JP4866789B2 - Performance analysis processing apparatus, performance analysis processing method, and performance analysis processing program - Google Patents

Description

  The present invention relates to a performance analysis processing apparatus, a performance analysis processing method, and a performance analysis processing program for analyzing the performance of a job net for executing a series of jobs.

  Batch processing is used to efficiently process a large amount of data collectively. A processing management apparatus for increasing the processing capacity of such batch processing has also been studied (for example, see Patent Document 1). The technique described in this document includes a job group that performs batch processing, a job monitoring device that gives priority to the job, and monitors the execution process of the job, and a scheduling device that schedules the job group. The job scheduling apparatus increases the priority of index rebuilding, and executes jobs in the order of index rebuilding and database backup. Back up the database after rebuilding the index.

In addition, when a file created by batch processing is used for business, the job may have to be completed by a predetermined end time. In particular, a plurality of jobs are associated with each other, and a subsequent job may be executed using a file generated by the preceding job. Therefore, a so-called job net is formed in which the end of the preceding job is the starting condition for the succeeding job.
JP 2002-259137 A (first page)

  Conventionally, in order to complete such a batch job by a set end time, tuning of a single job that requires a long time for job execution or renovation of an application has been performed. However, even if the time required for processing a single job is reduced, this does not necessarily directly improve the overall job net processing time. That is, a single job that requires a long time is not necessarily a critical path that determines the required time of the entire job.

  Here, when the number of jobs to be executed is large and there are a plurality of branches associated therewith, it is not easy to determine a critical path. Even if the processing time of the job included in the critical path is reduced, it may be difficult to complete the job by the end time unless the required time of the other critical path generated is short.

  The present invention has been made to solve the above-mentioned problems, and its purpose is to analyze the performance of a job net for executing a series of jobs and perform a performance analysis processing apparatus and performance for efficiently performing batch processing. To provide an analysis processing method and a performance analysis processing program.

In order to solve the above problem, the invention described in claim 1 is a job net information storage unit that records the job identifiers of jobs in a preceding / following relationship in association with each other, and a job executed for each job identifier. The performance analysis processing apparatus includes an execution result information storage unit that records the required time and a control unit that outputs an analysis result to the output unit, wherein the control unit is recorded in the job net information storage unit A job net expansion unit that generates a path expanded into a chain of job identifiers using a predecessor relationship and a time required for a job constituting the path are acquired from the execution result information storage unit and summed for each path. The path required time calculating means for calculating the path required time and the critical path with the longest path required time are identified, and the job constituting this critical path is identified. Get an identifier for the job identifier, and performance display processing means for outputting an input screen for improved value for shortening the time required for the job to the output portion via said input screen configured improved value Simulating means for updating the required time of the job with the job identifier using the path required time calculating means, the required time updated by the simulation means and the path including the job of the updated required time The required time acquired from the execution result information storage unit of the other jobs constituting the job is totaled to recalculate the required time for the path, and the performance display processing unit recalculates the required path for the path calculated by the required path time calculating unit. The gist is to repeat the process of deriving a new critical path using time.

According to a second aspect of the present invention, in the performance analysis processing apparatus according to the first aspect, the control means holds data relating to a reference time for evaluating a job constituting a critical path, and the performance display processing means The gist of the invention is to acquire a job identifier whose required time of each job is longer than a reference time from among the job identifiers constituting the critical path and display the acquired job identifier on the input screen.

According to a third aspect of the present invention, in the performance analysis processing apparatus according to the first or second aspect, the simulation unit acquires an improvement rate as an improvement value on the input screen, and the improvement is performed at a time required for each job. The gist is to calculate the predicted value by multiplying the rate and update the predicted value as the required time of the job .

According to a fourth aspect of the present invention, there is provided job net information storage means for recording job identifiers of jobs having a preceding / following relationship in association with each other, and execution result information for recording a time required for an executed job for each job identifier. A method for performing performance analysis processing using a performance analysis processing apparatus comprising storage means and control means for outputting an analysis result to an output unit, wherein the control means includes the job net information storage means A job net expansion stage for generating a path expanded in a chain of job identifiers using the preceding and succeeding relationship recorded in the path, and acquiring the required time of the job constituting the path from the execution result information storage means, The path required time calculation stage for calculating the total required time for each path, the critical path with the longest path required time is identified, and this criterion is determined. A performance display processing stage for obtaining a job identifier constituting a calpath and outputting an input screen of an improvement value for shortening the time required for the job to the job identifier to the output unit, and via the input screen A simulation stage is executed in which the required time of the job having the job identifier is updated using the set improvement value, and the path required time calculation stage includes the required time updated by the simulation stage and the updated required time. The required time acquired from the execution result information storage unit of other jobs constituting the path including the job is totaled to recalculate the required time for the path, and the performance display processing stage is performed in the required path time calculating stage. The gist is to repeat the process of deriving a new critical path using the recalculated path duration.

