JP2004005288A - Batch performance evaluating method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve problems in evaluating the performance of a batch job constructing a data warehouse, that the input of a calculation rule is unnecessary as a performance calculation rule is trivial, it is difficult to input a job processing data and a processing time in advance, and it is impossible to select a hardware from a CPU utilization time of every processing data and processing time zone. <P>SOLUTION: This batch performance evaluating method comprises steps of inputting a data flow, inputting a unit value of machine performance, setting a property, calculating a job single execution time, inputting a job schedule, calculating a job total execution time, inputting a job execution allowed time, and selecting a machine. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a performance estimation method and apparatus for estimating system performance at a design stage, and more particularly, to a batch job performance estimation method and apparatus for constructing a data warehouse system or the like.
[0002]
[Prior art]
As disclosed in Japanese Patent Application Laid-Open No. 2001-125784, (1) a data flow diagram, a record specification definition, and a method included in the data flow are disclosed as a performance estimation method and apparatus for estimating the performance of a system at a design stage. A processing pattern representing an outline of the processing of the process, the number of data items handled by the process, and a calculation rule for estimating the calculation time of the process are input to calculate a predicted execution time (CPU processing time) of each process. The CPU processing time for each processing time zone is calculated from the means for setting the processing time zone and the CPU processing time calculated in (1), and (3) hardware performance information and each calculated in (2). There is known a performance estimation method and apparatus for outputting advice information for hardware selection from CPU processing time for each processing time zone.
[0003]
[Problems to be solved by the invention]
The above-mentioned known example presupposes a method and an apparatus for estimating performance in the design of an online application, and when it is applied to a batch job for constructing a data warehouse, (1) batch processing for constructing a data warehouse requires input / output of data. The performance calculation rule is self-evident because it is tied to the performance, and there is no need to input the calculation rule. (2) The batch processing is executed based on the job execution order defined separately from the data flow. Since the time period to be executed is determined by the execution time of the job executed before the job, it is difficult to input the processing date and the processing time of the job in advance. (3) Hardware The selection is determined by how long the batch processing needs to be completed, so the CPU for each processing date and processing time zone From use time can not be performed selecting the hardware, there is a problem that.
[0004]
A first object of the present invention is to perform batch processing based on hardware performance unit price information and a data flow diagram, the number of data items handled by a job, a job schedule, and a target value of a job execution time in an estimation or design stage of a data warehouse system or the like. It is an object of the present invention to provide a performance estimating method that reduces and eliminates rework in the development process by calculating and presenting hardware that satisfies the processing time and the job execution time, and enables estimation and performance evaluation at the design stage.
[0005]
A second object of the present invention is to perform batch processing based on hardware performance unit price information and a data flow diagram, the number of data items handled by a job, a job schedule, and a target value of a job execution time in the estimation and design stages of a data warehouse system. An object of the present invention is to provide a performance estimating apparatus that calculates and presents hardware that satisfies a processing time and a job execution time, and that enables performance evaluation at a stage of estimation and design, thereby reducing rework in a development process.
[0006]
[Means for Solving the Problems]
A first object of the present invention is to provide a data flow input step for inputting a data flow expressing a flow of business data by arranging a table and a job, which is created at the time of designing system specifications, and a machine for inputting a performance unit price for each machine. A performance unit price input step, a property setting step for setting the number of input data items and output data items of the job included in the data flow, a number of input data items and output data number of the jobs included in the data flow, and a machine performance unit price input step. A job independent execution time calculation step of calculating an estimated execution time of each job from the machine performance unit price input using the job schedule input step of inputting a job schedule, and a job schedule input using the job schedule input step From the estimated execution time of each job Total job execution time calculation step for calculating the overall execution time, job execution allowable time input step for inputting the job execution allowable time, and execution prediction of the entire job for each machine input using the machine performance unit price input step This can be achieved by a performance estimation method having a machine selection step of outputting the recommended job execution machine information by calculating the time and comparing the estimated execution time of each machine with the allowable job execution time.
