Operation management device and method

Abstract

To provide an apparatus and method of transferring data allowed to transfer data at low cost while suppressing the shutdown of a system due to transfer of a job-net between cloud services.SOLUTION: A method of transferring data includes: acquiring an execution schedule of a job-net and execution time of each job respectively; predicting volume copy time from first cloud service to second cloud service respectively based on a volume for each job to use respectively; calculating start time of day of each job based on the execution schedule of the job-net and the execution time of each job; calculating start time of day of the volume copy of each volume based on the calculated start time of day of each job and the volume copy time of each volume; and starting the volume copy of each volume from the first cloud service to the second cloud service at the start time of day of the volume copy respectively calculated on the volume.SELECTED DRAWING: Figure 24

Description

Technical Field

[0001] The present invention relates to an operation management apparatus and method, and is suitable for application to, for example, an operation management apparatus that manages the transfer of data used in jobs between clouds.

Background Art

[0002] With the spread of cloud services, the number of cases where multiple cloud services are used for each application has been increasing. For example, when cost is emphasized, inexpensive cloud services are used for systems with low importance, and stable but expensive cloud services are used for systems with high importance. There are also usage patterns where emphasis is placed on functionality and a cloud service having a desired function is selected.

[0003] Regarding batch jobs such as aggregation and analysis, they have mostly been performed in an on-premises environment in the past, but in recent years, the number of cases where they are executed on the cloud has also been increasing.

[0004] In view of such a situation, it is considered that there is a need for a technology to transfer an existing batch job executed on a cloud service or on-premises to an appropriate cloud service.

[0005] In this case, a group of containers that execute a batch job on the cloud can be transferred across a plurality of cloud services in, for example, a few minutes or seconds on a Kubernetes cluster.

[0006] On the other hand, regarding the data used by a batch job, since the amount of data is large, it takes an enormous amount of time to transfer compared to containers. However, even if the group of containers that execute the batch job is transferred, the system cannot operate if the data used by the batch job does not exist at the transfer destination.

[0007] Furthermore, in relation to the migration of systems between cloud services, Patent Document 1 discloses a method for migrating virtual machines between multiple cloud services. [Prior art documents] [Patent Documents]

[0008] [Patent Document 1] U.S. Patent Publication No. 2019 / 0171477 [Overview of the project] [Problems that the invention aims to solve]

[0009] When applying the invention disclosed in Patent Document 1 to a real-world system, considering data migration between systems, it is likely that the data will be copied asynchronously to the destination cloud service. In this case, Patent Document 1 does not take into account the time required for data migration, and therefore there is a problem that the system will stop because batch jobs cannot be executed during data migration.

[0010] Another option is to sequentially copy the data from the cloud service to be migrated to the destination cloud service using synchronous copying. However, this method has drawbacks, such as extending the period during which cloud storage usage fees are incurred at both the source and destination cloud services, or incurring costs for expanding network bandwidth to facilitate data migration.

[0011] This invention has been made in consideration of the above points, and aims to propose a data migration device and method that can perform data migration at low cost while suppressing system downtime associated with the migration of job nets between cloud services. [Means for solving the problem]

[0012] To solve the above problems, the present invention provides a data migration device that migrates data from each volume used by each job constituting a job net executed in a first cloud service, the source of migration, to a second cloud service, the device comprising: a schedule acquisition unit that acquires the execution schedule of the job net; a job execution time prediction unit that predicts the execution time of each of the jobs; a volume copy time prediction unit that predicts the volume copy time for each of the volumes used by each of the jobs, which is the time required for volume copying from the first cloud service to the second cloud service; a volume copy start time calculation unit that calculates the start time of each job based on the execution schedule of the job net and the execution time of each of the jobs, and calculates the start time of volume copying for each of the volumes based on the calculated start time of each job and the volume copy time of each of the volumes, so that volume copying is completed by the start time of the corresponding job; and a volume copy execution unit that starts volume copying of each of the volumes from the first cloud service to the second cloud service at the volume copy start time calculated for each of the volumes.

[0013] Furthermore, the present invention provides a data migration method performed by a data migration device that migrates data from each volume used by each job constituting a job net executed in a first cloud service, the source of migration, to a second cloud service, the method comprising: a first step of obtaining the execution schedule of the job net, predicting the execution time of each job, and further predicting the volume copy time for each volume used by each job, which is the time required for volume copying from the first cloud service to the second cloud service; a second step of calculating the start time of each job based on the execution schedule of the job net and the execution time of each job, and calculating the start time of volume copying for each volume based on the calculated start time of each job and the volume copy time of each volume, so as to ensure that volume copying is completed by the start time of the corresponding job; and a third step of starting volume copying of each volume from the first cloud service to the second cloud service at the volume copy start time calculated for each volume.

[0014] According to the data migration device and method of the present invention, the volumes used by each job constituting the job net migrated from the first cloud service to the second cloud service can be volume-copied to the second cloud service by the start time of each job, thereby preventing system downtime in the second cloud service caused by the absence of data used by the jobs.

[0015] Furthermore, this data migration device and method can reduce the time required to prepare the volumes used by each job constituting the job net to be migrated in the first and second cloud services, thereby reducing the usage costs of the first and second cloud services associated with the migration of the job net. [Effects of the Invention]

[0016] According to the present invention, it is possible to realize a data transfer device and method capable of performing low-cost data transfer while suppressing system stoppage associated with the transfer of a job net between cloud services.

