JP2002297407A - Job schedule management device, job schedule management method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a job schedule management device that performs the precise scheduling for a job. SOLUTION: A job schedule management tool 11, regarding the job on which an execution request is issued, abstracts a piece of processing target data 17 to produce test data 14, and makes a client execute a job for processing the data. Depending on the time taken for the execution of the test data 14 and the size of the processing target data 17, the predicted time for the job processing is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a client / server system that performs parallel processing, and more particularly, to a job schedule management apparatus, a job schedule management method, and a program that execute scheduling of a job to be processed by a client.

[0002]

2. Description of the Related Art In a client / server system in which distributed processing is performed by a plurality of clients, data to be processed is processed by one or more application programs such as arithmetic processing, and as a result, a process of generating a file is performed by actual processing. Before starting the process, scheduling for distribution to each client machine in the form of JOB is performed. It is generally known that by performing this scheduling well, the processing time of the entire system can be reduced and the processing efficiency can be increased.

FIG. 19 shows parallel processing in a general client / server system. The client / server system targeted in this specification is a J /
A plurality of clients 102-1 to 102 executing OB
-N, and a job schedule server 101 that manages the schedule of job execution in each client.

[0004] A job schedule server 101 allocates a job instructed by an operator to each client 101 to perform scheduling, and sends a job execution command at the scheduled time. Each of the clients 102 receiving the job executes the specified job, and when the process is completed, notifies the schedule server 101 of the job execution end notification to the schedule server 101.

[0005]

In the scheduling management performed in such a general client / server system, the processing time of a job to be performed by the client 102 must be confirmed by actually executing the job. I don't know exactly. Therefore, there is a problem that the client 102 cannot complete the processing within the time allocated to the JOB, or the processing is completed much earlier than expected, resulting in a time loss.

In view of the above problems, the present invention can obtain a highly accurate predicted execution time of a job to be allocated to each client, and can efficiently process the job using the predicted execution time. It is an object to provide a schedule management method and a program.

[0007]

It is assumed that the present invention is directed to a client / server system comprising a plurality of clients and a server. The server includes a test data generation unit, a test processing execution unit, and a JOB prediction time calculation unit.

[0008] Test data is generated from test data generating means and data to be processed which is processed by the job instructed to execute. A test processing execution unit that causes one of the plurality of clients to execute the job with the test data as processing target data.

The JOB prediction time calculation means processes the data to be processed by the JOB based on the time required for the client to process the test data, the size of the data to be processed, and the size of the test data. Obtain the estimated time in the case of doing. Each of the clients includes a job execution unit that processes the processing target data in response to a job execution instruction from the server and returns a processing result to the server.

The server performs test data generation means for generating test data from data to be processed which is processed by the job instructed to execute, and performs test processing for causing the client to process the test data by the job. JOB prediction time calculation means for calculating a prediction time when the processing target data is processed by the JOB from the test processing result, the size of the processing target data, and the size of the test data. You can do it.

With this configuration, the server performs a test process using test data generated from the data to be processed.
Since the JOB prediction time calculation means obtains the prediction time using the result, a highly accurate prediction time can be obtained.
Further, the server further comprises a storage unit for storing information used for obtaining the predicted time, which is obtained from the result of the test processing, and when the information for the job for which the execution instruction has been issued is stored in the storage unit, The JOB prediction time calculation means may be configured to calculate the prediction time using the information.

With this configuration, when a test process has been performed on a job for which execution has been instructed in the past, the server can calculate the predicted time without performing a new test process. According to the present invention, the JOB prediction time calculation means calculates the JOB processing prediction time using the test data generated from the data to be processed. Therefore, highly accurate prediction can be performed, and the processing efficiency of the entire system can be improved.

[0013]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a hardware configuration of a computer (information processing apparatus) constituting a client and a computer constituting a JOB schedule server in the present embodiment.

As shown in FIG.
It comprises a CPU 1, an input device 2, a display device 3, a printing device 4, a storage device 5, a RAM 6, a transmission control unit 7, and the like. here,
The CPU 1 is a control unit that controls the entire computer. The input device 2 is a keyboard having a plurality of keys for inputting various data for JOB scheduling and execution instructions. In addition, this
A pointing device such as a mouse can be included. The display device 3 is a CRT or a liquid crystal display, and the printing device 4 is a printer.

The storage device 5 stores a program, data, and the like in advance, and the storage device 5 includes a recording medium 8 readable by the CPU 1 such as a magnetic or optical recording medium or a semiconductor memory. It is. The recording medium 8 includes a portable medium such as a CD-ROM and a memory card, and a fixed medium such as a hard disk. In addition, some or all of the programs and data stored in the recording medium 8 are transmitted from a server or client to a WAN,
It may be configured to receive data from the transmission control unit 7 via a network such as a LAN, and the recording medium 8 may be a recording medium of a server or a client constructed on the network.

The RAM 6 is a work memory used by the CPU 1 during execution of a program. Further, the transmission control unit 7 includes a network such as a WAN and a LAN,
It transmits and receives data to and from another device connected via the Internet.

FIG. 2 is a functional block diagram of a job schedule management device realized on the job schedule server according to the present embodiment. The job schedule server stores the job schedule management tool 1 in the storage device 5.
1, schedule file 12, past row log data 1
3. The test data 14 of each client, the system information 15 of each client, and the JOB script 16 are stored and held.