According to the fifth aspect of the present invention, job net information storage means for recording job identifiers of jobs in a preceding / following relationship in association with each other, and execution result information for recording a time required for an executed job for each job identifier. A program for performing a performance analysis process using a performance analysis processing apparatus comprising a storage means and a control means for outputting an analysis result to an output unit, wherein the control means includes the job net information storage means A job net expansion unit that generates a path expanded in a chain of job identifiers using the preceding and succeeding relationship recorded in the path, and obtains the required time of the job constituting the path from the execution result information storage unit, Path required time calculating means for calculating the required path time in total, and identifying the critical path with the longest path required time. A performance display processing means for obtaining a job identifier constituting a physical path and outputting an improved value input screen for shortening the time required for the job to the output portion, and setting via the input screen to function as a simulation means for has been updated the duration of the job of the job identifier with improved value, the path required time calculating means, a required time updated by the simulation means, updated durations jobs The required time acquired from the execution result information storage means of the other jobs constituting the path including the total is recalculated, and the performance required for the performance display processing means is re-calculated by the required path time calculation means. The gist is to repeat the process of deriving a new critical path using the calculated path required time.

(Function)
According to the present invention, the control unit generates a path expanded into a chain of job identifiers using the preceding / following relationship recorded in the job net information storage unit. Then, the execution times of the job identifiers constituting this path are acquired from the execution result information storage means, and the total required time is calculated for each path. The critical path with the longest path required time is identified, and the job identifiers constituting this critical path are acquired. Then, an improvement value input screen is output to the output unit for the job identifier, and the improvement value set on the input screen is supplied to the path required time calculation means. As a result, the improvement target jobs can be narrowed down to jobs constituting the critical path, and the improvement can be performed efficiently.

  According to the present invention, data relating to a reference time for evaluating a job constituting a critical path is held, and among the job identifiers constituting the critical path, a job identifier whose execution time of each job is longer than the reference time is acquired. Displayed on the input screen. Thereby, it is possible to narrow down the improvement target jobs to jobs that can be improved.

  According to the present invention, the improvement rate is acquired on the input screen, and the predicted value is calculated by multiplying the execution time of each job by the improvement rate. Thereby, the critical path can be recalculated by setting the improvement rate.

  According to the present invention, it is possible to provide a performance analysis processing apparatus, a performance analysis processing method, and a performance analysis processing program for analyzing the performance of a job net for executing a series of jobs and performing batch processing efficiently.

  Hereinafter, an embodiment embodying the present invention will be described with reference to FIGS. In the present embodiment, a performance analysis processing apparatus, a performance analysis processing method, and a performance analysis processing program for analyzing the performance of a job net and realizing an efficient job net will be described. In this embodiment, a job plan server 10, a log collection server 15, and a performance analysis device 20 connected to the network are used.

  The job plan server 10 includes a job execution plan log storage unit 11 that accumulates information related to job execution order. The job execution plan log storage unit 11 records a job execution plan log for specifying a preceding job and a succeeding job. This job execution plan log is registered immediately before the start of the job net. In the job execution plan log, data related to the preceding job identifier and data related to the subsequent job identifier are stored in association with each other.

  The log collection server 15 includes a log folder storage unit 16 that accumulates job execution plan logs and job execution result logs. The log folder storage unit 16 functions as a job net information storage unit and an execution result information storage unit. A daily folder is recorded for each job execution date, and a job execution plan log and a job execution result log are recorded for each day in the daily folder. The log collection server 15 acquires job execution plan data for the job execution date from the job plan server 10 and stores the job execution plan data in the daily folder of the job execution date. Further, the log collection server 15 monitors the job execution status and accumulates the job execution result log in the daily folder on the job execution date. In the job execution result log, data related to the start time and the required job time is stored for each job identifier of the job executed on the job execution date.

  The performance analysis apparatus 20 is a computer system that performs job net performance analysis. The performance analysis apparatus 20 includes a control unit 21 including a control unit (CPU) and a storage unit (RAM, ROM, etc.), an input unit 22 including a keyboard and a pointing device, an output unit 23 including a display, and the like.