Brief Description of the Drawings

[0017] [Figure 1] It is a block diagram showing the configuration of an information processing system according to the first, second, and fourth embodiments. [Figure 2] It is a block diagram showing the configuration of an operation management server according to the first and second embodiments. [Figure 3] It is a chart showing a configuration example of a business schedule holding table. [Figure 4] It is a chart showing a configuration example of a job volume mapping table. [Figure 5] It is a chart showing a configuration example of a job predicted execution time holding table. [Figure 6] It is a chart showing a configuration example of a volume copy predicted time holding table. [Figure 7] It is a chart showing a configuration example of a volume copy start time holding table. [Figure 8] It is a flowchart showing the processing procedure of business schedule acquisition processing. [Figure 9] It is a flowchart showing the processing procedure of job mapping processing. [Figure 10] It is a flowchart showing the processing procedure of job execution time prediction processing. [Figure 11] It is a flowchart showing the processing procedure of volume copy time prediction processing. [Figure 12] It is a flowchart showing the processing procedure of volume copy start time calculation processing. [Figure 13] It is a flowchart showing the processing procedure of volume copy execution processing. [Figure 14] It is a flowchart showing the processing procedure of job monitoring processing. [Figure 15]This flowchart shows the processing procedure for changing the volume copy start time. [Figure 16] This flowchart shows the processing procedure for the volume copy execution process according to the second embodiment. [Figure 17] This is a block diagram showing the configuration of an information processing system according to the third embodiment. [Figure 18] This block diagram shows the configuration of the operation management server according to the third embodiment. [Figure 19] This diagram shows an example of the configuration of a transaction count retention table. [Figure 20] This flowchart shows the processing procedure for job execution time prediction according to the third embodiment. [Figure 21] This block diagram shows the configuration of the operation management server according to the fourth embodiment. [Figure 22] This diagram shows an example of the configuration of a data transfer volume retention table for each time period. [Figure 23] This flowchart shows the processing steps for creating a table that holds the data transfer volume for each time period. [Figure 24] This flowchart shows the processing procedure for calculating the volume copy start time according to the fourth embodiment. [Modes for carrying out the invention]

[0018] An embodiment of the present invention will be described in detail below with reference to the drawings.

[0019] (1) First embodiment (1-1) Configuration of the information processing system according to this embodiment In Figure 1, 1 represents the overall configuration of the information processing system according to this embodiment. This information processing system 1 comprises a cloud service 2 that serves as the source of batch jobs (hereinafter referred to as the source cloud), a cloud service 3 that serves as the destination for the batch jobs (hereinafter referred to as the destination cloud), and an operation management server 4 that manages the operation of the source cloud 2 and the destination cloud 3. The source cloud 2, the destination cloud 3, and the operation management server 4 are interconnected via a network 5 such as the Internet.

[0020] Source Cloud 2 is a cloud service provided by a cloud system comprising one or more storage devices 10 that provide volume VOL as storage area to external business servers, etc. (not shown), and one or more server devices (not shown) that execute predetermined job nets, which are batch jobs such as aggregation of data stored in volume VOL. In Source Cloud 2, job net management files 11 for job nets that are periodically executed by such server devices are managed.

[0021] The jobnet management file 11 is a file that stores various information about the corresponding jobnet. For example, it stores information such as the start time of the corresponding jobnet, the job configuration of that jobnet (the content and order of each job 12 that makes up that jobnet), and which job 12 is associated with which Pod (not shown) in the Kubernetes environment, and which Pod is using which volume VOL.

[0022] Furthermore, the migration destination cloud 3 is a cloud service provided by a cloud system that includes one or more storage devices 10 that provide volume VOL to external business servers, etc. (not shown), and one or more server devices (not shown) that execute a predetermined job net, which is a batch job, on the data stored in volume VOL.

[0023] In the destination cloud 3, the system migration will be performed on a cluster basis from the source cloud 2. Specifically, the jobnet management files 11 for one or more jobnets that make up the system, and the data (volume data) stored in the volume VOL used by each job 12 that makes up the jobnet to be migrated will be migrated from the source cloud 2. Then, based on the migrated jobnet management files 11 and volume data, a jobnet with the same content as the jobnet that was running in the source cloud 2 will be executed in the destination cloud 3.

[0024] The operation management server 4 is a server device that manages the operations of the source cloud 2 and the destination cloud 3, and as shown in Figure 2, it is configured to include a CPU (Central Processing Unit) 20, memory 21, storage device 22, input device 23, output device 24, and communication device 25.

[0025] The CPU 20 is the processor that controls the operation of the entire operation management server 4. The memory 21 is composed of, for example, semiconductor memory and is used as the working memory for the CPU 20. The storage device 22 is composed of a large-capacity non-volatile storage device such as a hard disk drive or an SSD (Solid State Drive) and is used to store various programs and data that needs to be stored for a long period of time.

[0026] When the operation management server 4 starts up or as needed, the CPU 20 reads the necessary programs from the storage device 22 into memory 21. The CPU 20 then executes the programs read into memory 21, thereby performing various processes for the entire operation management server 4 as described later.

[0027] The input device 23 consists of a keyboard, mouse, etc., and is used by the operator of the operation management server 4 to input necessary information and instructions. The output device 24 consists of a liquid crystal display, an organic EL (Electro-Luminescence) display, etc., and is used to display necessary information. Alternatively, a touch panel integrating the input device 22 and output device 24 may be used instead.

[0028] The communication device 25 consists of, for example, a NIC (Network Interface Card) and performs protocol control during communication with the source cloud 2 and destination cloud 3 via the network 5 (Figure 1).

[0029] (1-2) Data migration function Next, the data migration function installed on the operation management server 4 will be described. This data migration function migrates the data of the volume VOL used by each job 12 that constitutes one or more job nets specified by the user, etc., among the job nets running in the source cloud 2, to the destination cloud 3 by the start time of the corresponding job 12. In the following, it will be assumed that the job nets to be migrated themselves will be migrated separately to the destination cloud 3 by their start time.

[0030] As a means to realize such data migration functionality, the storage device 22 of the operation management server 4 stores, as shown in Figure 2, a database 35 containing a business schedule storage table 30, a job volume mapping table 31, a job predicted execution time storage table 32, a volume copy predicted time storage table 33, and a volume copy start time storage table 34, as well as a business schedule acquisition unit 36, a job volume mapping unit 37, a job execution time prediction unit 38, a volume copy time prediction unit 39, a volume copy start time calculation unit 40, a volume copy execution unit 41, a job monitoring unit 42, and a volume copy start time modification unit 43.