Note that the job schedule server does not necessarily include all the elements shown in FIG.
The held configuration differs in each of the eleventh embodiments. For example, in the first embodiment, the past log data 1
3 and the system information 15 of each client.

The job schedule management tool 11
It manages the execution schedule of the OB, and the next job to be executed from the schedule file 12 and the job
, And sends a JOB execution command to the target client at the execution time set in the schedule file 12. At this time, the job schedule management tool 11 determines the start time of the job,
The OB, target client name, execution target data and its size are recorded in the log data 13. When the JOB execution ends on the client side, the JOB schedule management tool 11 sends a JOB execution end notification to the JOB schedule server, and sends the time (JOB end time),
The executed JOB and the target client name are described in the past log data 13.

Further, a job schedule management tool 11
Predicts the job execution time. As a method of this prediction, a test is performed as described later, a time required for the client to process the target JOB is predicted from the test result, and this is notified to the operator, and the job schedule registration is automatically performed based on the predicted value. Do it. At this time, JOB
The schedule management tool 11 uses not only the test result but also information such as the past processing time of the target JOB, the result of a test performed in the past, and the system information 15 of each client.

The schedule file 12 is a file in which the processing schedule of each job is recorded. J
The OB schedule management tool 11 refers to the schedule file 12 and issues a job execution command to the client. The job execution log data 13 is a log file that records an execution log of each job. The test data 14 is test data for performing a test for estimating a processing time to be described later, and is generated from processing target data that is actually processed by JOB. The system information 15 for each client records information indicating the processing capability of each client, for example, a CPU speed, HDD access speed, capacity, memory capacity, and the like in a database.

The job script 16 is a script file in which the contents of the job are described. FIG. 3 shows an example of the job script 16. In the JOB-A script shown in FIG. 9A, the job data is processed by the application DataConvert1.exe → DataRena
JO to process using me1.exe → Printout.exe
B. Also, the JO shown in FIG.
In the BB script, JOB-B transfers the data to be processed to the applications DataCreate1.exe and DataCreate2.ex.
e are executed in parallel, and the job is processed in parallel. In the job-C script shown in FIG. 10C, the job-C executes only the application DataCreate1.exe to process the data to be processed. This indicates that the job is to be processed. In the following description, J is used for simplification.
J running one application like OB-C
A description will be given for OB.

FIG. 4 is a block diagram showing the configuration of the client according to this embodiment. The client shown in the figure stores the job management tool 21 and the job script 2 in the storage device 5.
2, the application 23, the input data 24, and the output data 25 are stored and held.

Job management tool 21 on the client side
Receives the job execution command transmitted from the job schedule server, activates the application 23 corresponding to the instructed job, and processes the input data 24 from the job schedule server. Then, after the job processing is completed, the job management tool 21 transmits a job execution completion notification to the job schedule server together with the output data 25 as the processing result.

The job script 22 stores the designated job
This is a script file in which the processing contents of B are described. The application 23 is an application program executed by the client. Among the plurality of applications 23, the one designated by the job script 22 is executed by the job schedule management tool 21. The input data 24 is processing target data input and instructed from the job schedule server, and the output data 25 is a file in which a job execution result is stored. Next, a first embodiment using a system constituted by the client / JOB schedule server configured as described above will be described.

In the first embodiment, before allocating a JOB to a client, a test for generating test data 14 from the processing target data 17 and causing the client to process the test data 14 is executed.
Obtain the estimated processing time. FIG. 5 is a flowchart showing processing by the job schedule management tool 11 of the job schedule server according to the first embodiment.

This flowchart is for explaining a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management apparatus on a computer constituting the server according to the first embodiment of the present invention. It is. This program runs on the CPU
1 is stored in the recording medium 8 in the form of a readable program code. However, it is not necessary to store all functions in the recording medium 8.
A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In the figure, first, the job schedule management tool 11 prompts the operator to input a job name of a job to be scheduled, data 17 to be processed, and a full path of a file for storing data generated as a processing result (step S1). ), The job schedule server enters an input waiting state (step S2, No).

When all the items specified in step S1 are input by the operator (step S2, Yes), the job schedule server acquires the size of the processing target data 17 as step S3.
Further, test data is generated from the processing target data 17.

Next, in step S4, the job schedule server sends the test data generated in step S3 to the client, and uses this test data to execute J
OB is executed and time measurement is started. When the client receives a job completion notification, the measurement of the time is stopped in step S5, and the estimated time when the data actually used in the job is processed is calculated based on the measurement time and the ratio of the test data to the actual data size. Calculate, display the result on the screen, record and save.

The operator schedules the job based on the estimated time of the job processing displayed on the display device 3 and sets the job in the schedule file 12. For example, processing target data 1 as shown in FIG.
In the case of performing the processing for obtaining the processing result as shown in FIG. 7B from FIG. 7, this prediction time is calculated by extracting a part from FIG. 7A and generating test data as shown in FIG. Then, JOB is executed on the test data, and the effective time is measured.

Assuming that the size of the processing target data 17 in FIG. 6A is d1, the size of the test data is d2, and the time required for processing the test data is t2, the processing target data 17 is processed. Time t1 of JOB processing at the time
Can be expressed as t1 = t2 × (d1 / d3), and this result is the expected time of the JOB processing. As the actual JOB execution prediction data, the processing time per unit size of the data is used and stored. For example, if it takes 5 minutes to process 10 Kbytes of data, the job execution expected data of this job is 30
Seconds / K bytes.

