described in detail with reference to FIG. 1, FIG. 10 and FIG.

FIG. 11 shows a task execution computer decision rule table 9 created by the decision rule registration means 12.
FIG. 6 is a diagram showing an example of information in the information.

Referring to FIG. 11, a load monitoring execution flag 111 indicating whether or not to perform load monitoring is “ON” indicating that monitoring is to be performed (“OFF” when monitoring is not performed, and the default value is “OFF”). ), When the load monitoring is performed, the load monitoring time interval 112 which is the time interval for performing the monitoring operation is “1000”.
Milliseconds ”, C indicating whether to monitor CPU usage
“ON” indicating that the PU monitoring flag 113 performs monitoring
(If monitoring is not performed, the default value is “OFF” and the default value is “OF
F ”), the CPU load threshold value 64 indicating the threshold value of the CPU usage rate as a criterion for performing the distribution is“ 80% ”, and the CPU usage of the other computer 13 when the work is distributed to the other computer 13 Rate is C of own computer 1
This indicates that the CPU load margin 115 indicating how much lower must be made than the PU usage rate is “10%”.

FIG. 10 is a flow chart showing the flow of processing for business distribution based on the threshold value and the margin of the computer identification means 8.

Referring to FIG. 1 and FIG. 10, the computer specifying means 8 determines whether the CPU usage rate of the own computer 1 exceeds the CPU load threshold value (step 1).
001), when it is determined that it is exceeded, it is determined whether the CPU usage of the other computer 13 is lower than that of its own computer 1 (step 1002), and when it is determined that it is lower, the CPU usage of its own computer 1 is And the difference between the CPU usage rates of the other computers 13 (step 1003).
It is determined whether the difference is equal to or greater than the CPU load margin value (step 1004). If the difference is equal to or greater than the CPU load margin value, the task before the schedule is moved to another computer 13 having a CPU margin. The other computer 1
3 is stored in the computer information table 10 (step 1005), and if the difference is smaller than the value of the CPU load margin, all the tasks are executed by the own computer 1. The identification information of the own computer 1 and the information on business distribution are stored in the computer information table 10 (step 100).
6).

If there are two or more other computers 13, the above processing is repeated for each of them, and the task before the schedule is moved to the computer having the largest margin.

Thus, the operation of the second embodiment of the present invention is completed.

According to the second embodiment of the present invention, the task before the schedule is transferred to the other computer 13 only when the CPU usage rate of the other computer 13 is more than a certain value than the own computer 1. By doing so, it is possible to prevent business from being moved to a computer with a low CPU usage rate, and to avoid performance degradation caused when the business is moved to such a computer. Has the effect that can be.

Next, a third embodiment of the present invention will be described in detail with reference to the drawings.

According to the third embodiment of the present invention, the computer specifying means 8 according to the first embodiment determines whether or not the CPU usage rate of the own computer 1 exceeds the threshold.
The value of the ratio of the business to be processed by the own computer 1 stored in the computer information table 10 is reduced, and the destination computer determination means 11 moves the business to another computer 13 based on the reduced value. Is configured to determine.

Next, the operation of the third embodiment of the present invention will be described in detail with reference to FIG. 1, and FIGS.

FIG. 12 is a flow chart showing the flow of processing for business distribution based on fluctuations in the threshold value and processing rate of the computer specifying means 8.

Referring to FIG. 1 and FIG. 12, the computer specifying means 8 determines whether the CPU usage rate of the own computer 1 exceeds a threshold value (step 1201).
If it exceeds, the CPU usage rate of the other computer 13 is compared with the CPU usage rate of the own computer 1 (step 1202). If the CPU usage rate of the other computer 13 is lower than the CPU usage rate of the own computer 1, That is, if the other computer 13 has more room, the processing rate of the business on the own computer 1 is reduced (step 120).
3) Conversely, when the CPU usage rate of the other computer 13 is higher than the CPU usage rate of the own computer 1, the processing rate of the business of the own computer 1 is increased (step 1204).
The updated processing rate is stored in the computer information table 10 (step 1205). If the work is not distributed, the processing rate of the work of the own computer 1 is set to 100% in the computer information table 10 so that all the work is executed by the own computer 1 (step 1206).

When the processing rate of the business of the own computer 1 is reduced, it is necessary to prevent the load required for executing the business on the own computer 1 as it is and the load required for moving the business from being reversed. It is preferable to provide a lower limit value for executing a certain task on the computer 1. In that case, the processing rate of the business is 100
%.

FIGS. 13 and 14 show examples of the contents of the computer information table 10. FIG.

In FIG. 13, the own computer task processing rate 1
33 indicates “80%”. That is, if 100 tasks are input to the own computer 1 during a certain period, 80 tasks are processed by the own computer 1,
The remaining 20 cases are processed by the other computer 13. When the CPU usage rate of the own computer 1 is higher than the CPU usage rate of the other computer 13, the own computer task processing rate 133 is reduced, and for example, as shown by the own computer task processing rate 143 in FIG. % ".

The destination computer determining means 11 refers to the own computer service processing rate 133 in the computer information table 10 and determines that the own computer 1 executes a number of tasks corresponding to the service processing rate. Is determined to be moved to another computer 13.

When two or more other computers 13 exist, the above processing is repeated for each of them, and the processing rate of business is reduced for a computer with a heavy load.
In a computer with a light load, the business is moved between the computers so as to increase the processing rate of the business.

Thus, the operation of the third embodiment of the present invention is completed.

In the third embodiment of the present invention, it is decided to move a task between computers by increasing or decreasing the processing rate of a task set in each computer according to the load state of each computer. With this configuration, it is possible to set an appropriate processing rate of the business for each computer, and it is possible to operate each computer in an appropriate load state.

