JP5614300B2 - Processing time prediction apparatus, processing time prediction method, and processing time prediction program - Google Patents

Description

  The present invention relates to a processing time prediction apparatus, a processing time prediction method, and a processing time prediction program.

  Some business systems process data by online processing when the system is operating during the daytime, and update the database by batch processing when the system is stopped at night. In such a business system, it is necessary to efficiently perform batch processing before operating the system. Furthermore, if the system is always operating, there are restrictions on the execution of batch processing so that the operating system is not affected. For example, constraints such as assigning resources that do not affect the operation of the system and executing batch processing and setting a deadline at the end time of the batch processing are imposed.

  In order to efficiently process the batch process under such circumstances, it is desirable to predict the processing time required for the batch process before executing the batch process. Patent Documents 1 to 3 below disclose techniques for predicting the processing time required for batch processing.

  The technique described in Patent Document 1 measures the processing time for each job included in the target batch processing, and predicts the processing time required for the entire batch processing based on the past actual processing time. The technique described in Patent Document 2 predicts a job processing time with actual data based on a processing time when a job is executed using test data. The technique described in Patent Document 3 detects an error between the predicted processing time and the actual processing time, and recalculates the scheduled end time of processing based on this error.

JP 2009-230484 A JP 2002-297407 A JP-A-11-53227

  By the way, in the technologies described in Patent Documents 1 to 3 described above, for example, when an error occurs during job execution and rollback is performed, the number of data reprocessed by rollback cannot be grasped. Therefore, since the time required for rollback cannot be calculated with high accuracy, a divergence occurs between the actual processing time and the predicted processing time when an error occurs.

  The present invention has been made to solve the above-described problem, and is capable of improving the prediction accuracy of the processing time required for batch processing even if an error occurs, a processing time prediction method, and a processing. The purpose is to provide a time prediction program.

  The processing time prediction apparatus of the present invention is a processing time prediction apparatus that predicts the processing time required for batch processing including a plurality of jobs, and uses the number of processing data in units of jobs and the number of processing data in units of transactions. A transaction number calculation unit that calculates the number of transactions, a provisional processing time measurement unit that provisionally executes a process for one transaction for each job, and measures a provisional processing time that is a processing time for one transaction for each job; Using the number of transactions and the provisional processing time, a predicted processing time calculation unit that calculates a predicted processing time that is a processing time assumed for each job, and an actual processing time that is a processing time required when each job is executed are measured. An actual processing time measurement unit, a count unit that counts the number of errors that occur when each job is executed, An error rate calculation unit that calculates an error rate of the predicted processing time using the number of transactions, the actual processing time, and the number of errors, and the prediction of an unprocessed job using the error rate and the temporary processing time A predicted processing time recalculation unit that recalculates the processing time.

  The processing time prediction method of the present invention is a processing time prediction method for predicting the processing time required for batch processing including a plurality of jobs, and uses the number of processing data for each job and the number of processing data for each transaction. A transaction number calculating step for calculating the number of transactions, a provisional processing time measuring step for temporarily executing processing for one transaction for each job, and measuring a provisional processing time that is a processing time for one transaction for each job, Using the number of transactions and the provisional processing time, an estimated processing time calculating step for calculating an estimated processing time that is an estimated processing time for each job, and an actual processing time that is a processing time required for executing each job are measured. Counts the actual processing time measurement step and the number of errors that occur when each job is executed. An error rate calculating step of calculating an error rate of the predicted processing time using the number of transactions, the actual processing time, and the number of errors, and using the error rate and the provisional processing time, And a predicted processing time recalculation step for recalculating the predicted processing time of the processing job.

  The processing time prediction program of the present invention causes a computer to execute each step included in the processing time prediction method.

  According to the present invention, it is possible to improve the prediction accuracy of the processing time required for batch processing even if an error occurs.

It is a block diagram which illustrates functional composition of a processing time prediction device in an embodiment. It is a figure which illustrates the data structure of a job information table. It is a figure which illustrates the data structure of a process related information table. It is a figure which illustrates the data structure of the information table for prediction. It is a flowchart for demonstrating the process performed before starting a batch process. It is a flowchart for demonstrating the process sequence of a transaction number calculation process. It is a flowchart for demonstrating the process sequence of an estimated process time calculation process. It is a flowchart for demonstrating the process performed after starting a batch process. It is a flowchart for demonstrating the process sequence of an execution condition reflection process. It is a flowchart for demonstrating the process sequence of prediction process time recalculation process. It is a flowchart for demonstrating the process sequence of a process order adjustment process. It is a flowchart for demonstrating the process sequence of the deletion object job selection process.

  DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments of a processing time prediction apparatus, a processing time prediction method, and a processing time prediction program according to the present invention will be described with reference to the accompanying drawings. The processing time prediction device is a computer device that predicts processing time of batch processing including a plurality of jobs.

  Examples of batch processing including a plurality of jobs include batch processing including the following four jobs. (1) Job 1; data summarization processing for summing up sales for each store based on the sales data updated by online processing and registering it in the sales count table of each store. (2) Job 2; merchandise inventory quantity confirmation processing for calculating the current merchandise inventory quantity based on the sales total data of each shop and the merchandise inventory quantity of each shop and registering it in each shop information table. (3) Job 3: Product sales prediction processing that predicts future sales based on the sales total data of each store and the sales transition data of each store so far, and calculates the optimal number of product stocks. (4) Job 4: Product order processing for determining the number of product orders for each store based on the sales transition data so far. If there is no sales trend data, the forecast information up to the previous day is taken over and the stock quantity is the same as the previous day. In this case, the number of orders is the same as the number of products sold this time. In order to order a product to the outside, the product order processing needs to order the product by a predetermined date and time.

  The processing time prediction apparatus in the embodiment physically includes, for example, a CPU (Central Processing Unit), a storage device, and an input / output interface. The storage device includes, for example, a ROM (Read Only Memory) and HDD (Hard Disk Drive) that store programs and data processed by the CPU, and a RAM (Random Access Memory) mainly used as various work areas for control processing. ) Etc. are included. These elements are connected to each other via a bus. The CPU executes the program stored in the ROM, and processes the message received via the input / output interface and the data expanded in the RAM, thereby realizing the function of each unit in the processing time prediction apparatus described later. be able to.

  With reference to FIG. 1, the functional configuration of the processing time prediction apparatus in the embodiment will be described. Functionally, the processing time prediction device 1 functionally includes, for example, a transaction number calculation unit 11, a job provisional execution unit 12, a predicted processing time calculation unit 13, a job execution unit 21, an execution status reflection unit 22, and a prediction. A processing time recalculation unit 23 and a processing order adjustment unit 31 are included.

  The transaction number calculation unit 11 calculates the number of transactions executed by the job for each job. The number of transactions can be calculated, for example, by the following procedure.

  First, the transaction number calculation unit 11 refers to the job information table 41 and the processing related information table 42 and acquires the number of processing data for each job and the number of processing data for each transaction. Subsequently, the transaction number calculation unit 11 calculates the number of transactions for each job by dividing the number of processing data for each job by the number of processing data for each transaction. Subsequently, the transaction number calculation unit 11 stores the calculated number of transactions in the prediction information table 43.

  The data structure of the job information table 41 will be described with reference to FIG. The job information table 41 includes, for example, a job ID item, a use server ID item, and a use resource item as data items. The job ID item stores identification information that uniquely identifies a job. The use server ID item stores identification information that uniquely identifies a server used by a job. The use resource item stores information on data resources used by the job. The above-mentioned number of processing data for each job can be acquired from a data resource (for example, a database) that can be specified based on the use resource item.

  The data configuration of the processing related information table 42 will be described with reference to FIG. The processing related information table 42 includes, for example, a job ID item, a processing priority item, a required preceding job item, a transaction unit processing number item, a processing restriction item, a processing omissible flag item, and a processing order item as data items.

  The job ID item stores identification information that uniquely identifies a job. The processing priority item stores the job processing priority. The mandatory preceding job item stores the job ID of the preceding job that must be completed before the job processing is started. If there is no mandatory preceding job, leave it blank. The transaction unit processing count field stores the processing data count of each transaction. The process restriction item stores a process end time limit. If no end deadline is required, leave it blank. The process omissible flag item stores a flag indicating whether or not a job can be omitted. The processing order item stores the execution order of jobs in batch processing.

  The data configuration of the prediction information table 43 will be described with reference to FIG. The prediction information table 43 includes, for example, a job ID item, a transaction number item, a provisional processing time item, a prediction processing time item, a prediction error number item, an error rate item, and a processing state item as data items.