[0007]
The second object is to provide a data flow input unit for inputting a data flow expressing a flow of business data by arranging a table and a job created at the time of system specification design, and a machine performance for inputting a performance unit price for each machine. Unit price input means, property setting means for setting the number of input data items and the number of output data items of the job included in the data flow, and the number of input data items and the number of output data items of the job included in the data flow and the machine performance unit price input step are used. Job execution time calculating means for calculating an estimated execution time of each job from the machine performance unit price inputted by the user, a job schedule input means for inputting a job schedule, and a job schedule input using the job schedule input means. Calculate the execution time of the entire job from the estimated execution time of the job. Total job execution time calculating means, a job execution allowable time input means for inputting a job execution allowable time, and an estimated execution time of the entire job for each machine input using the machine performance unit price input step. The performance can be achieved by a performance estimating apparatus including a machine selecting unit that outputs recommended job execution machine information by matching the estimated execution time for each job with the allowable job execution time.
[0008]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 shows a functional configuration of the batch performance estimation method of the present invention.
[0009]
In FIG. 1, the batch performance estimation method includes a performance calculation step 1, a data flow description GUI 2, a data flow table 3, an entire job allowable execution time table 4, a per-machine performance unit price table 5, a job independent execution estimated time table 6, an entire job It has an expected execution time table 7 and a recommended job execution machine table 8, and further, the performance calculation step 1 calculates a job alone execution time calculation step 9, a whole job execution time calculation step 10, and calculates the entire job execution time for all machines. The data flow table 3 further includes a data table 31, an arrow table 32, a job table 33, and a job schedule table 34. The data flow input step, the property setting step, the job schedule input step, the machine performance unit price input step, and the entire job allowable execution time input step exist in the data flow description GUI2.
[0010]
FIG. 13 shows an example of the configuration of an apparatus for executing the performance calculation program 1 storing the present invention. The device includes a CPU 1301, a main memory 1302, a keyboard and mouse 1303, a display 1304, a bus 1305, a CD-ROM or FDD 1306, and a hard disk 1307.
[0011]
The performance calculation program 1 storing the present invention is read from the CD-ROM or the FDD 1306 and stored in the hard disk 1307 via the bus 1305. At the time of execution, the performance calculation program 1 is read into the main storage 1302 and executed by the CPU 1301.
[0012]
The data of the data flow table 3, the entire job permissible time table 4, and the per-machine performance unit price table 5 are input by a keyboard and mouse 1303 and stored in the hard disk 1307. Further, these data are temporarily stored in the main memory 1302.
[0013]
Further, the data of the estimated job execution time table 6, the estimated job execution time table 7, and the recommended job execution machine table 8 are stored in the main memory 1302 and then stored in the hard disk 1307 by the CPU 1301. The data of the estimated job execution time table 6, the estimated job execution time table 7, and the recommended job execution machine table 8 are processed by the CPU 1301 and displayed on the display 1304.
[0014]
FIG. 2 shows a configuration example of the data table 31. The data table 31 is a table for storing information of data in the data flow diagram, and includes a name field 311 for storing data names.
[0015]
FIG. 3 shows a configuration example of the arrow table 32. The arrow table 32 is a table that stores information of an arrow in a data flow, and stores a job name field 321 that stores the name of a job that becomes an input or output of an arrow, and stores a name of data that becomes an input or output of an arrow. A data name field 322, a processed record number field 323 storing the number of data input to the job or the number of data output by the job, and an input / output division field 324 indicating whether the data is input data to the job or output data from the job. Be composed.
[0016]
FIG. 4 shows a configuration example of the job table 33. The job table 33 is a table for storing information on jobs in the data flow, and includes a name field 331 for storing job names.
[0017]
FIG. 5 shows a configuration example of the job schedule table 34. The job schedule table 34 is a table that stores the execution order of jobs, and includes a previous job field 341 and a next job field 342.
[0018]
FIG. 6 shows a configuration example of the per-machine performance unit price table 5. The machine-specific performance unit price table 5 includes a machine name field 501, a record IO number per second field 502, and a processed record number per second field 503.
[0019]
FIG. 7 shows a configuration example of the job independent execution estimated time table 6. The estimated job execution time table 6 includes a job name field 601, a machine name field 602, and an estimated execution time field 603.
[0020]
FIG. 8 shows an example of the configuration of the estimated job execution time table 7. The entire job estimated execution time table 7 includes a machine name field 701 and an estimated execution time field 702.