Next, a fourth embodiment of the present invention will be described in detail with reference to the drawings.

In the fourth embodiment of the present invention, the load information table 7 in the first embodiment stores the CPU usage rate of each computer and an index (CPU performance value) indicating the CPU performance of the computer. , Computer specifying means 8
However, when the CPU usage rate of the own computer 1 exceeds the threshold value, the load information table 7 is referred to, and for each computer, the ratio of the time when the CPU is idle is multiplied by the CPU performance value of the computer, It is configured to determine that the business is executed on the computer having the largest value.

Next, the operation of the fourth embodiment of the present invention will be described in detail with reference to FIG. 1 and FIGS.

FIG. 15 is a flow chart showing the flow of business distribution processing based on the product of the CPU idle time and the CPU performance value of the computer specifying means according to the fourth embodiment of the present invention.

Referring to FIG. 1 and FIG. 15, the computer specifying means 8 determines whether the CPU usage rate of the own computer 1 exceeds the CPU load threshold value (step 1).
501), if exceeded, the CP of each computer
The ratio of the CPU idle time is calculated from the U usage rate and multiplied by the CPU performance value of each computer (step 150).
2) The multiplied product is compared between the own computer 1 and the other computer 13 (step 1503). If the product of the other computer 13 is larger than the product of the own computer 1, the other computer 13 performs processing by the CPU. It is determined that there is enough capacity, and information for moving the job to another computer 13 is stored in the computer information table 10 (step 1504). Conversely, when the product of the other computer 13 is smaller than the product of the own computer 1 Stores information for executing a task on the computer 1 in the computer information table 10 (step 1505).

Note that this CPU performance value is determined by the number of instructions executed per unit time (MIPS (Million Inst
function per second), or a relative performance value indicating how much CPU processing performance each computer has as compared to a performance value of a specific reference computer. That is, any performance value relating to the CPU can be used as long as the index is the same between the computers.

FIG. 16 shows the business execution computer decision rule table 9 created by the decision rule registration means 12.
FIG. 6 is a diagram showing an example of information in the information.

Referring to FIG. 16, CPU weighting flag 165 indicates "ON" indicating that comparison is to be performed using a product obtained by multiplying the CPU performance value by the ratio of CPU idle time ("OFF" when weighting is not performed). (Default value is "OFF")
Accordingly, the computer identification unit 8 determines to perform the processing of step 1502 described above.

FIG. 17 shows an example of the load monitoring information stored in the external storage device 5 by the own computer load information storage means 4 and the load monitoring information obtained from the external storage device 5 by the other computer load information reference means 6. It is.

Referring to FIG. 17, for example, an item of CPU performance value 175 is stored for each computer.
75 is “200” and computer number 171 is “2”
Indicates that the CPU performance value 175 is “30”.

FIG. 18 is a diagram showing an example of information on loads of all computers stored in the load information table 7 by the own computer load information collecting means 3.

Referring to FIG. 18, for example, load information 18
4, the computer number 181 is “1”, the CPU usage rate 182 is “90%”, and the CPU performance value 183 is “20”.
0 ”, and the load information 185 includes the computer number 1
81 is “2”, CPU usage rate 182 is “50%”, and C
This indicates that the PU performance value 183 is “30”.

FIG. 19 is a table showing an example of the values calculated in the above-described step 1502 of the computer specifying means 8 when the data as shown in FIG. 18 is stored in the load information table 7. FIG.

Referring to FIG. 19, the computer number “1” has a higher CPU usage rate, but the CPU performance value is larger than the computer number “2”. Is determined to have sufficient processing capacity of the CPU, and the task is determined to move to the computer number “1”.

If there are two or more other computers 13, the above processing is repeated for each of them, and it is determined that the task is to be transferred to the computer with the most CPU processing capacity.

Thus, the operation of the fourth embodiment of the present invention is completed.

In the fourth embodiment of the present invention, the CP of each computer is determined by the product of the idle time of the CPU and the CPU processing performance value.
By estimating the margin of the processing capacity of U, even between computers with different CPU processing performances, it is possible to move the work and distribute the load so that the margins of the processing capacity of the CPUs of the respective computers become equal. Has an effect.

Next, a fifth embodiment of the present invention will be described in detail with reference to the drawings.

According to the fifth embodiment of the present invention, the own computer load information collecting means 3 in the first embodiment collects information on a plurality of load monitoring items in addition to the CPU usage rate. The computer specifying means 8 determines whether a plurality of load monitoring items exceed a threshold based on a predetermined priority, compares the load monitoring items with the priority, and determines a computer with a light load. Then, it is configured to decide to move the business to the computer.

Next, the operation of the fifth embodiment of the present invention will be described in detail with reference to FIG. 1, FIG. 20 and FIG.

FIG. 20 is a flow chart showing a flow of a process of business distribution by a plurality of load monitoring items of the computer specifying means according to the fifth embodiment of the present invention. In this embodiment, it is assumed that two load monitoring items, that is, a CPU usage rate and a memory usage rate, are set, and the monitoring priority is determined in the order of the CPU usage rate and the memory usage rate.

Referring to FIG. 1 and FIG. 20, the computer specifying means 8 determines whether the CPU usage rate of the own computer 1 exceeds the CPU load threshold value (step 2).
001), if the threshold is exceeded, the other computer 13 determines whether there is room in the CPU (step 2002). If there is no room, the other computer 13 compares the next load monitoring item. It is determined whether the memory usage rate of the computer 1 exceeds the threshold value (step 2003). If the memory usage rate exceeds the threshold value, it is determined whether the other computer 13 has more memory (step 2004). If there is no room, the identification information of the own computer 1 is stored in the computer information table 10 in order to execute the task on the own computer 1 (step 2005). If it is determined in step 2002 or step 2004 that the other computer 13 has room, the identification information of the other computer 13 is stored in the computer information table 10 (step 2006).