[0031] The business schedule retention table 30 is a table used to manage the job configuration and start time of the job nets to be migrated, which are recognized based on the job net management file 11 of each job net held by the source cloud 2. As shown in Figure 3, it is configured to include a job net column 30A, a job column 30B, and a start time column 30C. In the business schedule retention table 30, one record (row) corresponds to one job net being executed in the source cloud 2.

[0032] The jobnet column 30A stores the identification information of the corresponding jobnet to be migrated, which is executed in the source cloud 2. The job column 30B stores the identification information of each job 12 (Figure 1) that makes up the corresponding jobnet, arranged in the order in which the jobs 12 are executed. The start time column 30C stores the start time of the corresponding jobnet that is set for that jobnet.

[0033] Therefore, in the example shown in Figure 3, for example, the job net named "CronJob1" is shown to be a job net that executes three jobs 12, "job1", "job2", and "job3", in that order at "15:00 every day".

[0034] Furthermore, the job volume mapping table 31 is a table used to manage the correspondence between the jobs 12 that constitute the job net to be migrated and executed in the source cloud 2, and the volume VOL (Figure 1) used when executing those jobs 12.

[0035] As shown in Figure 4, the job volume mapping table 31 is configured to include a job column 31A and a volume column 31B. In the job volume mapping table 31, each record (row) corresponds to the relationship between a job 12 that constitutes one of the job nets executed in the source cloud 2 and a volume VOL.

[0036] The job column 31A stores the identification information of the corresponding job 12, and the volume column 31B stores the identification information of the volume VOL used by that job 12. Therefore, in the example in Figure 4, "job1" is assigned the volume VOL "vol1" (indicating that the volume VOL "vol1" is used when executing the job 12 "job1").

[0037] Note that while Figure 4 illustrates a case where each job 12 has a one-to-one relationship with volume VOL, there are also cases where the relationships between job 12 and volume VOL are one-to-many, many-to-one, or many-to-many.

[0038] The Job Prediction Execution Time Storage Table 32 is a table used to manage the time required to execute each job 12 that constitutes each job net to be migrated, which is executed on the source cloud 2. As shown in Figure 5, it is configured to include a job column 32A and an execution time column 32B. In the Job Prediction Execution Time Storage Table 32, one record (row) corresponds to one job 12 that constitutes a job net to be migrated, which is executed on the source cloud 2.

[0039] The job column 32A stores the identification information of the corresponding job 12, and the execution time column 32B stores the execution time required to execute that job 12. This execution time may be the average of the execution times of that job 12 performed so far, or it may be an estimated time. For example, in the example in Figure 5, the execution time of job 12, "job1", is shown to be "90 min".

[0040] The volume copy prediction time retention table 33 is a table used to manage the volume size of the volume VOL used by each job 12 that constitutes the job net to be migrated, which is executed on the source cloud 2, and the estimated time required to copy that volume VOL to the destination cloud 3 (hereinafter, this will be referred to as the volume copy prediction time).

[0041] As shown in Figure 6, this volume copy prediction time retention table 33 is configured to include a volume column 33A, a volume size column 33B, and a copy prediction time column 33C. In the volume copy prediction time retention table 33, one record (row) corresponds to one volume VOL used by one job 12 that constitutes the job net to be migrated.

[0042] The volume field 33A stores the identification information of the corresponding volume VOL, and the volume size field 33B stores the volume size (capacity) of that volume VOL. The copy prediction time field 33C stores the estimated volume copy time required to copy that volume VOL from source cloud 2 to destination cloud 3. The method for calculating this volume copy prediction time will be described later.

[0043] Therefore, in the example shown in Figure 6, the volume VOL named "vol1" in source cloud 2 has a volume size of "30GB", and it is shown that copying the volume from source cloud 2 to destination cloud 3 takes "30 minutes".

[0044] The volume copy start time retention table 34 is a table used to manage the time at which volume copying should begin from source cloud 2 to destination cloud 3 for each volume VOL, as calculated by the volume copy start time calculation unit 40 (Figure 2) as described later. As shown in Figure 7, it is configured to include a volume column 34A and a volume copy start time column 34B.

[0045] The volume column 34A stores identification information for the volume VOL used by any of the jobs 12 that make up the job net to be migrated, and the volume copy start time column 34B stores the time calculated by the volume copy start time calculation unit 40 as the time when the volume copy of the corresponding volume VOL should start.

[0046] Therefore, in the example shown in Figure 7, the volume copy start time calculation unit 40 calculated that the volume copy of volume "vol1" from source cloud 2 to destination cloud 3 should start at "14:30".

[0047] On the other hand, the business schedule acquisition unit 36 ​​(Figure 2) is a program that has the function of acquiring the job configuration and start time of each job net from the job net management file 11 of each job net to be migrated, which is held by the source cloud 2. The business schedule acquisition unit 36 ​​stores this acquired information in the business schedule holding table 30 (Figure 3).

[0048] Furthermore, the job volume mapping unit 37 (Figure 2) is a program that has the function of identifying the volume VOL used by each job 12 that constitutes each job net, from the job net management file 11 of each job net to be migrated. The job volume mapping unit 37 stores the correspondence between each identified job 12 and the volume VOL used by that job 12 in the job volume mapping table 31 (Figure 4).

[0049] The job execution time prediction unit 38 (Figure 2) is a program that has the function of predicting the next execution time of each job 12 that constitutes the job net to be migrated, based on logs of past executions of each job 12 managed on the source cloud 2. The job execution time prediction unit 38 stores the predicted execution time of each job 12 (hereinafter referred to as the predicted execution time) in the job predicted execution time retention table 32 (Figure 5).

[0050] Furthermore, the volume copy time prediction unit 39 (Figure 2) is a program that has the function of predicting the volume copy time required to copy each volume VOL used by each job 12 to be migrated from the source cloud 2 to the destination cloud 3. The volume copy time prediction unit 39 stores the predicted volume copy time (volume copy prediction time) for each volume VOL in the volume copy prediction time retention table 33 (Figure 6).