In the first embodiment, the effective time of the entire JOB is predicted by extracting a part of the processing target data 17 and executing it as test data. This will give you accurate predictions,
The time allocated as the processing time when performing the scheduling does not run short, or the processing is completed earlier than the allocated time, so that there is no vacancy in the client, and the processing efficiency of the entire system can be improved. At this time, the size of the test data may be arbitrarily determined. In the case of this configuration, the test data can be made larger when it is desired to make a more accurate prediction, and can be made smaller when it is desired to obtain a predicted value sooner.

Next, a second embodiment will be described. In the second embodiment, the results of past tests performed in the first embodiment are stored in a database and stored, and the same JOB is processed using the stored JOB execution prediction data. Find the estimated time of

For example, based on the result of a test performed in the past, 30 seconds /
When a value of K bytes is recorded and 1 Mbytes of processing target data is processed using JOB-A, the predicted time of the JOB processing is 30 × 1024 (K bytes) = 30720 seconds = 512 minutes = 8 Data of time 32 minutes is obtained.

FIG. 7 is a flowchart showing processing by the job schedule management tool 11 of the job schedule server according to the second embodiment. This flowchart is for describing a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting the server according to the second embodiment of this invention. This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In the figure, the job schedule management tool 11 prompts the operator to input the job name of the job to be scheduled, the data 17 to be processed, and the full path of the file for storing the data generated as the processing result (step). S11), the job schedule server enters an input waiting state (step S12, No).

Then, the job schedule management tool 1
1 indicates that the operator has entered all the items specified in step S11 (step S12, Ye
s), a test for the target job is performed in the past as step S13, and the job for the target job is executed.
The database is searched to determine whether or not the B execution expected data is held. If the result is not held (step S13, No), the process proceeds to step S14.
Steps S14 to S5 corresponding to steps S3 to S5 in FIG.
The test is executed by performing the processing of S16, the estimated time of the JOB processing obtained as a result is displayed, recorded and saved, and the processing is terminated.

In step S13, if the job execution prediction data obtained from the past test results is stored in the job schedule server for the target job (step S13, Yes), the job schedule management tool 11 Processing target data 1 as S17
7 and the expected job execution data of the job, and as step S18, the predicted time of the current JOB processing is calculated from the size of the processing target data 17 obtained in step S17 and the expected execution data of the job. Is displayed on the display device 3 (step S19), and the process ends.

As a result, the operator proceeds to step S1
Scheduling is performed based on the predicted times displayed in steps S6 to S19 and set in the schedule file 12. In the second embodiment, when a test is already performed on a job to be scheduled, a predicted time of the job processing is calculated from predicted job execution data obtained from the result at this time. Perform scheduling. Thus, in the second embodiment, the number of times of performing the test can be reduced, and the processing overhead can be reduced.

Next, a third embodiment of the present invention will be described. In the third embodiment, the operator inputs the start time of a job, and the job schedule management tool 11 automatically sets the job scheduling using the start time.

For example, if the execution start time of JOB-A for processing data of 1 Mbytes is registered as 0:00, the result of the past execution and the result of the test performed on JOB-A is assumed as 30% as the expected execution data. If data of second / Kbyte is recorded, 30 × 1024 (Kbyte) = 30720 seconds = 5 as the estimated processing time of JOB-A when 1 Mbyte of processing target data is processed
Since 12 minutes = 8 hours and 32 minutes is obtained, the job schedule management tool 11 sets the end time of the job to 8:32.
Schedule file 12
Automatically register with.

FIG. 8 is a flowchart showing processing by the job schedule management tool 11 of the job schedule server according to the third embodiment. This flowchart is for describing a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting the server according to the third embodiment of this invention. This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. Part or all of it is WA
It may be realized by receiving from the transmission control unit 7 via a network such as N or LAN or the Internet.

In the figure, the job schedule management tool 11 includes a target job name, processing target data 17,
B. In addition to the full path of the file storing the data generated as the processing result, the user is prompted to input a job execution start time (step S21), and the job schedule server waits for input (step S22, No).

Then, a job schedule management tool 1
1 indicates that the operator has entered all the items specified in step S21 (step S22, Ye
s), as a step S23, the job schedule server searches the database of the job schedule server to determine whether or not the job execution expected data for the target job is held, and as a result, if not (step S23). S23, No), Step S2
At step 4, the operator manually inputs the job execution end time. Then, the job execution start time input in step S21 and the job execution end time input in step S24 are registered in the schedule file 12 as the job execution start time and execution end time, and the schedule is registered, and the process ends. .

In step S23, if the result of the past execution or the result of the test is recorded for the target job, and the job execution prediction data is stored in the job schedule server (step S23, Yes
s), the size of the processing target data 17 and the execution prediction data of the job are acquired as step S26, and the processing target data 1 obtained in step S27 is obtained as step S28.
7 and the JOB execution prediction data
The predicted time of the OB process is calculated. And step S27
The predicted time is added to the job execution start time input in step S24 to obtain a job execution end time, and using this, the schedule of this job is registered in the schedule file 12 (step S28), and the process ends. .

In the third embodiment, when scheduling a job, if the operator inputs the execution start time of the job, the optimum end time is automatically registered. Next, a fourth embodiment will be described. In the fourth embodiment, when there is a difference between the processing capabilities of the clients constituting the system, the execution time of the job is predicted using the system information 15 of each client registered in advance in the job schedule server. To do.