Although FIG. 20 shows the case where there are two load monitoring items, the above-described judgment is repeated similarly when there are three or more load monitoring items, and there is room for resources of other computers. In this case, the identification information of the other computer is stored in the computer information table 10.

FIG. 21 shows a task execution computer decision rule table 9 created by the decision rule registration means 12.
FIG. 6 is a diagram showing an example of information in the information.

Referring to FIG. 21, CPU monitoring flag 2
13 and the CPU load threshold value 214, a memory monitoring flag 215 indicating that the memory usage rate is to be monitored is "ON" indicating that monitoring is to be performed (if monitoring is not to be performed, "OFF" is set and the default value is " OFF "), indicating that the memory load threshold value 216 indicating the threshold value of the memory usage status as the criterion for performing the distribution is" 50% ",
This shows that the priority of monitoring the load is determined in the order of the CPU and the memory. When three or more load monitoring items are set, the memory load threshold 2 shown in FIG.
Subsequent to 16, the contents of another monitoring item and the threshold value thereof are stored.

If there are two or more other computers 13, the above processing is repeated for each of them, and it is determined that the business is moved to the computer with the most room.

As described above, the operation of the fifth embodiment of the present invention is completed.

In the fifth embodiment of the present invention, when a plurality of load monitoring items are set, even when it is determined that one business distribution item does not perform business distribution, another load monitoring item is determined. In the case where there is a computer which can afford, there is an effect that business can be distributed.

Next, a sixth embodiment of the present invention will be described in detail with reference to the drawings.

According to the sixth embodiment of the present invention, the own computer load information collecting means 3 in the first embodiment can perform both high-priority tasks using the CPU and low-priority tasks using the CPU. The CPU usage rate of each of the included tasks is calculated, and the computer identifying unit 8 determines the priority of using the CPU when the CPU usage rates of all the tasks in the own computer 1 exceed the threshold value. It is determined whether or not the CPU usage rate related to a task having a high task also exceeds the threshold value. If the CPU usage rate does not exceed the threshold value, it is determined that a task with a high priority using the CPU is not moved. If the CPU usage rate of a task having a higher priority to use the job exceeds the threshold value, the destination computer determination means 11 decides to move all tasks. It is configured to.

Next, the operation of the sixth embodiment of the present invention will be described in detail with reference to FIG. 1, and FIGS.

FIG. 24 shows a task execution computer decision rule table 9 created by the decision rule registration means 12.
FIG. 6 is a diagram showing an example of information in the information.

Referring to FIG. 24, a priority-based CPU monitoring flag 245 indicating whether or not to monitor the CPU usage rate of only a task having a high CPU usage priority indicates "ON" (monitoring is performed). If not performed, “OFF” and the default value is “OFF”). At this time, the own computer load information collecting means 3 is connected to a normal whole CPU.
The load monitoring information is collected, and the CPU usage time of a specific service having a high CPU usage priority such as an online service is calculated to obtain a CPU usage rate.

FIG. 22 is a flow chart showing the flow of a business distribution process in which the computer specifying means 8 monitors the CPU use priority according to the sixth embodiment of the present invention.

Referring to FIG. 1 and FIG. 22, the computer identification means 8 is a
It is determined whether the U usage rate exceeds the threshold (step 2).
201) If the threshold value is not exceeded, the identification number of the own computer 1 and the task distribution information are stored in the computer information table 10 so that all the tasks are executed by the own computer 1 (step 2206). If the CPU usage rate of the entire business exceeds the threshold, the CPU usage rate is smaller than the CPU usage rate of the own computer 1.
That is, it is determined whether or not there is another computer 13 which has room for the CPU (step 2202). If there is no other computer 13 which has room for the CPU, the computer information table 10 is set so that all tasks are executed by the own computer 1. Then, the identification number of the own computer 1 and the information of the business distribution are stored in the storage (step 2206). If there is another computer 13 which has room for the CPU, it is determined whether or not the CPU usage rate relating to the high-priority operation using the CPU exceeds the threshold (step 2203). In the computer information table 10, the identification number of each computer and the information on the distribution of tasks are stored in such a manner that tasks with a high CPU usage priority are executed by the own computer 1 and tasks with a low priority are moved to another computer 13. Is stored (step 2205). vice versa,
If the CPU usage rate of only the high-priority tasks using the CPU exceeds the threshold value, the computer information table 10 stores information about the other computers 13 so that all the tasks are moved to the other computers 13. The identification number and the information on business distribution are stored (step 2204).

FIG. 25 is a diagram showing an example of information in the computer information table 10 created by the computer identification means 8.

Referring to FIG. 25, for example, a normal CPU
The normal business distribution execution flag 252 indicating whether or not to move the business having the use priority is “ON” indicating that the business is to be moved (“OFF” when the business is not to be moved, and the default value is “ OFF "), a priority task distribution execution flag 25 indicating whether or not to move a task of high priority using the CPU.
3 is “OFF” which does not move the business (“ON” when the business is moved, and the default value is “OFF”);
The normal job execution computer number 254 which is the identification information of the computer most suitable for executing the job having the PU use priority is “2” which is the identification number of the other computer 13.
The priority job execution computer number 255, which is the identification information of a computer that is optimal for executing a job with a high priority using the CPU, indicates “1”, which is the identification number of the own computer 1. .