  The job ID item stores identification information that uniquely identifies a job. The transaction number item stores the number of transactions calculated by the transaction number calculation unit 11. The provisional processing time item stores provisional processing time (processing time for one transaction) measured by a provisional processing time measuring unit 121 described later. The prediction processing time item stores a prediction processing time calculated by a prediction processing time calculation unit 13 described later. The prediction error number item stores a prediction error number calculated by a prediction error number calculation unit 221 described later. The error rate item stores an error rate of the prediction processing time calculated by the error rate calculation unit 222 described later. The processing status item stores the processing status of the job. For example, “COMPLETED”, “RUNNING”, “NOT_RUNNING”, or the like can be used as the job processing state.

  The temporary job execution unit 12 shown in FIG. 1 temporarily executes processing for one transaction for each job. The temporary execution process is not a real process, but a temporary process for the purpose of measuring the processing time. Therefore, data is not written in a location that affects the actual processing. The temporary job execution unit 12 includes a temporary processing time measurement unit 121.

  The provisional processing time measurement unit 121 measures the provisional processing time by measuring the time required to provisionally execute the process for one transaction. The temporary processing time measuring unit 121 stores the measured temporary processing time in the temporary processing time item of the prediction information table 43.

  The predicted processing time calculation unit 13 calculates the predicted processing time, which is a processing time assumed for each job, using the number of transactions executed by each job and the provisional processing time that is the processing time per transaction. The prediction processing time calculation unit 13 stores the calculated prediction processing time in the prediction processing time item of the prediction information table 43.

  The job execution unit 21 loads and executes the job according to the processing order data in the processing related information table 42. The job execution unit 21 includes an actual processing time measurement unit 211 and a count unit 212.

  The actual processing time measuring unit 211 measures the actual processing time by measuring the time required when the job is executed, and stores it in the actual processing time item of the execution status information table 44 described later. The count unit 212 counts the number of errors that occur when each job is executed, and stores it in the error number item of the execution status information table 44 described later.

  Here, the execution status information table 44 includes, for example, a job ID item, an actual processing time item, an error number item, and a processing result item as data items. The job ID item stores identification information that uniquely identifies a job. The actual processing time item stores the actual processing time measured by the actual processing time measuring unit 211. The number of errors item stores the number of errors counted by the counting unit 212. The processing result item stores the status of the execution result of the job executed by the job execution unit 21. As the state of the execution result, for example, a normal end state, an abnormal end state, or a process interruption state can be used.

  The execution status reflecting unit 22 determines whether or not the job execution result is in a normal end state. If the job execution result is in a normal end state, the execution status of the job that has ended normally is reflected in an unprocessed job. Execute the process. On the other hand, if the process is abnormally terminated or interrupted, the process is terminated without executing the process reflected in the unprocessed job. The execution status reflecting unit 22 includes a prediction error number calculating unit 221 and an error rate calculating unit 222 as various functions for performing processing to be reflected in the unprocessed job.

  The predicted error number calculation unit 221 acquires the number of errors corresponding to the processed job from the execution status information table 44, and calculates the number of errors per job (average number of errors), thereby predicting the unprocessed job. Calculate the number of errors. The prediction error number calculation unit 221 stores the calculated prediction error number in the prediction error number item corresponding to the unprocessed job in the prediction information table 43.

  The error rate calculation unit 222 calculates the error rate of the prediction processing time and stores it in the error rate corresponding to the unprocessed job in the prediction information table 43. The error rate can be calculated, for example, by the following procedure.

  First, the error rate calculation unit 222 acquires the number of transactions corresponding to the processed job, the predicted processing time, and the provisional processing time from the prediction information table 43, and the number of errors corresponding to the processed job from the execution status information table 44. And get the actual processing time. Subsequently, the error rate calculation unit 222 calculates the number of transactions actually processed by adding the number of errors to the number of transactions. This is because rollback adopted when dealing with errors in batch processing resumes processing from the beginning of the transaction that was being processed when the error occurred, so the error count is actually added to the transaction count. This is because the number of transactions processed can be calculated.

  Subsequently, the error rate calculation unit 222 calculates the average value of the processing time spent per transaction by using the number of transactions actually processed and the actual processing time. Subsequently, the error rate calculation unit 222 calculates the error rate of the predicted processing time based on the difference between the average processing time spent per transaction and the provisional processing time. Subsequently, the error rate calculation unit 222 stores the calculated error rate of the prediction processing time in an error rate item corresponding to an unprocessed job in the prediction information table 43.