[0021]
FIG. 9 is a flowchart showing the processing executed in the job independent execution time calculation step 9. First, one record is acquired from the per-machine performance unit price table 5 (step 901), and it is determined whether or not the record has been acquired. This step is the step 11 of FIG. If a record has been obtained, one record is obtained from the job table 33 (step 903), and it is determined whether the record has been obtained (step 904). If the record is acquired, the record having the value of the record name field 331 acquired in step 903 in the job name field 321 is acquired from the arrow table 32 (step 905).
[0022]
Thereafter, in step 906, an expected execution time is calculated. The estimated execution time is obtained by dividing the sum of the number of processed records field 323 of the record obtained in step 905 by the value of the number of data IOs per second field 502 of the record obtained in step 901 and the value of the record obtained in step 905. Of these, the sum of the number of processed records field 323 of the record whose input / output section field 324 is “input” is calculated as the sum of the value obtained by dividing the number of processed records per second 503 of the record acquired in step 901. .
[0023]
Then, the predicted execution time obtained in step 906 is stored in the predicted execution time field 603 of the predicted job independent execution time table 6 in FIG. 6, and the value of the name field 331 of the record acquired in step 903 is stored in the predicted job execution time. The value of the machine name field 501 of the record acquired in step 901 is set in the job name field 601 of the table 6 in the machine name field 602 of the estimated job independent execution time table 6, respectively.
[0024]
After that, the entire job execution time calculation step 10 is executed, and the process returns to step 903. If a record cannot be obtained in step 904, the process returns to step 901. If no record can be obtained in step 11, the process ends. In Steps 901 and 903, the reading position of the record is increased every time one record is acquired. When all the records have been read, it is determined in Steps 11 and 904 that there is no record.
[0025]
FIG. 10 is a flowchart showing the processing performed in the overall job execution time calculation step 10. First, a graph representing the job execution order is created with reference to the job schedule table 34 shown in FIG. 5 (step 1001). In the example of FIG. 5, a graph is generated in which the job JOB001 is connected to the job JOB002.
[0026]
Next, all paths from the start node to the end node of the graph created in step 1001 are obtained (step 1002). The start node is JOB001 and the end node is JOB002, and the obtained path is a path from JOB001 to JOB002. Thereafter, 0 is set in an area indicating the calculation result of the execution time (step 1003).
[0027]
Next, it is determined whether or not the path still exists (step 1004). If the path does exist, the machine name of the record obtained in step 901 of FIG. 9 is stored in the machine name field 602, and in step 1002, For each job existing on the path that is currently being processed among the obtained paths, the estimated execution time is obtained from the job independent execution estimated time table 6 (step 1005).
[0028]
Thereafter, the estimated execution time of each job obtained in step 1005 is totaled (step 1006). If the value stored in the area indicating the calculation result of the execution time is smaller than the value obtained in step 1006 (step 1007), the value obtained in step 1006 is stored in the area (step 1008). If it is determined in step 1007 that the value stored in the area indicating the calculation result of the execution time is equal to or larger than the value obtained in step 1006, the process returns to step 1004. If it is determined in step 1004 that the execution time has been calculated for all the paths obtained in step 1002 (the determination result is N), the value of the machine name field 602 of the record obtained in step 901 in FIG. In the machine name field 701 of the estimated job execution time table 7 in FIG. 8, the value stored in the area indicating the calculation result of the execution time at the time of executing the step 1009 is set as the estimated execution time value of the estimated job execution time table 7. Each is stored in a field 702.
[0029]
FIG. 11 is an example of a process performed in the machine selection step 12 of FIG. In the machine selection step 12, a record that does not exceed the value of the entire job allowable execution time 4 and the value of the estimated execution time of the expected entire job execution time table 7 in FIG. The value of the machine name field 701 is obtained (step 1101).
[0030]
FIG. 12 is a diagram showing an example of the data flow description GUI 2 of FIG. The data flow description GUI 2 includes a data flow input unit 1201 for inputting a data flow, a property setting unit 1202 for setting properties of a table, a job, and an arrow in the data flow, and a job allowable execution time input for inputting a job allowable execution time. Means 1203, machine performance unit price input means 1204, job schedule input means 1209 for inputting a job schedule, job independent execution expected time result display means 1205 for displaying data of the job individual execution expected time table 6 in FIG. 7, and job in FIG. The system includes an expected job execution time result display unit 1207 for displaying data of the expected execution time table 7 and a recommended machine display unit 1208 for displaying data of the recommended job execution machine 8 in FIG.