[0051] The volume copy start time calculation unit 40 (Figure 2) is a program that has the function of calculating the start time (hereinafter referred to as the volume copy start time) for volume copying of each volume VOL used by each job 12 that constitutes the job net to be migrated to the destination cloud 3.

[0052] In practice, the volume copy start time calculation unit 40 calculates the volume copy start time for the volumes used by the job 12 to be migrated to the destination cloud 3 so that the volume copy of the volumes used by the job 12 to the destination cloud 3 is completed before the job 12 is executed in the destination cloud 3. The volume copy start time calculation unit 40 then stores the calculated volume copy start time for each volume in the volume copy start time retention table 34 (Figure 7).

[0053] The volume copy execution unit 41 is a program that has the function of executing a volume copy of the volume VOL corresponding to the volume copy start time of each volume VOL stored in the volume copy start time retention table 34 to the destination cloud 3.

[0054] Furthermore, the job monitoring unit 42 is a program that, after the migration of the job net from the source cloud 2 to the destination cloud 3, monitors the execution status of each job 12 that constitutes the job net being executed in the destination cloud 3, and has the function of predicting the actual execution time of these jobs 12 in the destination cloud 3.

[0055] Furthermore, the volume copy start time modification unit 43 is a program that has the function of changing the volume copy start time to the destination cloud 3 of the volume VOL used by the job executed after the job 12 (hereinafter referred to as the subsequent job) 12, if the actual execution time of the job 12 predicted by the job monitoring unit 42 deviates from the predicted execution time of the job 12 calculated by the job execution time prediction unit 38.

[0056] For example, if the actual execution time of job 12 predicted by the job monitoring unit 42 is shorter than the predicted execution time of job 12 calculated by the job execution time prediction unit 38, the volume copy start time of the volume VOL stored in the volume copy start time retention table 34 will be changed accordingly to advance the start time of the volume copy of the volume VOL used by subsequent jobs 12 to the destination cloud 3.

[0057] (1-3) Specific processing details of each program Next, we will explain the specific processing details of the various processes performed by the business schedule acquisition unit 36, job volume mapping unit 37, job execution time prediction unit 38, volume copy time prediction unit 39, volume copy start time calculation unit 40, volume copy execution unit 41, job monitoring unit 42, and volume copy start time modification unit 43 in relation to the data migration function.

[0058] (1-3-1) Business schedule acquisition process Figure 8 shows the flow of a series of processes (hereinafter referred to as the business schedule acquisition process) executed by the business schedule acquisition unit 36 ​​in relation to the data migration function. When the input device 23 (Figure 2) of the operation management server 4 is operated by a user and one or more job nets to be migrated are specified, and input is given indicating that the migration of those job nets should be executed, the business schedule acquisition unit 36 ​​starts the business schedule acquisition process shown in Figure 8.

[0059] The business schedule acquisition unit 36 ​​first obtains the start time of the job net from the job net management file 11 of the job net that has been instructed to migrate, which is managed in the source cloud 2 (S1). The business schedule acquisition unit 36 ​​also obtains the identification information of each job 12 that makes up the job net from the job net management file 11 of the job net to be migrated (S2).

[0060] The business schedule acquisition unit 36 ​​then stores the start time of each job net to be migrated, acquired in step S1, and the identification information of each individual job 12 that constitutes each job net to be migrated, acquired in step S2, into the business schedule holding table 30 (Figure 3) (S3), and then terminates this business schedule acquisition process.

[0061] (1-3-2) Job volume mapping process Figure 9 shows the flow of a series of processes (hereinafter referred to as the job mapping process) performed by the job volume mapping unit 37 (Figure 2) in relation to the data migration function. The job volume mapping unit 37 identifies the correspondence between each job 12 in the job net to be migrated and the volume VOL used by each of these jobs 12, following the processing procedure shown in Figure 9.

[0062] In practice, the job volume mapping unit 37 starts the job volume mapping process shown in Figure 9 once the business schedule acquisition process described above by the business schedule acquisition unit 36 ​​is completed.

[0063] The job volume mapping unit 37 then first refers to the job net management file 11 of each job net to be migrated, which was obtained by the business schedule acquisition unit 36, and for each of these job nets, it obtains the correspondence between each job 12 that constitutes that job net and the Pod that executes that job 12 (S10).

[0064] The job volume mapping unit 37 also refers to the job net management file 11 for each job net and obtains the correspondence relationship with volume VOL for each Pod whose relationship with job 12 was obtained in step S10 (S11).

[0065] Next, the job volume mapping unit 37 obtains the correspondence between each job 12 and the volume VOL used by that job 12 based on the information obtained in steps S10 and S11, and stores the obtained correspondence between each job 12 and volume VOL in the job volume mapping table 31 (Figure 4) (S12). After this, the job volume mapping unit 37 terminates the job mapping process.

[0066] For example, if the information processing system 1 (Figure 1) is configured to manage the information of the job net management file 11 for each job net executed on the source cloud 2 using a separately provided job operation management server (not shown), the job volume mapping unit 37 may use an API (Application Programming Interface) or the like to obtain information such as the job configuration, start time, and correspondence between each job 12 and volume VOL for each job net from the job operation management server.

[0067] (1-3-3) Job execution time prediction process Figure 10 shows the flow of a series of processes (hereinafter referred to as the job execution time prediction process) performed by the job execution time prediction unit 38 (Figure 2) in relation to the data migration function. The job execution time prediction unit 38 predicts the execution time for each job 12 of the job net to be migrated, following the processing procedure shown in Figure 10.

[0068] In practice, the job execution time prediction unit 38 starts this job execution time prediction process in parallel with the business schedule acquisition process and job volume mapping process performed by the business schedule acquisition unit 36 ​​and the job volume mapping unit 37.

[0069] The job execution time prediction unit 38 then first looks at the logs of past executions of each job 12 that make up each job net to be migrated, which are executed on the source cloud 2, and obtains the execution time (the time required for that job 12) when that job 12 was executed in the past (S20).