  The prediction processing time recalculation unit 23 acquires the prediction processing time and the error rate corresponding to the unprocessed job from the prediction information table 43. The predicted processing time recalculation unit 23 recalculates the predicted processing time of the unprocessed job using the acquired predicted processing time and error rate. The predicted processing time recalculation unit 23 stores the calculated predicted processing time in the predicted processing time corresponding to the unprocessed job in the prediction information table 43.

  The processing order adjustment unit 31 adjusts the job processing order based on the predicted job processing time, job processing constraints, and job processing priority. The processing order adjustment unit 31 includes a determination unit 311, a warning output unit 312, and a job exclusion unit 313.

  The determination unit 311 determines whether there is a job that exceeds the end deadline based on the scheduled end time of each job. This determination can be performed, for example, according to the following procedure.

  First, the processing order adjustment unit 31 calculates the scheduled start time and the scheduled end time of each job by mapping the predicted processing time of each job according to the processing order data in the processing related information table 42, and the processing order Generate a list. Subsequently, the determination unit 311 determines whether or not the scheduled end time of the job exceeds the end time limit for the job whose end time limit is stored in the process constraint of the process related information table 42.

  The warning output unit 312 outputs a warning when the determination unit 311 determines that there is a job that exceeds the end deadline. Examples of the warning output include an output for displaying a warning message and a warning mark, and an output of a warning sound.

  If the determination unit 311 determines that there is a job that exceeds the end deadline, the job exclusion unit 313 excludes a job whose processing omissibility flag in the processing related information table 42 is “YES” from the target of batch processing. . If there are a plurality of jobs whose processing omissibility flag is “YES”, for example, they can be sequentially excluded according to the following criteria. (1) Select a job having the lowest processing priority as an exclusion target. (2) Select a job that is not designated as a target of the mandatory preceding job as an exclusion target. (3) The job having the longest total time of the predicted processing time of the candidate job to be excluded and the predicted processing time of the job designated as the essential preceding job of the candidate job to be excluded is selected as the target to be excluded.

  Next, the operation of the processing time prediction apparatus in the embodiment will be described with reference to FIGS. First, with reference to FIG. 5, the process performed before starting a batch process is demonstrated.

  First, the transaction number calculation unit 11 executes a transaction number calculation process (step S101). Details of the transaction count calculation process will be described later.

  Subsequently, the provisional processing time measurement unit 121 executes provisional execution processing for provisionally executing processing for one transaction, and stores the measured provisional processing time in the provisional processing time item of the prediction information table 43 (step S102). .

  Subsequently, the prediction processing time calculation unit 13 executes a prediction processing time calculation process (step S103). The processing procedure of the predicted processing time calculation process will be described later.

  Subsequently, the determination unit 311 calculates the scheduled end time of each job based on the predicted processing time calculated in step S103, and determines whether there is a job that exceeds the end deadline (step S104). . If this determination is NO (step S104; NO), the processing of this operation is terminated.

  On the other hand, when it is determined in step S104 that there is a job that exceeds the end deadline (step S104; YES), the warning output unit 312 outputs a warning (step S105). And the process of this operation | movement is complete | finished.

  With reference to FIG. 6, the processing procedure of the transaction number calculation process executed in step S101 described above will be described.

  First, the transaction number calculation unit 11 determines whether there is a job whose transaction number has not yet been calculated among jobs to be batch processed (step S201). If this determination is NO (step S201; NO), the transaction number calculation process ends.

  On the other hand, when it is determined in step S201 that an uncalculated job exists (step S201; YES), the transaction number calculation unit 11 refers to the processing related information table 42 and corresponds to the job to be calculated. The number of transaction unit processes is acquired, and the number of job unit processes is acquired from a database that can be specified based on the use resource item of the job information table 41 (step S202).

  Subsequently, the transaction number calculation unit 11 calculates the transaction number of the calculation target job by dividing the job unit processing number acquired in step S202 by the transaction unit processing number (step S203).

  Subsequently, the transaction number calculation unit 11 stores the transaction number calculated in step S203 in the transaction number item of the prediction information table 43 (step S204). Then, the process proceeds to step S201.

  With reference to FIG. 7, the process sequence of the estimated process time calculation process performed by step S103 mentioned above is demonstrated.