The system information 15 is data indicating the processing capability of each client. As the system information 15, various values such as a value indicating the capability of the CPU, an access speed of the hard disk drive, and an internal memory capacity can be considered. In the present embodiment, the operating frequency of the CPU of each client is used as the system information 15. An example is shown below. In the following description, for simplicity of explanation, CP
It is assumed that the magnitude of the operating frequency of U simply indicates the processing capability of the client.

For example, it is assumed that JOB-A's JOB execution prediction data exists for the client A, and that the client B's JOB execution prediction data is not recorded. At this time, the job execution expected data of JOB-A at the client A is 30 seconds / Kbyte, and the size of the processing target data is 1M.
In the case of B, the predicted time of the processing of JOB-A by the client A is 30 × 1024 = 30720 seconds.
Then, the CPU of the client B is obtained from the system information 15.
When the data that the operating frequency of the job B is 1.5 times the operating frequency of the CPU of the client A is obtained, the predicted time of the JOB processing by the client B becomes 30720 / 1.5
= 2480 seconds = 341 minutes 20 seconds = 5 hours 41 minutes 20 seconds, and this value is used as the JOB execution prediction data on the display device 3.
Displayed above.

FIG. 9 shows a job schedule management tool 11 of the job schedule server according to the fourth embodiment.
Is a flowchart showing the process according to. This flowchart is for describing a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting the server according to the fourth embodiment of this invention. This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. Part or all of it is W
It may be realized by receiving from the transmission control unit 7 via a network such as AN or LAN or the Internet.

In the figure, the job schedule management tool 11 includes a target job name, processing target data 17,
B. In addition to the full path of the file storing the data generated as the processing result, input of the target client is prompted (step S31), and the job schedule server enters an input waiting state (step S32, No).

Then, a job schedule management tool 1
1 indicates that the operator has entered all the items specified in step S31 (step S32, Ye
s), as a step S33, search the database to determine whether the JOB schedule server holds JOB execution prediction data for the target JOB, and as a result, if not (step S33,
No), the process ends.

In step S33, if the result of the past execution or the result of the test is recorded for the target job, and the job execution prediction data is stored in the job schedule server (step S33, Yes).
s) Then, it is determined whether or not the job schedule server holds the system information 15 for the target client. If the system information 15 is not held as a result (No at step S34), the process ends.

If the system information 15 is held in step S34 (step S34, Yes), the job schedule management tool 11 proceeds to step S35 where the system information, the size of the processing target data 17 and its JO
The execution prediction data of B is acquired, and as step S36, the prediction time of the JOB processing is obtained from the size of the processing target data and the JOB execution prediction data.

Next, in step S37, the job schedule management tool 11 converts the estimated time of the job processing obtained in step S36 into the estimated time of the target client using the system information, and displays this on the display device 3. After that, the process ends. For example, if the operating frequency of the CPU of the client that is the source of the JOB execution prediction data used in step S36 is 1 GHz and the operating frequency of the CPU of the target client is 500 MHz, the predicted time calculated in step S36 Is doubled.

The operator performs scheduling based on the job execution prediction data displayed on the display device 3 and registers it in the schedule file 12. In the fourth embodiment, even when the processing capacities of the clients constituting the system are different from each other, the system information 15 is used to obtain a highly accurate execution prediction time from JOB execution prediction data for other clients. Thus, at the time of scheduling, it is possible to allocate a JOB that can utilize resources more effectively.

Next, a fifth embodiment will be described. In the fifth embodiment, when allocating a job, the schedule file 12 is searched and allocated to a client most suitable for executing the job. In the fifth embodiment, for example, when it is desired to execute a JOB with a predicted time of a certain JOB process of 3 hours as soon as possible, the JOB schedule management tool 11 refers to the schedule file 12 and executes a client that has been available for the last 3 hours. And assign the JOB to the client. At this time, when there are a plurality of vacant clients at the same time, the client is assigned to the client having the highest processing capability.

FIG. 10 shows a job in the fifth embodiment.
JOB schedule management tool for schedule server 1
3 is a flowchart showing a process by No. 1. This flowchart is for explaining a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting the server according to the fifth embodiment of the present invention.
This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In the figure, the job schedule management tool 11 includes a target job name, processing target data 17,
B. In addition to the full path of the file storing the data generated as the processing result, input of the target client is prompted (step S41), and the job schedule server waits for input (step S42, No).

Then, the job schedule management tool 1
1 indicates that all the items specified in step S41 are input by the operator (step S42, Ye
s) In step S43, a test is executed to obtain JOB execution predicted data from the result, or a predicted time of the JOB processing is calculated from the JOB execution predicted data stored in a database and displayed on a display device. 3 is displayed above.

Next, the job schedule management tool 11
Searches the schedule file 12 in step S44 to acquire schedule information of each client. FIG. 11 is a diagram illustrating a configuration example of the schedule file 12. In the example of FIG. 3, the scheduling file 12 records, as schedule information, the execution start time, end time, assigned server name, JOB name, and file name of the processing target data for each scheduled JOB. ing.

The job schedule management tool 11 executes this scheduling file 1 as step S45.
The scheduling status of each client is grasped from the schedule information in 2, the target JOB is scheduled at the optimal time of the optimal client, set in the schedule file 12, and the processing is terminated.

In the fifth embodiment, when allocating a job, the job schedule management tool 11 automatically performs scheduling from the scheduling file 12 to a client most suitable for executing the job. Processing efficiency can be improved.