[0070] Next, the job execution time prediction unit 38 calculates for each of the jobs 12 the average value of the execution time obtained in step S20 for a specified number of times when that job 12 was executed in the past (S21).

[0071] Furthermore, the job execution time prediction unit 38 stores the average execution time for each job 12 calculated in step S21 as the predicted execution time (predicted execution time) for the corresponding job 12 in the job predicted execution time storage table 32 (Figure 5) (S22).

[0072] The job execution time prediction unit 38 then terminates the job execution time prediction process.

[0073] (1-3-4) Volume copy time prediction process Figure 11 shows the flow of a series of processes (hereinafter referred to as the volume copy time prediction process) performed by the volume copy time prediction unit 39 (Figure 2) in relation to the data migration function. The volume copy time prediction unit 39 calculates the volume copy prediction time for each volume VOL used by each job 12 that constitutes the job net to be migrated, following the processing procedure shown in Figure 11.

[0074] In practice, the volume copy time prediction unit 39 starts the volume copy time prediction process shown in Figure 11 after the job volume mapping process by the job volume mapping unit 37 (Figure 9) and the job execution time prediction process by the job execution time prediction unit 38 (Figure 10) have been completed.

[0075] The volume copy time prediction unit 39 first refers to logs held by the source cloud 2 and destination cloud 3 of past data transfers between the source cloud 2 and destination cloud 3 to obtain the amount of data transferred and the transfer time at that time (S30).

[0076] Next, the volume copy time prediction unit 39 calculates the following based on the amount of data to be transferred and the transfer time acquired in step S30:

number

[0077] Next, the volume copy time prediction unit 39 calculates the volume VOL used by each job 12 that constitutes each job net to be migrated, using the following formula

number

[0078] (1-3-5) Volume copy start time calculation process On the other hand, Figure 12 shows the flow of a series of processes (hereinafter referred to as the volume copy start time calculation process) performed by the volume copy start time calculation unit 40 (Figure 2) in relation to the data migration function. The volume copy start time calculation unit 40 calculates the start time (volume copy start time) for volume copying from source cloud 2 to destination cloud 3 for each volume VOL used by each job in the job net to be migrated, according to the processing procedure shown in Figure 12.

[0079] In practice, the volume copy start time calculation unit 40 starts the volume copy start time calculation process shown in Figure 12 once the volume copy time prediction process (Figure 11) by the volume copy time prediction unit 39 is completed.

[0080] The volume copy start time calculation unit 40 first obtains the start time and identification information of each job 12 constituting the job net from the business schedule holding table 30 (Figure 3) for each job net to be migrated (S40).

[0081] Next, the volume copy start time calculation unit 40 obtains the predicted execution time for each of the jobs 12 for which identification information was obtained in step S40 from the job predicted execution time retention table 32 (Figure 5) (S41).

[0082] Next, the volume copy start time calculation unit 40 searches the job volume mapping table 31 (Figure 4) for the volumes VOL used by all jobs 12 for which identification information was obtained in step S40, and obtains the volume copy prediction time to the destination cloud 3 for each of the detected volumes VOL from the volume copy prediction time retention table 33 (Figure 6) (S42).

[0083] Furthermore, the volume copy start time calculation unit 40 calculates the start time for each of these jobs 12 in the destination cloud 3 based on the start time of each job net obtained in step S40 and the predicted execution time of each job 12 obtained in step S41 (S43).

[0084] For example, if the start time of a job net called "CornJob1" is "15:00 every day" as shown in Figure 3, and the predicted execution times of "job1", "job2", and "job3" that make up this job net are "90 min", "30 min", and "60 min" respectively as shown in Figure 5, then the volume copy start time calculation unit 40 sets the start time of "job1" to "15:00", which is the start time of the job net.

[0085] Furthermore, the volume copy start time calculation unit 40 calculates that for "job2", the start time is "16:30", which is 90 minutes after the start of the job net and is the execution time of "job1". In addition, the volume copy start time calculation unit 40 calculates that for "job3", the start time is "17:00", which is 30 minutes after the start of "job2" and is the execution time of "job2".

[0086] Next, the volume copy start time calculation unit 40 calculates the volume copy start time for each volume VOL, based on the start time of each job 12 calculated in step S43 and the volume copy prediction time for each volume VOL used by each job 12, obtained in step S42, so that the volume copy of the volume VOL used by each job 12 is completed by the start time of each job 12 (S44).

[0087] Specifically, the volume copy start time calculation unit 40 calculates the volume copy start time for volume VOL, which is obtained by subtracting the predicted volume copy time for volume VOL used by job 12 from the start time of job 12.

[0088] For example, if the volume VOL used by "job2," which constitutes the job net "CornJob1," is "vol2" as shown in Figure 4, and the estimated volume copy time for "vol2" is "20 min" as shown in Figure 6, then the volume copy start time for "vol2" is calculated as "16:10" by subtracting the estimated volume copy time for "vol2," which is "20 min," from the start time for "vol2," which is "16:30," as calculated above.

[0089] Next, the volume copy start time calculation unit 40 stores the volume copy start times for each volume VOL calculated as described above in the volume copy start time holding table 34 (Figure 7) (S45), and then terminates the volume copy start time calculation process.

[0090] (1-3-6) Volume copy execution process Figure 13 shows the flow of a series of processes (hereinafter referred to as the volume copy execution process) performed by the volume copy execution unit 41 in relation to the data migration function. The volume copy execution unit 41 copies each volume VOL used by each job 12 of the job net to be migrated to the destination cloud 3 according to the processing procedure shown in Figure 13.

[0091] In practice, the volume copy execution unit 41 starts the volume copy execution process shown in Figure 13 once the volume copy start time calculation process (Figure 12) by the volume copy start time calculation unit 40 is completed.