FIG. 23 is a flow chart showing the flow of business distribution processing for monitoring the CPU use priority of the destination computer determining means 11 according to the sixth embodiment of the present invention.

Referring to FIG. 1 and FIG. 23, the destination computer determining means 11 refers to the computer information table 10 and sets the normal business distribution execution flag 252 to ON.
It is determined whether or not it is OFF (step 2301).
(Step 2305). If the normal business distribution execution flag 252 is ON, then step 2
After 302, it is determined that a task having a normal CPU use priority is to be executed by another computer 13 (step 2304). If the CPU use priority of the task to be moved is high (step 2302), it is further determined whether or not the priority task distribution execution flag 253 is ON (step 2303). Is determined to be executed by the own computer 1 (step 2305), and when the flag is ON, it is determined to be executed by another computer 13 (step 2304).

If there are two or more other computers 13, the above processing is repeated for each of them, and it is determined that the business is to be moved to the computer with the largest load.

As described above, the operation of the sixth embodiment of the present invention is completed.

In the sixth embodiment of the present invention, it is determined whether or not to move a service by calculating the CPU usage rate for each CPU usage priority.
When the PU usage rate exceeds the threshold, the CP
If the CPU usage rate of a task with a high U use priority does not exceed the threshold, only a task with a low CPU use priority is moved to another computer. Typically,
Responsibilities are emphasized in tasks having a high CPU use priority, so that the tasks having a high CPU use priority are executed by the own computer and are not moved to another computer, so that the overhead associated with the transfer processing is reduced. Can be eliminated, and the performance is improved accordingly.

Next, a seventh embodiment of the present invention will be described in detail with reference to the drawings.

According to the seventh embodiment of the present invention, when the own computer load information collecting means 3 in the first embodiment collects the CPU usage rate, the computer load information collecting means 3 in two or more periods is used. The CPU usage rate is obtained, the computer specifying means 8 determines whether the CPU usage rate exceeds a threshold value determined for each period, and the destination computer determination means 11 determines whether the elapsed time is short. The service is configured to determine whether to move the task based on the CPU usage rate for a short period, and to determine whether to move the task based on the CPU usage rate for a long period for a task with a long elapsed time. I have.

Next, the operation of the seventh embodiment of the present invention will be described in detail with reference to FIG. 1, and FIGS.

FIG. 29 shows the task execution computer decision rule table 9 created by the decision rule registration means 12.
FIG. 6 is a diagram showing an example of information in the information.

Referring to FIG. 29, a load monitoring execution flag 291 indicating whether or not to perform load monitoring is "ON" indicating that monitoring is to be performed ("OFF" when monitoring is not performed, and the default value is "OFF"). ), A time interval for performing a monitoring operation when performing load monitoring, a short-term load monitoring time interval 292 for performing short-term monitoring is “1000 milliseconds”, and a long-term load for performing long-term monitoring. If the monitoring time interval 293 is "2
“0000 milliseconds”, “ON” indicating that the CPU monitoring flag 294 indicating whether or not to monitor the use state of the CPU is to be monitored (“OFF” when the monitoring is not performed, and the default value is “OFF”). Short-term CP
Short-term CPU load threshold 29 indicating U usage threshold
5 is “80%”, and the long-term C
Long-term CPU load threshold value 2 indicating the threshold value of PU usage rate
96 indicates “50%”.

In this embodiment, the case where the CPU is monitored will be described in detail. However, other load monitoring items may be similarly controlled by setting monitoring time intervals having different periods. .

FIG. 26 is a flowchart showing the flow of processing of monitoring the CPU load of the own computer load information collecting means 3 during different periods according to the seventh embodiment of the present invention.

Referring to FIGS. 1 and 26, when receiving the notification from the interval timer 2, the own computer load information collecting means 3 sets the CPU monitoring flag 294 to "O".
In the case of "N", the CPU usage rate in a short period is obtained (step 2601). Next, the own computer load information collecting means 3 determines whether there is a definition for obtaining a long-term CPU usage rate (step 2602). The CPU usage rate during the period is obtained (step 2603), and if there is no definition, the process is terminated.

FIG. 30 is a diagram showing an example of information in the computer information table 10 created by the computer identification means 8.

Referring to FIG. 30, a load monitoring flag 301 indicating whether or not to perform load monitoring is "ON" indicating that monitoring is being performed (if monitoring is not performed, "OFF" and the default value is "OFF"). "OFF"), "ON" indicating that the short-term task distribution execution flag 302 indicating that the task with a short elapsed time is to be transferred to the other computer 13 is to be moved ("OFF" if not moved, and the default value is "OFF") , The long-term work distribution execution flag 303 indicating that the work with a long elapsed time is to be transferred to the other computer 13 does not move.
F ”(“ ON ”to move, default is“ OF ”
F "), when a task with a short elapsed time is moved, the short-term task execution computer number 304 indicating which computer executes the task indicates the identification number" 2 "of the other computer 13, and the task with a long elapsed time is indicated. When moving,
The long-term job execution computer number 305 indicating which computer executes the job is "1" which is the identification number of the own computer.

FIG. 27 is a flow chart showing the flow of a job distribution process in the case where the CPU load of the computer specifying means 8 is monitored for different periods according to the seventh embodiment of the present invention.