In the above example, in the fifth embodiment, the JOB schedule server confirms the scheduling status of each client at the time of scheduling, and allocates JOBs in preference to the clients whose schedules are vacant. Instead of the scheduling status at the time when the processing is being performed, an arbitrary time may be set as the start time or the end time, the schedule status at that time may be checked, and the job may be allocated to the client.

Next, a sixth embodiment will be described. In the sixth embodiment, when scheduling a job,
It is checked whether the same JOB has been executed in the past, and if it has been executed, execution prediction is performed from the time required for processing at that time.

For example, when processing the target data 17 having a size of 2 MB in JOB-A, it is checked whether or not JOB-A has been executed in the past by referring to the past log data 13.
As a result, when it is found from the past log data 13 that the result processing executed on the processing target data 17 having a size of 1 MB in the past took one hour, the scheduling is performed with the estimated time of the current JOB-A processing set to two hours. I do.

FIG. 12 shows a job in the sixth embodiment.
JOB schedule management tool for schedule server 1
3 is a flowchart showing a process by No. 1. This flowchart is for explaining a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting a server according to the sixth embodiment of the present invention.
This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In the figure, the job schedule management tool 11 prompts the user to input a target job name, processing target data 17, and a full path of a file for storing data generated as a processing result (step S51). It enters a waiting state (step S52, N
o).

Then, the job schedule management tool 1
1 indicates that all items specified in step S51 are input by the operator (step S52, Ye
s), referring to the past log data 13 as step S53, it is determined whether the same JOB as the target JOB has been executed in the past.

FIG. 13 is a diagram showing a configuration example of the past log data 13. In the example shown in the figure, the past log data 13 records the start and end of the execution of a JOB executed in the past as an event.
The target job name, the file name of the processing target data, the size of the processing target data (K bytes), and whether this event is for the start (START) or end (END) of the job are recorded. In the example of FIG.
JOB-A for 5678 Kbytes of data to be processed
Is executed, the execution time is 02: 34: 56-0:
00: 00 = 2 hours, 34 minutes, 56 seconds.

The job schedule management tool 11 searches the past log data 13 by the job name, and
It is checked whether the same job as the target job has been executed in the past. As a result, if the same job has not been executed in the past (step S53, No), a test similar to that of the first embodiment is performed on the job as step S54, and the estimated time of the job processing obtained from the result is displayed on the display device. 3 and the process ends.

In step S53, if the same job as the target job has been executed in the past (step S5
3, Yes), the job schedule management tool 11
In step S55, the size and the execution time (end time−
(Start time), and the process proceeds to step S56.
The estimated time of the current JOB processing is calculated from the size and the execution time of the processing target data 17 obtained at 5t, and this is displayed on the display device 3 (step S56), and the processing ends.

As a result, the operator proceeds to step S5
4 or the target JOB based on the predicted time displayed in S56
Of the schedule file 12
Set to. In the sixth embodiment, if there is a past execution record for a job to be scheduled, the estimated time of the current job processing is calculated from the execution time obtained from the result at this time. Is performed, the execution prediction can be performed with high accuracy, the number of times of the test can be reduced, and the processing overhead can be reduced.

Next, a seventh embodiment will be described. In the seventh embodiment, a schedule set and input by an operator is checked. For example, if the operator is 2
JOB-A for processing target data 17 of MB size
When the scheduling time is set to 2 hours and registered in the schedule file 12, the job schedule management tool 11 determines the processing time based on the JOB-A processing and test results obtained in the past.
If the estimated time of the process of -A is 3 hours, a warning is displayed to urge the operator to reset the schedule.

FIG. 14 shows a job in the seventh embodiment.
JOB schedule management tool for schedule server 1
3 is a flowchart showing a process by No. 1. This flowchart is for describing a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting a server according to the seventh embodiment of the present invention.
This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In the figure, a job schedule management tool 11 performs job scheduling for an operator.
B, JOB name, file name of processing target data 17, JO
B. Enter the full path of the file storing the data generated as the processing result, the execution start time and the execution end time of the job, and register the schedule (step S6).
1).

Next, the job schedule management tool 11
Is the job execution prediction data of the job registered by the operator in step S62.
The database is searched to determine whether it is held in the schedule server, and as a result, if it is not held (step S62, No), the process ends.

In step S62, if the result of the past execution or the result of the test is recorded for the same job as the job registered in the schedule, and the job execution prediction data is stored in the job schedule server. (Step S62, Yes), as step S63, the job schedule management tool 11 stores the job execution prediction data of the held job and the processing target data 1
The estimated time of the job processing is obtained from the size of “7”.

Then, the job schedule management tool 11
Is the J obtained in step S63 as step S64.
The predicted time of the OB processing is compared with the processing time of the job obtained from the execution start time and the execution end time of the job set by the operator in step S61. as a result,
If the processing time of the job registered and set in step S61 is shorter (step S64, Yes), step S64 is executed.
At step 65, a warning is displayed on the display device 3, and the process returns to step S61 to cause the operator to set the schedule again for the job. Step S
If the processing time of the job registered in the schedule at 61 is longer or the same (step S64, N
o), since the schedule registered and set in step S61 has no problem, the process ends.

In the seventh embodiment, the operator
When the OB schedule is registered, the schedule contents are checked, and if there is a problem, a warning is issued and the operator is re-registered. Therefore, the set processing time is short, and the processing ends before the job is completed. And other troubles can be prevented.