[0092] The volume copy execution unit 41 then waits for the current time to become the volume copy start time of any of the volumes, based on a clock (not shown) and the volume copy start times of each volume VOL stored in the volume copy start time column 34B (Figure 7) of each record in the volume copy start time holding table 34 (S50).

[0093] Then, when the volume copy execution unit 41 obtains a positive result in step S50 because the current time eventually becomes the volume copy start time for one of the volumes VOL, it starts copying the volume VOL from source cloud 2 to destination cloud 3 for the volume VOL whose volume copy start time has arrived (S51).

[0094] Specifically, the volume copy execution unit 41 controls the source cloud 2 and destination cloud 3 to migrate the data of the volume VOL from the source cloud 2 to the destination cloud 3.

[0095] Furthermore, once the volume copy execution unit 41 has finished copying the target volume VOL in step S51, it deletes that volume VOL in the source cloud 2 (S52). The volume copy execution unit 41 then returns to step S50, and thereafter repeats the process from step S50 onward in the same manner as described above.

[0096] Furthermore, the volume copy start time calculation unit 40 may, after the completion of the volume copy start time calculation process in Figure 12, create a job net for each volume VOL, treating the series of processes in steps S51 and S52 in Figure 13 as a single job net, and register the volume copy start time of the corresponding volume VOL calculated in the volume copy start time calculation process as the start time of that job net in the source cloud 2 and destination cloud 3. By doing so, the volume copy for each volume VOL in Figure 13 can be automated.

[0097] (1-3-7) Job monitoring process Figure 14 shows the flow of a series of processes (hereinafter referred to as the job monitoring process) performed by the job monitoring unit 42 (Figure 2) in relation to the data migration function. The job monitoring unit 42 starts this job monitoring process after a certain period of time has elapsed since any of the jobs 12 constituting the job net that have been migrated to the destination cloud 3 were started in the destination cloud 3.

[0098] The job monitoring unit 42 first determines whether there is a job (successor job) 12 that follows the currently running job (hereinafter referred to as the running job) 12 (S60). If the job monitoring unit 42 obtains a negative result in this determination, it terminates the job monitoring process.

[0099] In response, if the job monitoring unit 42 obtains a negative result in step S60, it obtains the number of data items that the running job 12 should process based on the data stored in volume VOL that the running job 12 is currently using (S61).

[0100] The job monitoring unit 42 also obtains the number of data items that the running job 12 has processed up to the present based on the logs of the processes that the running job 12 is performing (S62).

[0101] Next, the job monitoring unit 42 predicts the actual execution time of the running job 12 (S63) based on the number of data that the running job 12 should process, obtained in step S61, the number of data that the running job 12 has processed up to the present, obtained in step S62, and the elapsed time from when the running job 12 started until the present.

[0102] Specifically, the job monitoring unit 42 uses the following formula

number

[0103] Next, the job monitoring unit 42 performs the following:

number

[0104] In response, if the job monitoring unit 42 obtains a positive result in the judgment in step S64, it gives an instruction to the volume copy start time change unit 43 to change the volume copy start time of volume VOL used by the subsequent job 12 of the currently executing job 12 (hereinafter referred to as the volume copy start time change instruction) (S65), and then terminates this job monitoring process.

[0105] (1-3-8) Volume copy start time change process Figure 15 shows the flow of a series of processes (hereinafter referred to as the volume copy start time change process) executed by the volume copy start time change unit 43 (Figure 2) upon receiving such a volume copy start time change instruction.

[0106] When the volume copy start time change unit 43 receives such a volume copy start time change instruction from the job monitoring unit 42, it executes a volume copy start time change process in accordance with the processing procedure shown in Figure 15, which changes (advances or delays) the volume copy start time of subsequent jobs 12 stored in the volume copy start time retention table 34 (Figure 7) according to the expected end time of the currently running job 12.

[0107] In practice, when the volume copy start time change unit 43 receives the volume copy start time change instruction from the job monitoring unit 42, it starts the volume copy start time change process shown in Figure 15. First, it obtains from the job monitoring unit 42 the predicted actual execution time of the running job 12, which was calculated in step S63 of the job monitoring process (S70).

[0108] Next, the volume copy start time modification unit 43 searches the job volume mapping table 31 for the volume VOL used by the running job 12, and obtains the predicted volume copy time (volume copy prediction time) for the detected volume VOL from the volume copy prediction time retention table 33 (Figure 6) (S71).

[0109] Next, the volume copy start time modification unit 43 calculates the predicted start time of the subsequent job 12 by adding the predicted actual execution time of the running job 12 obtained in step S70 to the start time of the running job 12 (S72).

[0110] Furthermore, the volume copy start time modification unit 43 obtains the estimated time required to copy the volume VOL used by the subsequent job 12 from the source cloud 2 to the destination cloud 3 (volume copy estimated time) from the volume copy estimated time retention table 33, and subtracts the obtained volume copy estimated time from the start time of the subsequent job 12 calculated in step S72 to calculate the volume copy start time after the volume VOL has been modified (S73).

[0111] The revised volume copy start time calculated at this point is the final time required to copy the volume VOL used by the subsequent job 12 to the destination cloud 3 before the subsequent job 12 starts. Therefore, to allow for some leeway, the volume copy start time for the volume VOL used by the subsequent job 12 may be set earlier by the necessary amount of time compared to the volume copy start time calculated in step S73.

[0112] Next, the volume copy start time modification unit 43 modifies the volume copy start time of the volume (the volume used by the subsequent job 12) VOL stored in the volume copy start time retention table 34 (Figure 7) by overwriting it with the volume copy start time calculated in step S73 (S74). After this, the volume copy start time modification unit 43 terminates this volume copy start time modification process.

[0113] (1-4) Effects of this embodiment As described above, the operation management server 4 of this embodiment can copy the volume VOL used by each job 12 that constitutes the job net migrated from the source cloud 2 to the destination cloud 3 to the destination cloud 3 by the start time of each job 12. Therefore, it is possible to prevent system downtime in the destination cloud 3 caused by the absence of the volume VOL used by the jobs 12 that constitute the migrated job net.