Referring to FIG. 1 and FIG. 27, the computer identification means 8 executes the computer 1 to create the computer information table 10 as shown in FIG.
It is determined whether or not the short-term CPU usage rate exceeds the threshold value of the short-term CPU usage rate in step 2701 (step 2701).
02 is turned off so that work with a short elapsed time is not moved to another computer 13 (step 270).
4). Conversely, if the threshold value is exceeded, the CPU usage rate of the other computer 13 for a short period of time is reduced
It is determined whether or not the CPU is lower than that of the other computer 13 (step 2702).
Does not exist, the short-term business distributed execution flag 302
To turn off the work with a short elapsed time
No. 3 is not moved (step 2704).
If there is another computer 13 which has room for the CPU, the short-time task distribution execution flag 302 is turned on to move tasks with a short elapsed time to the other computer 13 (step 2703). The identification number is stored in the computer information table 10 (step 2705).

Next, it is determined whether or not a long-term CPU usage rate is determined (step 2706). If not, the process is terminated. When the long-term CPU usage rate is obtained, the long-term CPU usage rate of the own computer 1 becomes
It is determined whether or not the threshold value of the long-term CPU usage rate has been exceeded (step 2707). (Step 2711). Conversely, if the threshold is exceeded, it is determined whether the other computer 13 has a lower long-term CPU usage rate than the own computer 1 and has room for the CPU (step 2708). If a certain other computer 13 does not exist, the long-time business distributed execution flag 303 is turned off so that a business with a long elapsed time is not moved to the other computer 13 (step 2711). If there is a computer with sufficient CPU, turn on the long-term business distributed execution flag 303
Then, the task with a long elapsed time is moved to another computer 13 (step 2709), and the identification number of the computer with a low CPU load is stored in the computer information table 10 (step 2710).

FIG. 28 is a diagram showing C in different periods of the destination computer determining means 11 according to the seventh embodiment of the present invention.
9 is a flowchart illustrating a flow of a process of business distribution when PU load monitoring is performed.

Referring to FIG. 1 and FIG. 28, the destination computer determining means 11 refers to the computer information table 10 created as described above for the work before schedule input to the own computer 1. hand,
The following processing is performed.

First, it is determined whether the elapsed time of the entered work is short or long (step 2801). If the elapsed time is short, it is determined whether or not the short-term task execution flag 302 is ON (step 2802). If it is ON, the computer designated by the short-term task execution computer number 304 stored in the computer information table 10 is used. It is determined that the job is to be moved to (step 2803).
When the short-term business distribution execution flag 302 is OFF,
It is determined that the job is to be executed by the own computer 1 without moving (step 2805).

On the other hand, if the elapsed time of the input job is long, it is determined whether the long-term job distribution execution flag 303 is ON (step 2804). If it is ON, the flag is stored in the computer information table 10. It is determined that the task is to be transferred to the computer designated by the long-term task execution computer number 305 (step 2806). If the long-term task distribution execution flag 303 is OFF, it is determined that the task is not to be moved and executed by the own computer 1 (step 2805).

In this embodiment, two monitoring periods are set when monitoring the CPU. However, even if the load monitoring items are other than the CPU, there are three or more measuring periods. However, only the above processing is repeated, and the basic processing contents are the same.

If there are two or more other computers 13, the above processing is repeated for each of them, and it is determined that the business is to be moved to the computer with the largest load.

As described above, the operation of the seventh embodiment of the present invention is completed.

In the seventh embodiment of the present invention, when monitoring a load, two or more measurement periods are set, monitoring data of a short period is used for a job with a short elapsed time, and monitoring is performed for a business with a long elapsed time. Uses long-term monitoring data. When it is determined whether or not to move a task with a long elapsed time based on the monitoring data for a short period of time, due to the temporary high load state of the task, the task whose load is not so high on average is classified as another task. There is a possibility of performing inappropriate business distribution processing such as deciding to move to a computer, but by configuring as in this embodiment,
This has the effect of preventing such inappropriate distribution of duties.

Next, an eighth embodiment of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 1, an eighth embodiment of the present invention is similar to the first embodiment, except that
And a recording medium 14 for recording a program for causing the other computer 13 to execute the processes according to the above-described first to seventh embodiments.

The program is read from the recording medium 14 into the own computer 1 and the other computer 13 and controls the operations of the own computer 1 and the other computer 13. The own computer 1 and the other computer 13
By the control of the program, the processing by the own computer 1 and the other computer 13 in the first to seventh embodiments, that is, the own computer load information collecting means 3,
Own computer load information storage means 4, other computer load information reference means 6, computer identification means 8, destination computer determination means 11, and determination rule registration means 1
The same processing as in step 2 is executed.

Thus, the processing of the eighth embodiment is completed.

[0160]

As described above, according to the present invention,
By leveling the load across multiple computers,
This has the effect that tasks can be executed efficiently.

Also, by storing the load information of each computer in a shared file, each computer can monitor each other's load information. This has the effect of preventing the performance of the computer system from deteriorating.

[Brief description of the drawings]

FIG. 1 is a block diagram showing first to eighth embodiments of the present invention.

FIG. 2 is a flowchart showing a flow of processing of a decision rule registration unit 12 according to the first embodiment of the present invention.

FIG. 3 is a flowchart showing the flow of processing of the own computer load information collecting means 3 according to the first embodiment of the present invention.

FIG. 4 is a flowchart illustrating a processing flow of a computer specifying unit 8 according to the first embodiment of the present invention.

FIG. 5 is a flowchart showing a processing flow of a destination computer determining unit 11 according to the first embodiment of the present invention.

FIG. 6 is a diagram showing an example of information in a task execution computer determination rule table 9 according to the first embodiment of the present invention.

FIG. 7 is a diagram illustrating an example of load monitoring information stored in an external storage device 5 according to the first embodiment of this invention.

FIG. 8 is a diagram showing an example of information in a load information table 7 according to the first embodiment of the present invention.