Next, an eighth embodiment will be described. In the eighth embodiment, an average value of past results is obtained for each job, and this is used as job execution prediction data. For example, a certain J
It is assumed that JOB execution prediction data obtained from the result obtained by executing OB is 30 seconds / Kbyte. Also this J
Assuming that the OB has been tested and JOB executed three times in the past, and the JOB execution expected data obtained from the average value at that time is 26 seconds / K byte, the eighth embodiment will be described as follows.
(26 × 3 + 30) / 4 = 27 seconds / K byte is new J
The OB execution prediction data is used in the next scheduling.

FIG. 15 shows a job in the eighth embodiment.
JOB schedule management tool for schedule server 1
3 is a flowchart showing a process by No. 1. This flowchart is for explaining a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting the server according to the eighth embodiment of the present invention.
This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In FIG. 15, the job schedule management tool 11 performs the processing according to each of the above-described embodiments as step S71, and executes a test or a job.
Next, the job schedule management tool 11 executes step S
As 72, the JOB execution prediction data for the JOB that has been tested or
The database is searched to determine whether or not it is held by the B schedule server. If the corresponding job execution prediction data is stored in the job schedule server (step S72, Yes), the test of step S71 is performed as step S73. The average value of the result obtained by the / JOB execution and the past execution result recorded in the database is obtained, and this is recorded and set in the database as JOB execution prediction data to be used in the next scheduling, and the process is terminated.

In step S72, if the corresponding job execution prediction data is not stored in the job schedule server (step S72, No), step S74.
Then, the JOB execution prediction data obtained from the result obtained by the test / JOB execution in step S71 is recorded and set in the database as JOB execution prediction data to be used in the next scheduling, and the process is terminated.

In the eighth embodiment, the JOB execution prediction data used for prediction is obtained in consideration of a plurality of past results, so that the accuracy of the prediction result can be improved.
Next, a ninth embodiment will be described. In the ninth embodiment, the processing target data 17 to be processed by JOB is large.
If it is predicted that it will take a long time to execute the OB, the processing target data 17 is divided into a plurality of pieces and processed by a plurality of clients. Note that the JOB handled in the ninth embodiment performs processing that does not affect the processing result even if the processing target data 17 is divided, such as extracting specific data from the processing target data 17 or performing processing. It is assumed that

For example, the processing time when the client 1 executes JOB-A on the processing target data 17 by the method described above, such as executing a test or using information recorded in the JOB schedule server. If it is predicted that it will take 12 hours, in the ninth embodiment, the job schedule server uses
7 is divided into a plurality of, for example, three, and three clients 3 and
JOB-A uses the divided data as the processing target data 17.
Is executed. Thus, each client can complete JOB-A in four hours.

After each client finishes JOB-A, the processing results of JOB-A from each client are collected, and the target processing result can be obtained by combining with the JOB schedule server. FIG. 16 is a flowchart showing processing by the job schedule management tool 11 of the job schedule server according to the ninth embodiment.

This flowchart is for explaining a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management apparatus on a computer constituting a server according to the ninth embodiment of the present invention. It is. This program runs on the CPU
1 is stored in the recording medium 8 in the form of a readable program code. However, it is not necessary to store all functions in the recording medium 8.
A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In the ninth embodiment, when a job scheduling request is issued, the job schedule management tool 11 determines in step S81 the execution job name and the file name of the processing file data 17 in accordance with the method described above. , A predicted time of the job processing is obtained.

As a result, if the estimated time is longer than the allowable time for the job processing, the job schedule management tool 11 divides the storage file of the processing target data 17 into a plurality of files based on the estimated time of the job as step S82. The division position is, for example, when the file is C
If the file is an SV format data file, the file is divided at a location where the processing does not affect the processing, such as division at a line feed position. Then, the divided file is used as the processing target data 17 by a plurality of clients with the JO.
Distribute B.

When the job ends at each client and the processing result is sent, the job schedule management tool 11
Collects these processing results in step S83 and combines them into one file in consideration of the position of the divided file in the file before division.

As a result, the designated processing target data 1
7 is obtained. In the ninth embodiment, when a certain JOB is processed within a specific time, even if the processing target data 17 is large and cannot be processed within the time when processed by a single client, The data 17 can be divided and processed by a plurality of clients, and the job can be completed in time.

Next, a tenth embodiment will be described.
In the tenth embodiment, in the ninth embodiment, the number of divisions of the processing target data 17, that is, the number of clients to which JOBs are distributed is determined from the estimated time and the registration time.

For example, when a schedule for executing a job in two hours is set, the job of the job is
If the predicted time of the B process is 8 hours and 30 minutes, 8.5
Since /2=4.25 (rounded up to 5 after the decimal point), the file of the processing target data 17 is divided into five and the job is distributed to five clients.

FIG. 17 shows the JO in the tenth embodiment.
13 is a flowchart illustrating a process performed by a job schedule management tool 11 of the B schedule server. This flowchart is for explaining a program (JOB schedule management tool 11) for realizing each function as a JOB schedule management device on a computer constituting a server according to the tenth embodiment of the present invention. This program will be described using an example in which the program is stored in the recording medium 8 in the form of a program code that can be read by the CPU 1. However, it is not necessary to store all functions in the recording medium 8. A part or all of them may be realized by receiving them from the transmission control unit 7 via a network such as a WAN or LAN or the Internet.