  First, the predicted processing time calculation unit 13 determines whether there is a job for which the predicted processing time has not yet been calculated among jobs to be batch processed (step S301). If this determination is NO (step S301; NO), the prediction processing time calculation process ends.

  On the other hand, when it is determined in step S301 that an uncalculated job exists (step S301; YES), the prediction processing time calculation unit 13 refers to the prediction information table 43 and corresponds to the calculation target job. The information for prediction to acquire is acquired (step S302).

  Subsequently, the prediction processing time calculation unit 13 calculates the prediction processing time to be calculated using the number of transactions and the provisional processing time included in the prediction information acquired in step S302 (step S303).

  Subsequently, the prediction processing time calculation unit 13 stores the prediction processing time calculated in step S303 in the prediction processing time item of the prediction information table 43 (step S304). Then, the process proceeds to step S301.

  Next, with reference to FIG. 8, the process performed after starting a batch process is demonstrated.

  First, the job execution unit 21 sequentially executes jobs according to the processing order data in the processing related information table 42, and sets the actual processing time measured by the actual processing time measurement unit 211 as an actual processing time item in the execution status information table 44. The number of errors counted by the counting unit 212 is stored in the error number item of the execution status information table 44 (step S401).

  Subsequently, the execution status reflection unit 22 executes execution status reflection processing (step S402). The processing procedure of the execution status reflection process will be described later.

  Subsequently, the prediction processing time recalculation unit 23 executes a prediction processing time recalculation process (step S403). The processing procedure of the predicted processing time recalculation process will be described later.

  Subsequently, the process order adjustment unit 31 executes a process order adjustment process (step S404). And the process of this operation | movement is complete | finished. Here, the processing procedure of the processing order adjustment processing will be described later.

  With reference to FIG. 9, the processing procedure of the execution status reflection process executed in step S402 described above will be described.

  First, the execution status reflection unit 22 refers to the execution status information table 44 and acquires execution status information corresponding to the processed job (step S501).

  Subsequently, the execution status reflection unit 22 determines whether the status of the job execution result is a normal end status based on the processing result item of the execution status information acquired in step S501 (step S502). . If this determination is NO (step S502; NO), the execution status reflection process is terminated.

  On the other hand, when it is determined in step S502 that the job execution result is in the normal end state (step S502; YES), the predicted error number calculation unit 221 performs each process from the execution status information table 44. The number of errors corresponding to the job is acquired, the average value of the number of errors is calculated, and the calculated average value is stored in the prediction error number item corresponding to the unprocessed job in the prediction information table 43 (reflected in the number of prediction errors) (Step S503).

  Subsequently, the error rate calculation unit 222 calculates an error rate of the prediction processing time, and stores the calculated error rate in an error rate item corresponding to an unprocessed job in the prediction information table 43 (reflected in the error rate). (Step S504). Then, the execution status reflection process ends.

  With reference to FIG. 10, the processing procedure of the predicted processing time recalculation process executed in step S <b> 403 described above will be described.

  First, the predicted processing time recalculation unit 23 determines whether there is an unprocessed job for which the predicted processing time has not yet been recalculated (step S601). If this determination is NO (step S101; NO), the prediction processing time recalculation process is terminated.

  On the other hand, when it is determined in step S601 that an unrecalculated unprocessed job exists (step S601; YES), the prediction processing time recalculation unit 23 refers to the prediction information table 43 and recalculates. Information for prediction corresponding to the target job is acquired (step S602).

  Subsequently, the prediction processing time recalculation unit 23 recalculates the prediction processing time of the recalculation target job using the prediction processing time and the error rate included in the prediction information acquired in step S602 (step S603). .

  Subsequently, the prediction processing time recalculation unit 23 stores the prediction processing time calculated in step S603 in the prediction processing time item corresponding to the recalculation target job in the prediction information table 43 (reflects in the prediction processing time). (Step S604). Then, the process proceeds to step S601.

  With reference to FIG. 11, the process sequence of the process order adjustment process performed in step S404 described above will be described.

  First, the processing order adjustment unit 31 refers to the processing order data item in the processing related information table 42 to generate and store a processing order list of unprocessed jobs (step S701).

  Subsequently, the determination unit 311 determines, for the unprocessed jobs in the processing order list, the scheduled end time of the unprocessed job for the job whose end limit is stored in the processing constraint item of the processing related information table 42. It is determined whether or not it exceeds (step S702). If this determination is NO (step S702; NO), the process proceeds to step S710 described later.