[0031]
The data input by the data flow input unit 1201 and the property setting unit 1202 is a record 310 of the data table 31 when the input target is the table 1210, and a record 320 of the arrow table 33 when the input target is the arrow 1212. In the case of the job 1211, it is stored as the record 330 of the job table 33.
[0032]
Taking the data flow shown in FIG. 12 as an example, in this data flow, table 1 (1210a), table 2 (1210b), table 3 (1210c), table 4 (1210d), job JOB001 (1211a), and job JOB002 (1211b), an arrow 1212a indicating that data flows from the table 1210a to the job 1211a, an arrow 1212b indicating that data flows from the job 1211a to the table 1210b, and an arrow 1212c indicating that data flows from the table 1210b to the job 1211b. , An arrow 1212d indicating that data flows from the job 1211b to the table 1210c, and an arrow 1212e indicating that data flows from the job 1211b to the table 1210d. The data of 210a is stored in the record 31a of the data table 31, the data of the table 1210b is stored in the record 31b of the data table 31, the data of the table 1210c is stored in the record 31c of the data table 31, and the data of the table 1210d is stored in the data table 31. In the record 31d, the data of the job 1211a is stored in the record 33a of the job table 33, the data of the job 1211b is stored in the record 33b of the job table 33, and the data of the arrow 1212a is stored in the record 32a of the arrow table 32. The data is stored in the record 32b of the arrow table 32, the data of the arrow 1212c is stored in the record 32c of the arrow table 32, and the data of the arrow 1212d is stored in the record 32d of the arrow table 32. Data 212e is the record 32e of Arrow table 32, are respectively stored by the data flow input step.
[0033]
The data input to the machine performance unit price input means 1204 is stored as a record 51 of a machine-specific performance unit price table in a machine performance unit price input step.
[0034]
In the example of FIG. 6, there are two machines SV001 and SV002, and the number of records IO per second and the number of records processed per second are 10, 100,000, 100, and 100,000, respectively.
[0035]
The data input to the job allowable execution time input unit 1203 is stored in the job entire execution allowable time table 4 of FIG. 1 by the job allowable execution time input step. The data input to the job schedule input unit 1209 is input as records 341 and 342 of the job schedule table 34 in FIG. In the example of FIG. 5, the record 34a indicates that the job JOB002 is executed after the job JOB001.
[0036]
Hereinafter, an example will be described in which the single execution expected time of each job is calculated from the data flow shown in FIG. 12, the job schedule shown in FIG. 5, and the performance unit price for each machine shown in FIG.
[0037]
First, in step 901 of the job independent execution expected time calculation step 9 shown in FIG. 9, the record 51a is acquired from the per-machine performance unit price table 5 of FIG. Since the record has been obtained, the result of step 11 is Y. Next, in step 903, the record 33a is obtained from the job table 33 of FIG. Since a record has been obtained, step 904 becomes Y.
[0038]
Next, step 905 is executed. Since the value of the name field 331 of the record 33a acquired in step 903 is “JOB001”, a record having “JOB001” in the job name field 321 is stored in the arrow table 32 of FIG. To get from. In this case, the record 32a and the record 32b are obtained.
[0039]
Next, in step 906, the estimated job execution time is calculated. First, the sum of the processed record number fields 323 of the records 32a and 32b is calculated. In this case, it is 11000.
[0040]
Next, a value is obtained by dividing this by the value 10 of the number of record IOs per second field of the record 51a acquired in step 901. In this case, 11000 ÷ 10, and the result is 1100.
[0041]
Next, of the records 32a and 32b, the record whose value of the input / output section 324 is "input" is selected. In this case, the record 32a corresponds. A value is obtained by dividing the value of the processed record number field 323 of the record 32a by the number of processed records per second of the record 51a. In this case, 10,000 ÷ 100,000, and the result is 0.1. Next, the previously calculated 1100 and 0.1 are added to obtain the estimated job execution time. In this example, the decimal portion is rounded down, and 1100 is set as the value of the estimated job execution time.