The control unit 21 executes processing for evaluating a job net. For this reason, the control unit 21 of the performance analysis apparatus 20 executes the performance analysis processing program read from the recording medium, thereby performing a job net development stage, a path required time calculation stage,
Perform simulation stage processing. For this purpose, the control unit 21 functions as a job net expanding unit 211, a job required time calculating unit 212, a path required time calculating unit 213, a longest time path specifying unit 214, a job improvement setting unit 215, and an improvement time calculating unit 216. .

The job net expanding unit 211 expands a job into a chained path using the preceding / following relationship of each job.
The job required time calculation unit 212 calculates the job required time from the job execution result log.

The required pass time calculation means 213 calculates the required pass time for each pass using the required job time.
The longest time path specifying means 214 specifies the critical path using the required time for the path.

  The job improvement setting unit 215 holds data relating to a reference time for specifying a single job included in the specified long-time path. A single job to be improved is extracted using this reference time, and a screen for setting an improvement rate is displayed.

  The improvement time calculation means 216 recalculates the job required time in order to perform a simulation using the set improvement rate. In the present embodiment, the job improvement setting unit 215 and the improvement time calculation unit 216 function as a simulation unit.

Next, a procedure for performing performance analysis processing using the system configured as described above will be described with reference to FIG.
First, the performance analysis processing program is started in the performance analysis apparatus 20. In this case, the control unit 21 of the performance analysis apparatus 20 performs analysis target date designation processing (step S1-1). Specifically, the job net expansion unit 211 of the control unit 21 displays an analysis target date input screen on the output unit 23. Then, the job net expanding unit 211 acquires the analysis target date input on the analysis target date input screen by using the input unit 22 by the user.

  Next, the control unit 21 of the performance analysis apparatus 20 executes a job net development process (step S1-2). Specifically, the job net expansion unit 211 of the control unit 21 accesses the log folder storage unit 16 and acquires a job execution plan log from the daily folder of the analysis target date. Then, the job net expanding means 211 expands using the preceding and succeeding dependency relationships of all job nets.

  Specifically, first, the job net expanding unit 211 specifies a terminal job included in the analysis target range. Next, the preceding job is specified from the last job to the starting job by using the preceding / following dependency relationship. Then, a path in which the job identifier of the preceding job is arranged with the job identifier of the subsequent job is generated and recorded in the memory. Here, each time the job net expansion unit 211 detects a plurality of jobs having a preceding dependency relationship, it copies the path of the job identifier of “end job to start job” generated so far and creates a new independent Recorded in the memory as a completed path. The above processing is repeated for each path until the start job is reached.

Next, the control unit 21 of the performance analysis apparatus 20 executes a job start time and required time totaling process (step S1-3). Specifically, the required job time calculation unit 212 of the control unit 21 accesses the log folder storage unit 16 and acquires a job execution result log from the daily folder. Then, the job required time calculation unit 212 calculates the job required time from the job execution result log, and expands the start time and the job required time in the memory in association with the job identifier.

  Next, the control unit 21 of the performance analysis apparatus 20 performs a calculation process for the time required for each path (step S1-4). Specifically, the required path time calculation means 213 of the control unit 21 acquires the required job time for the single job identifiers that constitute each path, and calculates the required path time by summing the required job time for each path. .

  Next, the control unit 21 of the performance analysis apparatus 20 executes sort processing according to the time required for each pass (step S1-5). Specifically, the longest time path specifying unit 214 of the control unit 21 rearranges the paths in descending order using the required time for the path. Here, the path with the longest time becomes the critical path.

  Next, the control unit 21 of the performance analysis apparatus 20 executes performance display processing (step S1-6). Specifically, the job improvement setting unit 215 of the control unit 21 specifies a path having a long required time from the top. The job improvement setting unit 215 extracts a single job that is included in the specified long-time path and that is longer than the reference time. Then, the job improvement setting means 215 displays the performance evaluation screen 500 shown in FIG. The performance evaluation screen 500 includes a job list 501 that displays a list of job identifiers of single jobs extracted from long-time paths. In this case, an improvement rate input field 502 for setting an improvement rate as the improvement value of each job is displayed in association with each job identifier of the specified single job. Furthermore, the performance evaluation screen 500 includes a simulation execution button and a reset button.

  Further, the job improvement setting unit 215 generates a job count list by time 503 in which the number of jobs executed by time slot is displayed, and displays it on the performance evaluation screen 500. Further, the performance evaluation screen 500 outputs a total time display 504 in which the single job execution time is totaled for each path.