  On the other hand, if it is determined in step S702 that there is an unprocessed job whose scheduled end time exceeds the end deadline (step S702; YES), the warning output unit 312 outputs a warning (step S703). ).

  Subsequently, the job exclusion unit 313 executes a deletion target job selection process (step S704). The processing procedure of the deletion target job selection process will be described later.

  Subsequently, the determination unit 311 determines whether or not a deletion target job has been selected (step S705). If this determination is NO (step S705; NO), the warning output unit 312 outputs a message indicating that there is no job to be deleted (step S706), and the process proceeds to step S710 described later.

  On the other hand, when it is determined in step S705 that the deletion target job has been selected (step S705; YES), the job exclusion unit 313 excludes the selected deletion target job from the stored processing order list. Then, the processing order adjustment unit 31 regenerates the processing order list and saves it by overwriting (step S708).

  Subsequently, the determination unit 311 determines, for the unprocessed jobs in the processing order list, the scheduled end time of the unprocessed job for the job whose end limit is stored in the processing constraint item of the processing related information table 42. It is determined whether or not it exceeds (step S709). If this determination is YES (step S709; YES), the process proceeds to step S704 described above.

  On the other hand, if it is determined in step S709 that there is no unprocessed job whose scheduled end time exceeds the end deadline (step S709; NO), the processing order adjustment unit 31 stores the stored processing. An order list is output (step S710). Then, the processing order adjustment process ends.

  With reference to FIG. 12, the process procedure of the deletion target job selection process executed in step S704 described above will be described.

  First, the job exclusion unit 313 determines whether there is a job in which “YES” is stored in the process omissibility flag item of the process related information table 42 among the unprocessed jobs in the process order list. (Step S801). If this determination is NO (step S101; NO), the job exclusion unit 313 outputs an error message (step S811), and ends the deletion target job selection process.

  On the other hand, when it is determined in step S801 that there is a job in which “YES” is stored in the process omissible flag item (step S801; YES), the job exclusion unit 313 sets the process omissible flag item. Processing related information corresponding to a job storing “YES” is acquired from the processing related information table 42 (step S802).

  Subsequently, the job exclusion unit 313 determines whether there are a plurality of jobs that are candidates for deletion based on the processing related information acquired in step S802 (step S803). If this determination is NO (step S803; NO), the job exclusion unit 313 outputs the job corresponding to the processing related information acquired in step S802 as a deletion target job (step S810), and selects the deletion target job. The process ends.

  On the other hand, when it is determined in step S803 that there are a plurality of jobs that are candidates for deletion (step S803; YES), the job exclusion unit 313 determines the processing priority included in the processing related information corresponding to each job. And the job with the lowest processing priority is selected as a candidate job to be deleted (step S804).

  Subsequently, the job exclusion unit 313 determines whether there are a plurality of deletion target candidate jobs selected in step S804 (step S805). When this determination is NO (step S805; NO), the job exclusion unit 313 outputs the deletion target candidate job selected in step S804 as a deletion target job (step S810), and ends the deletion target job selection process. To do.

  On the other hand, when it is determined in step S805 that there are a plurality of deletion target candidate jobs (step S805; YES), the job exclusion unit 313 includes a plurality of deletion target candidate jobs in the processing related information table 42. Only jobs that are not designated as mandatory preceding jobs in the mandatory preceding job item are selected as candidate jobs to be deleted (step S806).

  Subsequently, the job exclusion unit 313 determines whether a deletion target candidate job has been selected in step S806 (step S807). If this determination is NO (step S807; NO), the job exclusion unit 313 outputs an error message (step S811), and ends the deletion target job selection process.

  On the other hand, if it is determined in step S807 that a deletion target candidate job has been selected (step S807; YES), the job exclusion unit 313 determines whether there are a plurality of selected deletion target candidate jobs. (Step S808). If this determination is NO (step S808; NO), the job exclusion unit 313 outputs the deletion target candidate job selected in step S806 as a deletion target job (step S810), and ends the deletion target job selection process. To do.

  On the other hand, when it is determined in step S808 that there are a plurality of deletion target candidate jobs (step S808; YES), the job exclusion unit 313 determines that the deletion target candidate job among the plurality of deletion target candidate jobs exists. The total time of the predicted processing time and the predicted processing time of the job specified as the mandatory preceding job of the deletion target candidate job is calculated, and the job having the longest total time is selected as the deletion target candidate job (step S809). .