[0114] Furthermore, this operation management server 4 reduces the time required to prepare the volume VOL used by each job 12 constituting the job net to be migrated in both the source cloud 2 and the destination cloud 3, thereby reducing the usage costs of both the source cloud 2 and the destination cloud 3 associated with the migration of the job net.

[0115] Therefore, this operation management server 4 enables low-cost data migration while suppressing system downtime associated with the migration of job nets between cloud services.

[0116] (2) Second embodiment In Figure 2, 50 represents the operation management server 50 according to the second embodiment, which is applied to the information processing system 1 in Figure 1, replacing the operation management server 4 of the first embodiment. This operation management server 50 differs from the operation management server 4 of the first embodiment in that, in addition to the data migration function described above, it is equipped with a cloud bursting function; otherwise, it is configured the same as the operation management server 4 of the first embodiment.

[0117] Here, cloud bursting is a function that, when the load on one cloud system becomes extremely high, allows processes that should be executed on that cloud system to be migrated and executed on another cloud system. Below, we will explain the case where, when the load on source cloud 2 becomes extremely high, job nets that should be executed on source cloud 2 are migrated and executed on destination cloud 3.

[0118] Furthermore, this job net itself will be migrated to the destination cloud 3 separately by its start time. Below, we will explain a series of processes (hereinafter referred to as the cloud bursting data migration support process) that temporarily migrate the volume VOL used by each job constituting this job net to the destination cloud 3.

[0119] This cloud bursting data migration support process is initiated when the input device 23 (Figure 2) of the operation management server 4 is operated by a user or when a request is made from the source cloud 2 that has become extremely overloaded.

[0120] Then, when the data migration support process for cloud bursting is started, the business schedule acquisition unit 36, the job volume mapping unit 37, the job execution time prediction unit 38, the volume copy time prediction unit 39, and the volume copy start time calculation unit 40 sequentially execute the same processes as those executed by the functional unit in relation to the data migration function described above in the first embodiment.

[0121] Furthermore, once the volume copy start time calculation process described above by the volume copy start time calculation unit 40 (Figure 12) is completed, the volume copy execution unit 51 (Figure 2) starts the volume copy execution process shown in Figure 16.

[0122] The volume copy execution unit 51 then first waits for the current time to become the volume copy start time of any of the volumes, based on a clock (not shown) and the volume copy start times of each volume VOL stored in the volume copy start time column 34B (Figure 7) of each record in the volume copy start time holding table 34 (S80).

[0123] Then, when the volume copy execution unit 51 obtains a positive result in step S80 because the current time eventually becomes the volume copy start time for one of the volumes VOL, it executes a volume copy of the volume VOL whose volume copy start time has arrived from source cloud 2 to destination cloud 3 (S81).

[0124] Specifically, the volume copy execution unit 41 controls the source cloud 2 and destination cloud 3 to migrate the data of the volume VOL from the source cloud 2 to the destination cloud 3.

[0125] Furthermore, once the volume copy execution unit 51 has finished copying the target volume VOL in step S51, it waits for the job 12 that uses that volume VOL to finish in the destination cloud 3 (S82).

[0126] Then, when the job 12 is completed, the volume copy execution unit 51 deletes the volume VOL from the destination cloud 3 (S83), and then terminates the volume execution process.

[0127] In this embodiment, the volume copy execution process is performed for each volume VOL. Specifically, if the volume copy execution unit 51 obtains a positive result in step S80, it starts the next volume execution process and executes the volume execution process that obtained a positive result in step S80 and the newly started volume execution process in parallel.

[0128] With the operation management server 50 of this embodiment having the above configuration, the cloud bursting function, like the data migration function, can be provided at low cost while suppressing system downtime associated with the migration of job nets between cloud services.

[0129] (3) Third Embodiment Figure 17, which uses the same reference numerals as Figure 1 to indicate corresponding parts, shows the configuration of the information processing system 60 according to the third embodiment. This information processing system 60 is configured similarly to the information processing system 1 according to the first embodiment, except that the content of the job execution time prediction process, which predicts the execution time of each job 12 constituting the job net to be migrated in the operation management server 61, differs from the content of the job execution time prediction process according to the first embodiment described above in Figure 10.

[0130] In the following, the information processing system 60 will have a business server 62 that performs business transactions and stores the transaction data obtained in this way in volume VOL within source cloud 2. The job net that executes batch jobs such as aggregation on the business transaction data stored in volume VOL will be the job net to be migrated from source cloud 2 to destination cloud 3. In the following, the job net to be migrated will be migrated separately to destination cloud 3 by its start time.

[0131] Figure 18, which uses the same reference numerals as Figure 2 to indicate corresponding parts, shows an example of the configuration of the operation management server 61 according to this embodiment. As shown in Figure 18, in the operation management server 61 of this embodiment, in addition to the business schedule retention table 30, job volume mapping table 31, job predicted execution time retention table 32, volume copy predicted time retention table 33, and volume copy start time retention table 34, the transaction count retention table 70 is stored in the database 71.

[0132] This transaction count table 70 is used to manage the number of transactions that each business server 62 in the source cloud 2 uses for each volume VOL. As shown in Figure 19, it is configured to include a volume column 70A and a transaction count column 70B. In the transaction count table 70, one record (row) corresponds to one volume VOL in which the business server 62 stores transaction data.

[0133] The volume column 70A stores the identification information of the corresponding volume VOL, and the transaction count column 70B stores the number of transactions (transaction count) that used that volume VOL in the corresponding job 12. Therefore, in the example in Figure 19, it is shown that there were "30,000" transactions that used, for example, the volume VOL "vol1" in the corresponding job. The transaction count for each volume VOL is obtained from the business server 62.

[0134] Figure 20 shows the specific processing details of the job execution time prediction process performed by the job execution time prediction unit 72 (Figure 18) in this embodiment. The job execution time prediction unit 72 in this embodiment predicts the execution time for each job 12 of the job net to be migrated according to the processing procedure shown in Figure 20.