In the figure, the job schedule management tool 11 firstly gives the operator the name of the execution job, the file name of the processing file data 17,
Enter the start / end time of the JOB processing
B is registered in the schedule. Next, as step S92, the job schedule management tool 11 obtains the estimated time of the job processing by the method described above. As a result, if the predicted time is longer than the processing time for the JOB processing obtained from the execution start / end time, J
The OB schedule management tool 11 performs, at step S93, the estimated time of the JOB processing obtained at step S92 /
The processing target data 17 is divided into the number obtained from the registered processing time. For example, when the time allowed for the processing of the job is 4 hours and the predicted time is 12 hours, the processing target data 17
Is divided into three, and the job is distributed to each client.

When each client completes the execution of the job, the job schedule management tool 11 proceeds to step S
Processing results are collected as 94, and these are combined to obtain a processing result for the requested processing target data 17. In the tenth embodiment, when a job is to be executed within a predetermined time, the processing is distributed to a plurality of clients in consideration of the predetermined time, so that the processing can be surely completed within the time. Can be.

Next, an eleventh embodiment will be described.
In the eleventh embodiment, the size of the processing target data 17 and J
A plurality of relations with the OB processing time are prepared, and these are properly used depending on the application executed in the job to obtain a predicted value.

In the examples so far, the size of the processing target data 17 is proportional to the processing time.
Has been calculated based on the assumption that the relationship between the size of d and the processing time t is t = αd (α is a proportional constant). However, the processing time is actually determined by the processing executed by JOB. The relationship between the size of the target data 17 and its processing differs. For example, when processing of 1 × 1, 1 × 2,... 9 × 9 is performed using the processing target data 17, 1, 2,.
If the size of the processing target data 17 is d, the relation of the processing time t is t = αd ² .

In the present embodiment, the job schedule management tool 11 stores a plurality of expressions indicating the relationship between the processing target data 17 and the processing time. Then, the formula corresponding to each job is checked, and the formula used at the time of prediction is changed. Further, by making fine adjustments to each coefficient of the expression indicating this relationship based on the result, for example, taking statistics of the result obtained by executing the test / JOB, more accurate prediction can be performed.

In the client / server of this embodiment,
In some cases, application software is exchanged using a storage medium such as a floppy (registered trademark) disk.
Therefore, the present invention provides a job schedule management device and a job schedule management device.
The present invention is not limited to the B schedule management method, and may be configured as a computer-readable storage medium or a program for causing a computer to perform the functions of the above-described embodiments of the present invention when used by the computer.

In this case, the “storage medium” includes, for example, FIG.
As shown in FIG. 8, a removable storage medium 36 such as a CD-ROM, a floppy disk (or an MO, a DVD, or a removable hard disk) detachable from a medium drive device 37 or a network line 33 A storage means (database or the like) 32 in an external device (a server or the like) to be transmitted, or a main body 34 of the information processing device 31
(RAM, hard disk, etc.) 35 and the like. The program stored in the portable storage medium 36 or the storage means (database or the like) 32 is loaded into a memory (RAM or hard disk or the like) 35 in the main body 34 and executed.

[0103]

As described above in detail, according to the present invention, a job is executed using test data generated from data to be processed, and the job is executed based on the time required for the processing. Since the processing prediction time is calculated, highly accurate prediction can be performed. Therefore, the processing efficiency of the entire system can be improved.

Further, since the test process is not performed more than necessary, the overhead of the process required for the prediction can be reduced.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a hardware configuration of a client and a job schedule server according to an embodiment.

FIG. 2 is a functional configuration block diagram of a job schedule server according to the embodiment.

FIG. 3 is a diagram illustrating an example of a JOB script.

FIG. 4 is a configuration block diagram of a client according to the embodiment.

FIG. 5 is a flowchart illustrating a process performed by a job schedule management tool on a job schedule server according to the first embodiment.

FIG. 6 is a diagram illustrating generation of test data.

FIG. 7 is a flowchart illustrating a process performed by a job schedule management tool on a job schedule server according to the second embodiment.

FIG. 8 is a flowchart illustrating a process performed by a job schedule management tool on a job schedule server according to the third embodiment.

FIG. 9 is a flowchart illustrating a process performed by a job schedule management tool on a job schedule server according to a fourth embodiment.

FIG. 10 is a diagram illustrating a flowchart illustrating processing performed by a job schedule management tool on a job schedule server according to the fifth embodiment.

FIG. 11 is a diagram illustrating a configuration example of a schedule file.

FIG. 12 is a diagram illustrating a flowchart showing processing by a job schedule management tool of a job schedule server according to a sixth embodiment.

FIG. 13 is a diagram illustrating a configuration example of past log data.

FIG. 14 is a diagram illustrating a flowchart illustrating processing by a job schedule management tool of a job schedule server according to a seventh embodiment.

FIG. 15 is a diagram illustrating a flowchart showing processing by a job schedule management tool of a job schedule server according to the eighth embodiment.

FIG. 16 is a diagram illustrating a flowchart showing processing by a job schedule management tool of a job schedule server according to the ninth embodiment.

FIG. 17 is a diagram illustrating a flowchart showing processing by a job schedule management tool of a job schedule server according to the tenth embodiment.

FIG. 18 is a diagram illustrating an example of a medium.