  Subsequently, the job exclusion unit 313 outputs the deletion target candidate job selected in step S809 as a deletion target job (step S810), and ends the deletion target job selection process.

  As described above, according to the processing time prediction apparatus 1 in the embodiment, the processing time can be predicted by temporarily executing the processing for one transaction before the job is actually executed. It is possible to calculate the predicted processing time of an unprocessed job while minimizing the effect on the job.

  In addition, since the error rate of the predicted processing time can be calculated based on the execution status information when the job is actually executed and reflected in the prediction information, the predicted processing time can be recalculated. An error from the actual processing time can be reflected.

  Furthermore, since the number of predicted errors can be calculated based on the execution status information when the job is actually executed and reflected in the prediction information, the predicted processing time can be recalculated. It is possible to reflect the situation.

  Therefore, even when an error occurs, it is possible to improve the prediction accuracy of the processing time required for batch processing.

[Modification]
Note that the above-described embodiment is merely an example, and does not exclude various modifications and technical applications that are not explicitly described in the embodiment. That is, the present invention can be implemented by being modified into various forms without departing from the spirit of the present invention.

  For example, in the above-described embodiment, the process order adjustment process by the process order adjustment unit 31 is performed during the process after the start of the batch process, but may be performed during the process before the start of the batch process. As a result, when there is a job that exceeds the end deadline in the state before the start of batch processing, it is possible to exclude an optional job or the like from the batch processing and output a warning.

  In the embodiment described above, the processing order list after the processing order adjustment is output, but the present invention is not limited to this. For example, the processing order adjusted by the processing order adjustment process may be dynamically reflected (updated).

  Further, in the above-described embodiment, in the rollback processing at the time of error occurrence, it is assumed that the processing is restarted from the beginning of the transaction that was being processed at the time of error occurrence. There is also a job to skip without re-executing. When such a job exists, it can be dealt with by adding a process that does not add the number of errors to the number of transactions when calculating the error rate of the predicted processing time. Specifically, the job execution status is monitored, and when a job skips a transaction when an error occurs, it can be dealt with by adding a function that does not add the number of errors to the number of transactions.

  In the above-described embodiment, when the predicted error number calculation unit 221 calculates the predicted error number, the number of errors per job (average value of the number of errors) is calculated. However, the present invention is not limited to this. For example, the predicted error number may be calculated by calculating the success rate of transaction processing (number of errors / number of successful transactions) for each job.

  In the above-described embodiment, when the job provisional execution unit 12 provisionally executes a job, the process for one transaction is provisionally executed for each job. You don't have to wait until you have everything. For example, if the amount of data for one transaction is large and the processing time for one data is long, starting the provisional execution process after all the data for one transaction has been prepared will affect the actual processing. Can also be considered. Therefore, in such a case, the provisional execution process may be started when one piece of data handled in the provisional execution process is prepared. In this case, a data monitoring server may be provided for monitoring whether or not one piece of data handled in the provisional execution process has been prepared. As a result, provisional execution can be performed without affecting the actual processing.

  A part or all of the above embodiments can be described as in the following supplementary notes, but the present invention is not limited to the following.

  (Appendix 1) A processing time prediction device that predicts the processing time required for batch processing including a plurality of jobs, and calculates the number of transactions for each job using the number of processing data for each job and the number of processing data for each transaction. A transaction number calculation unit, a provisional processing time measurement unit that temporarily executes processing for one transaction for each job, and measures a provisional processing time that is a processing time for one transaction for each job, the number of transactions and the provisional time Using the processing time, a predicted processing time calculation unit that calculates a predicted processing time that is a processing time assumed for each job, and an actual processing time measurement unit that measures a real processing time that is required when each job is executed A counting unit that counts the number of errors that occur when each job is executed, and the transaction An error rate calculation unit that calculates an error rate of the predicted processing time using the actual processing time and the number of errors; and the predicted processing time of an unprocessed job using the error rate and the temporary processing time. A processing time prediction apparatus comprising: a prediction processing time recalculation unit that calculates again.

  (Supplementary Note 2) A predicted end time that is a time at which each job is scheduled to end is calculated using the predicted processing time, and whether or not the predicted end time exceeds an end deadline that can be set in advance for each job. The processing time prediction apparatus according to claim 1, further comprising: a determination unit that determines whether or not and a warning output unit that outputs a warning when any of the predicted end times exceeds the end deadline.