FIG. 9 is a diagram illustrating an example of information in a computer information table 10 according to the first embodiment of the present invention.

FIG. 10 is a flowchart illustrating a flow of a process of business distribution based on a threshold and a margin of the computer identification unit 8 according to the second embodiment of the present invention.

FIG. 11 is a diagram illustrating an example of information in a task execution computer determination rule table 9 according to the second embodiment of this invention.

FIG. 12 is a flowchart illustrating a flow of a process of business distribution due to a change in a threshold value and a processing rate of the computer specifying unit 8 according to the third embodiment of the present invention.

FIG. 13 is a diagram showing an example of the contents of a computer information table 10 according to the third embodiment of the present invention.

FIG. 14 is a diagram showing an example of the contents of a computer information table 10 after a change according to the third embodiment of the present invention.

FIG. 15 is a flowchart illustrating a flow of a task distribution process based on a product of a CPU idle time and a performance value of the computer identification unit 8 according to the fourth embodiment of the present invention.

FIG. 16 is a diagram illustrating an example of information in a task execution computer determination rule table 9 according to the fourth embodiment of the present invention.

FIG. 17 is a diagram illustrating an example of load monitoring information stored in an external storage device 5 according to the fourth embodiment of the present invention.

FIG. 18 is a diagram illustrating an example of information in a load information table 7 according to the fourth embodiment of the present invention.

FIG. 19 is a diagram illustrating, in a table form, a calculation example for determination by a computer specifying unit 8 according to the fourth embodiment of the present invention.

FIG. 20 is a flowchart showing a flow of a process of distributing a task by a plurality of load monitoring items of the computer identification unit 8 according to the fifth embodiment of the present invention.

FIG. 21 is a diagram illustrating an example of information in a task execution computer determination rule table 9 according to the fifth embodiment of the present invention.

FIG. 22 is a flowchart illustrating a flow of a business distribution process in which the computer specifying unit 8 monitors the CPU use priority according to the sixth embodiment of the present invention.

FIG. 23 is a flowchart showing a flow of a business distribution process for monitoring the CPU use priority of the destination computer determining unit 11 according to the sixth embodiment of the present invention.

FIG. 24 is a diagram illustrating an example of information in a task execution computer determination rule table 9 according to the sixth embodiment of the present invention.

FIG. 25 is a diagram illustrating an example of information in a computer information table 10 according to the sixth embodiment of the present invention.

FIG. 26 is a flowchart showing a flow of a process of monitoring the CPU load of the own computer load information collecting means 3 in different periods according to the seventh embodiment of the present invention.

FIG. 27 is a flowchart illustrating a flow of a task distribution process in the case where the computer identification unit 8 monitors the CPU load in different periods according to the seventh embodiment of the present invention.

FIG. 28 is a flowchart illustrating a flow of a task distribution process when the destination computer determining unit 11 monitors CPU loads for different periods according to the seventh embodiment of the present invention.

FIG. 29 is a diagram illustrating an example of information in a job execution computer determination rule table 9 according to the seventh embodiment of the present invention.

FIG. 30 is a diagram illustrating an example of information in a computer information table 10 according to the seventh embodiment of the present invention.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 own computer 2 interval timer 3 own computer load information collecting means 4 own computer load information storage means 5 external storage device 6 other computer load information reference means 7 load information table 8 computer specifying means 9 business execution computer determination rule table 10 computer information table 11 destination computer determination means 12 determination rule registration means 13 other computer 14 recording medium

Claims (24)