FIG. 19 is a diagram showing parallel processing in a general client / server system.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 CPU 2 Input device 3 Display device 4 Printing device 5 Storage device 6 RAM 7 Transmission control unit 11 JOB schedule management tool 12 Schedule file 13 Past line log data 14 Test data at each client 15 System information 16 JOB script 21 JOB management tool 22 Job script 23 Application 24 Input data 25 Output data 31 Information processing device 32 Storage means 33 Network line 34 Information processing device main body (computer) 35 Memory 36 Portable recording medium

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5B042 HH20 MA00 MA05 5B098 AA10 GA03 GA08

Claims (18)

[Claims] 1. A job in a client / server system, comprising a plurality of clients and a server, wherein the server allocates a job for which an execution instruction is issued to the client.
A schedule management device, comprising: a test data generation unit configured to generate test data from processing target data processed by a job instructed to execute; and a test process for causing a client to process the test data by the job. , The result of the test process,
JOB schedule management apparatus, comprising: a JOB prediction time calculation unit for calculating a prediction time when the processing target data is processed by the JOB from the size of the processing target data and the size of the test data. . 2. The apparatus according to claim 1, further comprising a storage unit configured to store information used for obtaining the estimated time, which is obtained from a result of the test processing, wherein the information for the job for which the execution instruction is given is stored in the storage unit. 2. The job schedule management device according to claim 1, wherein the job prediction time calculation means obtains the prediction time using the information. 3. The J according to claim 1, wherein the test data generation unit extracts a part of the processing target data to generate the test data.
OB schedule management device. 4. A system information storage unit for storing system information indicating a processing capability of each of the plurality of clients, wherein the JOB predicted time calculation unit calculates the predicted time using the system information. J according to any one of claims 1 to 3, characterized in that:
OB schedule management device. 5. The system according to claim 4, wherein the system information is an operating frequency of a CPU of the client.
2. The job schedule management device according to 1. 6. A schedule storage means for recording a job assignment state of the plurality of clients, and a job execution instruction based on the job assignment state and the predicted time stored in the schedule storage means in response to a job execution instruction.
The job schedule management device according to any one of claims 1 to 5, further comprising a job assignment unit that assigns B. 7. A past log data storage unit for holding a record of the execution of a job executed by a client, wherein the job estimated time calculation unit responds to the instruction to execute the job in the execution record of the job. The job schedule management device according to any one of claims 1 to 6, wherein if there is an execution record for the set job, the predicted time is obtained using the execution record. 8. A scheduling means for allocating a job to a client based on an input content inputted to the job by an operator, and a warning means for checking the input content and warning the operator if there is a problem. The job schedule management device according to any one of claims 1 to 7, comprising: 9. The input content includes time information about the start and end of execution of a job, and the warning means issues a warning when the execution time of the job obtained from the time information is shorter than the predicted time. The job schedule management device according to claim 8, wherein: 10. When the execution of the job and the test processing have been performed a plurality of times in the past for the job for which the execution instruction has been issued, or when both the execution of the job and the test processing have been performed. 10. The method according to claim 2, wherein the JOB prediction time calculation means calculates an average value from information used for calculating the prediction time and the execution record, and calculates the JOB prediction time from the average value. Any one of them
The job schedule management device according to the paragraph. 11. A processing target data dividing means for dividing the processing target data into a plurality of pieces, and a JOB division allocating means for allocating to the plurality of clients a JOB instructed to be executed with the divided processing target data as a processing target. The job schedule management device according to any one of claims 1 to 10, wherein: 12. The processing target data dividing means, wherein the predicted time and the JO
12. The number of divisions to be divided is determined based on a processing time registered for B.
The job schedule management device according to any one of the above. 13. The job estimated time calculation means has a plurality of relational expressions between the size of the data to be processed and the time required to process the data to be processed, and executes the job by the application used by the job that has been instructed to execute. 13. The job schedule management device according to claim 1, wherein the predicted time is obtained by switching a relational expression. 14. The JOB prediction time calculation means, wherein:
14. The job schedule management device according to claim 13, wherein the corresponding relational expression is changed based on a result of executing B. 15. The job schedule management device according to claim 1, further comprising monitor means for displaying the predicted time obtained by the job predicted time calculation means. 16. A job schedule management method performed by a server in a client / server system in which a server allocates a job for which execution is instructed to a client, comprising a plurality of clients and a server. Generating test data from the data to be processed, instructing the client to execute a job for which the execution instruction has been issued for the test data, the time required for the client to process the test data, the processing target A predicted time when the processing target data is processed by the JOB is obtained from a size of the data and a size of the test data, and based on the predicted time, a JOB that has received the execution instruction for the processing target data is obtained. Of execution to one of the plurality of clients JOB schedule management method, characterized in that. 17. When processing of test data has been performed in the past by a JOB for which execution has been instructed, the test data is not generated, and the time required for processing of the past test data is used. 18. The JO according to claim 16, wherein an estimated time when the processing target data is processed by the JOB is obtained.
B schedule management method. 18. A job in which an execution instruction is issued when used by an information processing apparatus constituting a server in a client / server system in which the server allocates a job to which execution is instructed to a client, comprising a plurality of clients and a server. Reading out the processing target data processed from the memory from the memory, generating test data from the read processing target data, instructing the client to execute the JOB instructed to execute the test data, and the client From the time required for processing the test data, the size of the data to be processed, and the size of the data for test, a predicted time when the data to be processed is processed by the JOB is obtained, and based on the predicted time Of the job instructed to execute the data to be processed Program for causing allocation row for one of said plurality of clients, to record in the memory to the information processing apparatus.