[0042]
Then, in step 907, this value, the job name, and the machine name are written as the record 61a of the estimated job execution time table 6 in FIG. After that, the entire job execution time calculation step 10 is executed, and the process returns to step 903. An execution example of the entire job execution time calculation step 10 will be described later.
[0043]
Thereafter, the same processing is performed on the record 33b of the job table 33 of FIG. 4, and the record 61b of the estimated job independent execution time table 6 of FIG. 7 is added.
[0044]
Thereafter, the same processing is performed on the record 51b of the per-machine performance unit price table 5 of FIG. 6 to generate the records 61c and 61d of FIG. These records 61a to 61d are displayed on the job independent execution expected time result display means 1205 in FIG.
[0045]
Next, an example of obtaining the estimated execution time of the entire job will be described. First, in step 1001 of FIG. 10, a graph indicating the job execution order is generated from the job schedule table of FIG. In this case, as shown in FIG. 14, a graph including a node 1401 indicating the job JOB001, a node 1402 indicating the job JOB002, and an edge 1403 connecting the nodes in the direction from the node 1401 to the node 1402 is obtained.
[0046]
Next, in step 1002, a path from the start node to the end node of the graph is obtained. In this case, the number of paths is one and is as shown in FIG.
[0047]
Next, the execution time is set to 0 (step 1003), one path obtained in step 1002 is taken out (step 1004), and in step 1005, the single execution time of the job on the path is obtained. In FIG. 14, the independent execution time is obtained for each of the jobs of the job nodes 1401 and 1402. In the case of the job node 1401, the value “SV001” of the machine name field 501 of the record 51a (FIG. 6) obtained in step 901 is included in the machine name field 602 of FIG. 7, and the name of the job node 1401 is “JOB001”. Is obtained from the job-only expected execution time table 6 in FIG. 7. In this case, the record 61 a corresponds to the record, and then the value 1100 of the expected execution time field 603 of the record 61 a is obtained. Similarly, the estimated execution time is acquired for the job node 1402, and in this case, it is 560.
[0048]
Next, these values are summed to obtain 1660 as the execution time of the entire job (step 1006). Next, step 1007 is executed. At this point, since the execution time is 0, the determination result is Y. In step 1008, 1660 is set as the execution time. If there are two or more paths in step 1002, these processes are further repeated (the result of step 1004 is N) to obtain the time with the largest value. In this example, the execution time is 1660.
[0049]
Thereafter, in step 1009, the execution time 1660 is set in the field 702 of the record 71a of the estimated job execution time table in FIG. 7, and the machine name SV001 is set in the machine name field 701.
[0050]
Thereafter, the entire job execution time calculation step 10 is called again for the machine SV002 from the flow of FIG. 9, and the record 71b of FIG. 8 is generated.
[0051]
Finally, an example of calculating a job execution recommended machine in the machine selection step 12 will be described.
[0052]
First, when 1800 is input from the job allowable execution time input unit 1203 in FIG. 12 as the entire job allowable execution time, this value is stored in the entire job allowable execution time table 4. In the machine selection step 12, step 1101 is executed. From the record 71 in FIG. 8, the record in which the value of the predicted execution time field 702 does not exceed 1800 and has the largest value is selected. In this case, the record 71a corresponds, and the job execution recommended machine is the value “SV001” of the machine name field 701 of the record 71a.
[0053]
Next, when 200 is input as the permissible execution time of the entire job, the value of the predicted execution time field 702 does not exceed 200, and 71b is selected as the record having the largest value, and the job execution recommended machine is SV002. It becomes. If 20 is entered as the overall job execution allowable time, there is no record in which the value of the execution time predicted value field 702 does not exceed 20, and there is no corresponding machine. In the present embodiment, in the estimation and design stages of the data warehouse system, the processing time of the batch processing and the target value of the data flow diagram, the number of data handled by the job, the job schedule, and the job execution time, Hardware that satisfies the job execution time can be calculated and presented. As a result, it is possible to reduce rework in the development process by enabling estimation and performance evaluation at the design stage.