[Claims] 1. A computer system comprising a plurality of computers sharing an external storage device, wherein each of the plurality of computers collects load information on at least one resource of the own computer at regular intervals. Load information collection means, own computer load information storage means for storing the load information of the own computer collected by the own computer load information collection means in the external storage device, and load information of another computer group from the external storage device The other computer load information reference means to be extracted; and a load margin based on the load information of the own computer collected by the own computer load information collection means and the load information of another computer group extracted by the other computer load information reference means. Computer to identify computers with Determining means, and a computer which is executable by any one of the plurality of computers is input to its own computer. And a means for distributing tasks between a plurality of computers. 2. The self-computer load information collecting means collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying means determines that the usage rate of the resource of the own computer is set in advance. If the threshold value is exceeded and there is a computer in the group of other computers that has a lower usage rate of the resource than the own computer, it is determined that the computer is a computer with a sufficient load. 2. The business distribution system according to claim 1, wherein the business operation is distributed between a plurality of computers. 3. The self-computer load information collecting means collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying means determines that the usage rate of the resource of the own computer is set in advance. If the threshold value is exceeded and there is a computer in the group of other computers that has a lower usage rate of the resource than a value determined separately from the own computer, the computer is a computer with a sufficient load. 2. The business distribution system between a plurality of computers according to claim 1, wherein the system is specified as being present. 4. The self-computer load information collecting means collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying means determines that the usage rate of the resource of the own computer is set in advance. If there is a computer that exceeds the threshold and has a lower usage rate of the resource than the own computer in the other computer group, the computer is executed by the own computer in the business group input to the own computer. The business processing rate, which is the percentage of business, is reduced, and when the load of the own computer is higher than that of the other computer group, the business processing rate of the own computer is increased. If a task executable on any of the plurality of computers is input to the own computer, the computer identification unit changes the own computer. Referring to business processing rate of Yuta, a number of operations corresponding to the task processing rate running on its own computer,
2. The system according to claim 1, wherein a decision is made to move the rest to another computer. 5. The own computer load information collecting means collects the CPU usage rate of the own computer at regular time intervals, and the computer specifying means sets the CPU usage of the own computer.
If usage exceeds a preset threshold,
2. The computer according to claim 1, wherein the computer having the maximum value is multiplied by the idle time of the CPU and the performance value of the CPU for each of the plurality of computers, and the computer having the maximum value is identified as a computer having a sufficient load. Business distribution system between multiple computers. 6. If the computer specifying means includes a computer having a lower load on any of the resources than the own computer in the other computer group, the computer specifying means sets the computer as a computer having a sufficient load. 2. The business distribution system between a plurality of computers according to claim 1, wherein the system is specified as being present. 7. The self-computer load information collection means collects, at regular intervals, a CPU usage rate of a high-priority task using a CPU executed on the own computer and a CPU usage rate of the entire self-computer. The computer identification means is the C of the entire computer;
When the PU usage rate exceeds the threshold value, it is determined whether the CPU usage rate related to a job having a higher priority to use the CPU also exceeds the threshold value. In this case, the destination computer determining means determines that the other tasks except for the tasks having the higher priority using the CPU are to be transferred to another computer, and the CPU usage rate for the tasks having the higher priority using the CPU is also determined. 2. The business distribution system according to claim 1, wherein when the value exceeds the value, the destination computer determination means determines that all the business is to be transferred to another computer. 8. The own computer load information collecting means collects the usage rate of a specific resource of the own computer for a plurality of sections having different lengths, and the computer specifying means collects the resource usage information of a short section of the own computer. When the usage rate exceeds a preset threshold value of the usage rate of the resource in the section, and in the section, there is a computer having a usage rate of the resource lower than that of the own computer in the other computer group. The computer identifies the computer as a computer suitable for executing a task with a short elapsed time, and the usage rate of the resource in a long section of the own computer is a threshold of the usage rate of the resource in the predetermined section. If the value of the resource exceeds the value and there is a computer having a lower usage rate of the resource than the own computer in the other computer group in the section, The computer is identified as a computer suitable for executing a task with a long elapsed time, and the destination computer determining means, when a task executable on any of the plurality of computers is input to its own computer, If the elapsed time of the task is short, it is determined that the task is executed by a computer suitable for executing the task whose elapsed time specified by the computer identifying means is short, and if the elapsed time of the task is long, 2. The business distribution system according to claim 1, wherein it is determined that the business specified by the computer specifying means is executed by a computer suitable for executing the business having a long elapsed time. 9. A self-computer load information collecting step in which each of a plurality of computers sharing an external storage device collects load information related to at least one resource of the own computer at regular time intervals; A self-computer load information storing step of storing the load information of the own computer collected in the computer load information collecting step in the external storage device; and other computer loads in which each computer retrieves the load information of another computer group from the external storage device. An information referencing step, wherein each computer has a margin of load based on the load information of the own computer collected in the own computer load information collecting step and the load information of another computer group extracted in the other computer load information referring step. With A computer specifying step of specifying a computer; and when each of the computers inputs a task executable in any of the plurality of computers to its own computer, a destination to which the computer specified in the computer specifying step moves the task. And a destination computer determining step of determining a destination computer. 10. The own computer load information collecting step collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying step determines whether the usage rate of the resource of the own computer is set in advance. If the threshold value is exceeded and there is a computer in the group of other computers that has a lower usage rate of the resource than the own computer, it is determined that the computer is a computer with a sufficient load. The method according to claim 9, wherein the task is distributed among a plurality of computers. 11. The own computer load information collecting step collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying step determines whether the usage rate of the resource of the own computer is set in advance. If the threshold value is exceeded and there is a computer in the group of other computers that has a lower usage rate of the resource than a value determined separately from the own computer, the computer is a computer with a sufficient load. 10. The method according to claim 9, wherein the identification is performed.
A method for distributing business among a plurality of computers described in the above. 12. The self-computer load information collecting step collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying step determines whether the usage rate of the resource of the own computer is set in advance. If there is a computer that exceeds the threshold and has a lower usage rate of the resource than the own computer in the other computer group, the computer is executed by the own computer in the business group input to the own computer. The business processing rate, which is the percentage of business, is reduced, and when the load of the own computer is higher than that of the other computer group, the business processing rate of the own computer is increased. When a task executable on any of a plurality of computers is input to the own computer, the computer identification step is performed. And determining, by referring to the changed business processing rate of the own computer, the number of business operations corresponding to the business processing rate on the own computer and moving the rest to another computer. 9. The method for distributing business among a plurality of computers according to item 9. 13. The own computer load information collecting step collects the CPU usage rate of the own computer at regular time intervals, and the computer specifying step includes the C