  (Supplementary Note 3) A job exclusion unit that excludes a job corresponding to an optional job that can be set in advance for each job from the target of batch processing when any of the predicted end times exceeds the end time limit, The processing time prediction apparatus according to supplementary note 2, characterized by comprising:

  (Supplementary Note 4) A processing time prediction method for predicting the processing time required for batch processing including a plurality of jobs, and calculating the number of transactions for each job using the number of processing data for each job and the number of processing data for each transaction. A transaction number calculating step, a provisional processing time measuring step of temporarily executing processing for one transaction for each job, and measuring a provisional processing time that is a processing time for one transaction for each job, the number of transactions and the provisional time A predicted processing time calculation step for calculating a predicted processing time that is a processing time assumed for each job by using the processing time, and an actual processing time measurement step for measuring a real processing time that is a processing time required for executing each job. And a counts that counts the number of errors that occur when each job is executed An error rate calculating step of calculating an error rate of the predicted processing time using the number of transactions, the actual processing time, and the number of errors, and using the error rate and the provisional processing time, A prediction processing time recalculation step for recalculating the prediction processing time of the job.

  (Additional remark 5) The processing time prediction program for making a computer perform each step of Additional remark 4.

  DESCRIPTION OF SYMBOLS 1 ... Processing time prediction apparatus, 11 ... Transaction number calculation part, 12 ... Temporary job execution part, 13 ... Prediction processing time calculation part, 21 ... Job execution part, 22 ... Execution status reflection part, 23 ... Prediction processing time recalculation part , 31 ... Processing order adjustment unit, 41 ... Job information table, 42 ... Processing related information table, 43 ... Information table for prediction, 44 ... Execution status information table, 121 ... Temporary processing time measurement unit, 211 ... Actual processing time measurement unit , 212... Count unit, 221... Prediction error number calculation unit, 222... Error rate calculation unit, 311.

Claims (5)

A processing time prediction device for predicting processing time required for batch processing including a plurality of jobs,
A transaction number calculation unit for calculating the number of transactions for each job using the number of processing data for each job and the number of processing data for each transaction;
A provisional processing time measurement unit that provisionally executes processing for one transaction for each job and measures a provisional processing time that is processing time for one transaction for each job;
A predicted processing time calculation unit that calculates a predicted processing time that is a processing time assumed for each job, using the number of transactions and the provisional processing time;
An actual processing time measurement unit that measures actual processing time that is processing time required for execution of each job;
A counting unit that counts the number of errors that occur during the execution of each job;
Using the number of transactions, the actual processing time, and the number of errors, an error rate calculating unit that calculates an error rate of the predicted processing time;
A predicted processing time recalculation unit that recalculates the predicted processing time of an unprocessed job using the error rate and the provisional processing time;
A processing time prediction apparatus comprising: Determination that calculates a predicted end time that is a time at which each job is scheduled to end using the predicted processing time, and determines whether the predicted end time exceeds an end time limit that can be set in advance for each job. And
A warning output unit for outputting a warning when any of the predicted end times exceeds the end deadline;
The processing time prediction apparatus according to claim 1, further comprising: A job exclusion unit that excludes a job corresponding to an optional job that can be set in advance for each job from the batch processing target when any of the predicted end times exceeds the end deadline;
The processing time prediction apparatus according to claim 2, further comprising: A processing time prediction method for predicting processing time required for batch processing including a plurality of jobs,
A transaction number calculating step for calculating the number of transactions for each job using the number of processing data for each job and the number of processing data for each transaction;
Temporary processing time measurement step for temporarily executing processing for one transaction for each job and measuring temporary processing time that is processing time for one transaction for each job;
A predicted processing time calculating step of calculating a predicted processing time, which is a processing time assumed for each job, using the number of transactions and the provisional processing time;
An actual processing time measuring step for measuring an actual processing time which is a processing time required for executing each job;
A counting step that counts the number of errors that occur during the execution of each job;
An error rate calculating step of calculating an error rate of the predicted processing time using the number of transactions, the actual processing time, and the number of errors;
A predicted processing time recalculation step of recalculating the predicted processing time of an unprocessed job using the error rate and the provisional processing time;
The processing time prediction method characterized by including.   A processing time prediction program for causing a computer to execute each step according to claim 4.

Images