  And the control part 21 of the performance analysis apparatus 20 performs the determination process about the completion | finish of an analysis process (step S1-7). Specifically, when the performance of a single job cannot be improved, a reset button on the performance evaluation screen is selected. In this case (in the case of “YES” in step S1-7), the control unit 21 ends the analysis process. In this case, the function of returning to the state immediately after the critical path is derived from the execution plan log and the execution result log before the simulation is implemented is realized.

  On the other hand, when the simulation execution button is selected (in the case of “NO” in step S1-7), the control unit 21 of the performance analysis apparatus 20 executes a required time prediction process (step S1-8). Specifically, the improvement time calculation means 216 of the control unit 21 calculates the predicted value by multiplying the job required time of each job identifier by the improvement rate set in the improvement rate input field 502. Then, the improvement time calculation means 216 updates the required job time developed in the memory using this predicted value.

  And the control part 21 of the performance analyzer 20 repeats the process from step S1-4. In this case, the control unit 21 of the performance analysis apparatus 20 calculates the required path time using the required job time updated in the memory. In this case, since the time required for the improved job is shortened, the time required for the path including the job (including the critical path) is also shortened. As a result, the critical path is recalculated.

As described above, according to the present embodiment, the following effects can be obtained.
In the above embodiment, the control unit 21 of the performance analysis apparatus 20 executes the job net development process (step S1-2) and calculates the time required for each path (step S1-4). And the control part 21 of the performance analysis apparatus 20 performs a sort process according to the required time of each path | pass (step S1-5). Thereby, the critical path can be specified.

  Furthermore, the control unit 21 of the performance analysis apparatus 20 executes performance display processing (step S1-6). Here, the job improvement setting unit 215 displays the performance evaluation screen 500 on the output unit 23. The performance evaluation screen 500 includes a job list 501 that displays a list of job identifiers of single jobs included in a long-time path. As a result, the improvement target can be set while confirming the single jobs constituting the critical path. When many jobs have a preceding / following relationship, it may be difficult to determine which job should be adjusted. In the present invention, since the improvement of the job constituting the critical path is preferentially performed, the entire time can be efficiently reduced.

  In the above embodiment, the job improvement setting unit 215 extracts a single job that is included in the specified long-time path and is longer than the reference time. This makes it possible to display a list of jobs whose performance should be improved preferentially in order to reduce the overall processing time.

  In the above embodiment, the job improvement setting unit 215 generates the hourly job number list 503 that displays the number of jobs executed by time period, and displays the job number list 503 on the performance evaluation screen 500. By using this hourly job number list 503, it becomes possible to analyze the load balance by comparing the number of hourly job executions and the CPU status.

In addition, you may change the said embodiment into the following aspects.
In the above embodiment, the improvement rate is set as the improvement value, but the present invention is not limited to this, and the required time itself can be set.

  In the above embodiment, the control unit 21 of the performance analysis apparatus 20 executes the analysis target date designation process (step S1-1). Here, the job net expanding unit 211 acquires the analysis target date input by the user on the analysis target date input screen using the input unit 22. The method for specifying the analysis target date is not limited to the designation by the user, and the performance analysis apparatus 20 may be configured to specify it. In this case, the performance analysis apparatus 20 holds data related to the job end time. Then, the control unit 21 of the performance analysis device 20 identifies the date on which the job could not be completed by the job end time as the analysis target date.

  In the above embodiment, the job improvement setting unit 215 generates the hourly job number list 503 that displays the number of jobs executed by time period, and displays the job number list 503 on the performance evaluation screen 500. In this case, when the number of simultaneous jobs exceeds the number of simultaneously executable, it is possible to perform a simulation by inserting a waiting time until the number of simultaneous jobs becomes equal to or less than the number of simultaneously executable.

In the above embodiment, the control unit 21 of the performance analysis apparatus 20 performs a required time prediction process (step S1-8). Specifically, the improvement time calculation means 216 of the control unit 21 multiplies the job required time of each job identifier by the improvement rate set in the improvement rate input field 502. Then, the improvement time calculation means 216 updates the required job time stored in the memory. Instead, the control unit 21 of the performance analysis device 20 can set the improvement rate by the performance analysis device 20 with the improvement rate set in the improvement rate input field 502 as a maximum value. In this case, the target end time is stored in the control unit 21 of the performance analysis apparatus 20. Then, the improvement rate is sequentially multiplied from the longest job among the single jobs constituting the critical path, and the estimated end time is calculated each time. Then, the improvement rate of each single job when the target end time reaches the predicted end time is output. Thereby, setting of an excessive improvement rate can be suppressed and optimization of an improvement rate can be aimed at.