If the PU usage rate exceeds a preset threshold value, the CPU of each of the plurality of computers is
10. The method according to claim 9, further comprising: multiplying the idle time by the performance value of the CPU and identifying a computer having the maximum value as a computer having a sufficient load. 14. The computer identifying step, if there is a computer having a lower load on any of the resources than the own computer in the other computer group, the computer is selected as a computer having a sufficient load. 10. The method according to claim 9, wherein the task is specified. 15. The own computer load information collecting step collects, at regular intervals, a CPU usage rate of a service having a high priority using a CPU executed on the own computer and a CPU usage rate of the entire self computer. The computer identifying step, when the CPU usage rate of the entire computer exceeds a threshold value,
It is also determined whether the CPU usage rate of a task having a high priority to use the CPU exceeds the threshold value. If the CPU usage rate does not exceed the threshold value, other tasks except for the task having a high priority to use the CPU Move to another computer in the destination computer determination step,
When the CPU usage rate of a task having a high priority to use a PU also exceeds the threshold, the moving destination computer determining step determines that all tasks are to be moved to another computer. The method for distributing business among a plurality of computers according to claim 9. 16. The own-computer load information collecting step collects the usage rate of a specific resource of the own computer for a plurality of sections having different lengths, and the computer specifying step includes the step of collecting the resource of a short section of the own computer. When the usage rate exceeds a preset threshold value of the usage rate of the resource in the section, and in the section, there is a computer having a usage rate of the resource lower than that of the own computer in the other computer group. The computer identifies the computer as a computer suitable for executing a task with a short elapsed time, and the usage rate of the resource in a long section of the own computer is a threshold of the usage rate of the resource in the predetermined section. Value, and there is a computer in the other computer group in which the usage rate of the resource is lower than that of the own computer in the section. In the above, the computer is specified as a computer suitable for executing a task with a long elapsed time, and the destination computer determining step is executed when a task executable on any of the plurality of computers is input to the own computer. In the case, when the elapsed time of the task is short, it is determined that the task is executed on a computer suitable for executing the task whose elapsed time specified in the computer identifying step is short, and the elapsed time of the task is determined. 10. The business between a plurality of computers according to claim 9, wherein when the time is long, it is determined that the business specified by the computer specifying step is executed by a computer suitable for executing the business having a long elapsed time. Dispersion method. 17. A self-computer load information collection process for collecting load information relating to at least one resource of the self-computer at regular intervals for each of a plurality of computers sharing an external storage device; Own computer load information storage processing for storing the load information of the own computer collected in the processing in the external storage device; other computer load information reference processing for extracting load information of another computer group from the external storage device; A computer identification process for identifying a computer having a sufficient load based on the load information of the own computer collected in the load information collection process and the load information of another computer group extracted in the other computer load information reference process; No need for multiple computers In the case where the executable task is input to the own computer, a program for performing a destination computer determining process of determining the computer identified in the computer identifying process as a destination computer to which the task is to be moved is recorded. Recording medium characterized by the above-mentioned. 18. The self-computer load information collecting process collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying process determines that the usage ratio of the resource of the own computer is set in advance. If the threshold value is exceeded and there is a computer in the group of other computers that has a lower usage rate of the resource than the own computer, it is determined that the computer is a computer with a sufficient load. The recording medium according to claim 17, characterized in that: 19. The self-computer load information collecting process collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying process determines that the usage ratio of the resource of the own computer is set in advance. If the threshold value is exceeded and there is a computer in the group of other computers that has a lower usage rate of the resource than a value determined separately from the own computer, the computer is a computer with a sufficient load. 18. The recording medium according to claim 17, wherein the recording medium is specified. 20. The self-computer load information collecting process collects a usage rate of a specific resource of the own computer at regular intervals, and the computer specifying process determines that the usage ratio of the resource of the own computer is set in advance. If there is a computer that exceeds the threshold and has a lower usage rate of the resource than the own computer in the other computer group, the computer is executed by the own computer in the business group input to the own computer. If the task processing rate, which is the percentage of the task, is reduced, and the own computer has a higher load than the other computer group, the task processing rate of the own computer is increased. When a task executable on any of the plurality of computers is input to the own computer, the computer identification process is changed by the own computer. Referring to business processing rate of computers, the number of operations corresponding to the task processing rate running on its own computer,
18. The recording medium according to claim 17, wherein a decision is made to move the rest to another computer. 21. The self-computer load information collecting process collects the CPU usage rate of the self-computer at regular intervals, and the computer specifying process includes the CPU of the self-computer.
If usage exceeds a preset threshold,
18. The computer according to claim 17, further comprising: multiplying the idle time of the CPU and the performance value of the CPU for each of the plurality of computers, and specifying a computer having the maximum value as a computer having a sufficient load. recoding media. 22. If the computer identification process includes a computer having a lower load on any of the resources than the own computer in the other computer group, the computer may be replaced by a computer having a sufficient load. 18. The recording medium according to claim 17, wherein the recording medium is specified. 23. The self-computer load information collection process collects, at regular intervals, a CPU usage rate of a high-priority task using a CPU executed on the own computer and a CPU usage rate of the entire self-computer. The computer identification processing is performed by the C of the entire computer.
When the PU usage rate exceeds the threshold value, it is determined whether the CPU usage rate related to a job having a higher priority to use the CPU also exceeds the threshold value. In the moving destination computer determination processing, the other computer except for the high-priority task using the CPU is moved to another computer, and the CPU usage rate for the high-priority task using the CPU is also determined. 18. The recording medium according to claim 17, wherein when the value exceeds the value, the destination computer determination processing determines to move all tasks to another computer. 24. The self-computer load information collecting process collects the usage rate of a specific resource of the own computer for a plurality of sections having different lengths, and the computer specifying process collects the resource of a short section of the own computer. When the usage rate exceeds a preset threshold value of the usage rate of the resource in the section, and in the section, there is a computer having a usage rate of the resource lower than that of the own computer in the other computer group. The computer identifies the computer as a computer suitable for executing a task with a short elapsed time, and the usage rate of the resource in a long section of the own computer is a threshold of the usage rate of the resource in the predetermined section. If the value exceeds the value and there is a computer having a lower usage rate of the resource than the own computer in the other computer group in the section, If the computer is identified as a computer suitable for executing a task with a long elapsed time, the destination computer determination processing is executed when a task executable on any of the plurality of computers is input to the own computer. If the elapsed time of the task is short, it is determined that the task specified by the computer identifying process is to be executed by a computer suitable for executing the task with a short elapsed time. 18. The recording medium according to claim 17, wherein it is determined that the business specified by the computer specifying process is executed by a computer suitable for executing the business having a long elapsed time.