[0054]
【The invention's effect】
As described above, according to the present invention, at the estimation or design stage of a data warehouse system, based on hardware performance unit price information and a data flow diagram, the number of data items handled by a job, a job schedule, and a target value of a job execution time, It has been shown that hardware that satisfies the processing time of batch processing and the job execution time can be calculated and presented. As a result, it is possible to reduce the rework in the development process by enabling estimation and performance evaluation at the design stage.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration example of a performance estimation method and apparatus for implementing the present invention.
FIG. 2 is a diagram for explaining a configuration example of a data table 31;
FIG. 3 is a diagram for explaining a configuration example of an arrow table 32;
FIG. 4 is a diagram for explaining a configuration example of a job table 33.
FIG. 5 is a diagram for describing a configuration example of a job schedule table 34.
FIG. 6 is a diagram for describing a configuration example of a performance unit price table by machine 5;
FIG. 7 is a diagram for explaining a configuration example of a job independent execution estimated time table 6.
FIG. 8 is a diagram for describing a configuration example of a job entire execution estimated time table 7.
FIG. 9 is a flowchart illustrating a process performed in a job independent execution time calculation step 9;
FIG. 10 is a flowchart illustrating a process performed in an entire job execution time calculation step 10;
FIG. 11 is a flowchart showing a process performed in a machine selection step 12;
FIG. 12 is a diagram for describing an example of a data flow description GUI2.
FIG. 13 is a diagram illustrating a configuration example of a system in which the present invention is implemented.
FIG. 14 is a diagram illustrating an example of a network indicating the execution order of jobs.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Performance calculation step, 2 ... Data flow description GUI, 3 ... Data flow table, 31 ... Data table, 32 ... Arrow table, 33 ... Job table, 34 ... Job schedule table, 4 ... Permissible job execution time table, 5 ... Machine performance unit price table, 6 ... Job independent execution time table, 7 ... Job total execution time table, 8 ... Job execution recommended machine table, 9 ... Job individual execution time calculation step, 10 ... Job total execution time calculation step , 11: Step of determining whether or not the entire job execution time has been calculated for all machines, 12: Machine selection step.

Claims (6)

A method of estimating system performance when designing a system,
A data flow input step of inputting a data flow expressing a flow of business data by arranging a table and a job created at the time of system specification design;
A machine performance unit input step for inputting a performance unit price for each machine;
A property setting step of setting the number of input data items and the number of output data items of a job included in the data flow;
A job independent execution time calculating step of calculating an expected execution time of each job from the number of input data and the number of output data of the jobs included in the data flow and the machine performance unit price input using the machine performance unit input step;
A batch performance estimation method characterized by having: The batch performance estimation method according to claim 1, further comprising:
A job schedule input step for inputting a job schedule;
An overall job execution time calculating step of calculating the overall job execution time from the job schedule input using the job schedule input step and the expected execution time of each job;
A batch performance estimation method characterized by having: The batch performance estimation method according to claim 2, further comprising:
A job execution allowable time input step for inputting a job execution allowable time;
The estimated execution time of the entire job is calculated for each machine input using the machine performance unit price input step, and job execution recommended machine information is output by matching the estimated execution time of each machine with the permissible job execution time. Machine selection step;
A batch performance estimation method characterized by having: A device for estimating system performance when designing a system,
Data flow input means for inputting a data flow expressing a flow of business data by arranging a table and a job, which are created at the time of system specification design
A machine performance unit input step for inputting a performance unit price for each machine;
Property setting means for setting the number of input data items and the number of output data items of a job included in the data flow;
A job independent execution time calculating unit that calculates an expected execution time of each job from the number of input data and the number of output data of the jobs included in the data flow and the machine performance unit price input using the machine performance unit input unit;
A batch performance estimating apparatus comprising: The batch performance estimating apparatus according to claim 4, further comprising:
A job schedule input unit for inputting a job schedule;
An overall job execution time calculation unit that calculates the overall job execution time from the job schedule input using the job schedule input unit and the estimated execution time of each job;
A batch performance estimating apparatus comprising: The batch performance estimating apparatus according to claim 5, further comprising:
A job execution allowable time input unit for inputting a job execution allowable time;
The estimated execution time of the entire job is calculated for each machine input using the machine performance unit price input step, and job execution recommended machine information is output by matching the estimated execution time of each machine with the permissible job execution time. Machine selection means,
A batch performance estimating apparatus comprising:

Images