1 is a schematic diagram of a system according to an embodiment of the present invention. Explanatory drawing of the process sequence of this embodiment. Explanatory drawing of the screen displayed on the output part of a performance analyzer.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Job plan server, 11 ... Job execution plan log storage part, 15 ... Log collection server, 16 ... Log folder storage part, 20 ... Performance analysis apparatus, 21 ... Control part, 211 ... Job net expansion means, 212 ... Job requirement Time calculation means, 213... Path required time calculation means, 214... Longest time path identification means, 215... Job improvement setting means, 216... Improvement time calculation means, 22.

Claims (5)

A job net information storage means for recording the job identifiers of jobs in a preceding / following relationship in association with each other;
Execution result information storage means for recording the time required for the executed job for each job identifier;
A performance analysis processing apparatus comprising a control means for outputting an analysis result to an output unit,
The control means is
Job net expansion means for generating a path expanded in a chain of job identifiers using the preceding and succeeding relationship recorded in the job net information storage means;
Path required time calculating means for acquiring the required time of the job constituting the path from the execution result information storage means and calculating the required time for each path by totaling;
The critical path with the longest path required time is specified, a job identifier constituting the critical path is acquired, and an input screen for an improvement value for shortening the required time for the job is displayed for the job identifier. Performance display processing means for outputting to
Simulation means for updating the time required for the job with the job identifier using the improvement value set via the input screen,
The required path time calculation means includes the required time updated by the simulation means and the required time acquired from the execution result information storage means of other jobs that constitute the path including the job with the updated required time. Total to recalculate the path duration,
The performance analysis processing device, wherein the performance display processing means repeats a new critical path derivation process using the path required time recalculated by the path required time calculating means . The control means holds data relating to a reference time for evaluating a job constituting a critical path,
The said performance display process means acquires the job identifier whose required time of each job is longer than reference | standard time among the job identifiers which comprise a critical path, and displays it on the said input screen. Performance analysis processor. The simulation means obtains an improvement rate as an improvement value on the input screen, calculates a predicted value by multiplying the required time of each job by the improvement rate, and updates the predicted value as a required time of the job The performance analysis processing device according to claim 1, wherein A job net information storage means for recording the job identifiers of jobs in a preceding / following relationship in association with each other;
Execution result information storage means for recording the time required for the executed job for each job identifier;
A method for performing a performance analysis process using a performance analysis processing apparatus having a control means for outputting an analysis result to an output unit,
The control means is
A job net expansion stage for generating a path expanded in a chain of job identifiers using the preceding / following relationship recorded in the job net information storage means;
The required time of the job constituting the path is acquired from the execution result information storage means, and the required path time calculating step for calculating the required time for the path by totaling for each path;
The critical path with the longest path required time is specified, a job identifier constituting the critical path is acquired, and an input screen for an improvement value for shortening the required time for the job is displayed for the job identifier. Performance display processing stage to output to,
Performing a simulation step of updating the time required for the job with the job identifier using the improvement value set via the input screen;
In the path required time calculation step, the required time updated by the simulation step and the required time acquired from the execution result information storage unit of other jobs that constitute the path including the job of the updated required time are obtained. Total to recalculate the path duration,
In the performance display processing step, a new critical path derivation process is repeated using the path required time recalculated in the path required time calculating step . A job net information storage means for recording the job identifiers of jobs in a preceding / following relationship in association with each other;
Execution result information storage means for recording the time required for the executed job for each job identifier;
A program for performing a performance analysis process using a performance analysis processing apparatus having a control means for outputting an analysis result to an output unit,
The control means;
A job net expansion unit that generates a path expanded in a chain of job identifiers using the preceding / following relationship recorded in the job net information storage unit;
Path required time calculating means for acquiring the required time of the job constituting the path from the execution result information storage means and calculating the required path time for each path;
The critical path with the longest path required time is specified, a job identifier constituting the critical path is acquired, and an input screen for an improvement value for shortening the required time for the job is displayed for the job identifier. Performance display processing means to output to
Function as simulation means for updating the time required for the job with the job identifier using the improvement value set via the input screen;
The required path time calculation means includes the required time updated by the simulation means and the required time acquired from the execution result information storage means of other jobs that constitute the path including the job with the updated required time. Total to recalculate the path duration,
The performance display processing means repeats a new critical path derivation process using the required path time recalculated by the required path time calculating means .

Images