[0135] In practice, the job execution time prediction unit 72 starts this job execution time prediction process in parallel with the business schedule acquisition process (Figure 8) and job volume mapping process (Figure 9) performed by the business schedule acquisition unit 36 ​​and the job volume mapping unit 37.

[0136] The job execution time prediction unit 72 then first looks at the logs of past executions of each job 12 that make up the job net to be migrated, which are held by the source cloud 2, and obtains the execution time and the number of transactions at that time (S90). The number of transactions at this time is also obtained from the business server 62.

[0137] Next, the job execution time prediction unit 72 calculates the following formula for each job 12, based on the execution time and number of transactions obtained in step S90 when that job 12 was executed in the past:

number

[0138] Next, the job execution time prediction unit 72 calculates the predicted execution time for the next job 12 for each job 12 by multiplying the execution time per transaction calculated in step S91 by the number of transactions for the volume VOL used by that job 12, which is stored in the transaction count retention table 70 (S92).

[0139] Furthermore, the job execution time prediction unit 72 stores the predicted time for each job 12 calculated in step S92 into the job execution time prediction table 32 (Figure 5) (S93), and then terminates the job execution time prediction process.

[0140] With the operation management server 61 of this embodiment having the above configuration, it is possible to predict the execution time of each job 12 with greater accuracy compared to the first embodiment, where the average execution time of previously performed jobs 12 is used as the execution time of that job 12.

[0141] (4) Fourth embodiment Figure 21, which uses the same reference numerals as Figure 2, shows an operation management server 80 according to a fourth embodiment, which is applied to the information processing system 1 in Figure 1 instead of the operation management server 4 according to the first embodiment. This operation management server 80 is configured similarly to the operation management server 4 according to the first embodiment, except that it is configured to more accurately calculate the volume copy time and volume copy start time for each volume VOL to be copied, taking into account that the data transfer speed between the source cloud 2 and the destination cloud 3 differs depending on the time of day.

[0142] In practice, in the operation management server 80 of this embodiment, in addition to the business schedule retention table 30, job volume mapping table 31, job predicted execution time retention table 32, volume copy predicted time retention table 33, and volume copy start time retention table 34, a time-based data transfer volume retention table 81, as shown in Figure 22, is stored in the database 82.

[0143] The time-of-day data transfer volume retention table 81 is a table used to manage the amount of data transferred per unit time between the source cloud 2 and the destination cloud 3 for each time period. As shown in Figure 22, it is configured to include a time period column 81A and a data transfer volume column 81B. In the time-of-day data transfer volume retention table 81, one record (row) corresponds to one time period.

[0144] The time zone column 81A stores information representing the corresponding time zone. The data transfer volume column 81B stores the average value of the data transfer volume per unit time measured by the volume copy time prediction unit 83 of this embodiment for the corresponding time zone. Therefore, in the example in Figure 22, for example, it is shown that the data transfer volume per unit time during the time zone "00:00-01:00" is "20Gbps".

[0145] Figure 23 shows the process of creating a time-period data transfer volume retention table executed by the volume copy time prediction unit 83 of this embodiment. The volume copy time prediction unit 83 creates the time-period data transfer volume retention table 81 described above for Figure 22 according to the processing procedure shown in Figure 23.

[0146] In practice, once the job volume mapping process described above for Figure 9 by the job volume mapping unit 37 and the job execution time prediction process described above for Figure 10 by the job execution time prediction unit 38 are completed, the volume copy time prediction unit 83 starts the process of creating the time-based data transfer volume retention table shown in Figure 23.

[0147] The volume copy time prediction unit 83 first refers to logs held by the source cloud 2 and destination cloud 3 of past data transfers between the source cloud 2 and destination cloud 3, and obtains the amount of data transferred per unit time for each time period and the time required for that data transfer (transfer time) (S100).

[0148] Next, the volume copy time prediction unit 83 calculates the amount of data transferred per unit time for each time period using equation (1) above, based on the amount of data transferred and the transfer time for each time period acquired in step S100 (S101).

[0149] Next, the volume copy time prediction unit 83 creates a time-period data transfer volume retention table 81 using the data transfer volume per unit time for each time period calculated in step S101 (S102), and then terminates the process of creating this time-period data transfer volume retention table.

[0150] The volume copy time prediction unit 83 then performs the volume copy time prediction process described above for Figure 11. At this time, it uses the time-period data transfer volume retention table 81 created in the time-period data transfer volume retention table creation process to predict the volume copy time for each volume VOL that is to be copied.

[0151] On the other hand, Figure 24 shows the processing details of the volume copy start time calculation process performed by the volume copy start time calculation unit 84 (Figure 21) according to this embodiment. The volume copy start time calculation unit 84 calculates the start time of volume copying for each volume VOL used by each job 12 of the job net to be migrated, according to the processing procedure shown in Figure 24.

[0152] In practice, the volume copy start time calculation unit 84 starts the volume copy start time calculation process shown in Figure 24 once the volume copy time prediction process (Figure 11) by the volume copy time prediction unit 83 is completed.

[0153] The volume copy start time calculation unit 84 first obtains the start time and identification information of each job 12 constituting the job net from the business schedule storage table 30 (Figure 3) for each job net to be migrated from the source cloud 2 to the destination cloud 3 (S110).

[0154] Next, the volume copy start time calculation unit 84 obtains the execution time for each of the jobs 12 whose identification information was obtained in step S110 from the job prediction execution time retention table 32 (Figure 5) (S111).

[0155] Next, the volume copy start time calculation unit 84 searches the job volume mapping table 31 (Figure 4) for the volumes VOL used by all jobs 12 for which identification information was obtained in step S110, and obtains the predicted volume copy time to the destination cloud 3 for each of the detected volumes VOL from the volume copy prediction time retention table 33 (Figure 6) (S112).

[0156] Furthermore, the volume copy start time calculation unit 84 calculates the start time for each job net to be migrated based on the start time of the job net obtained in step S110 and the execution time of each job 12 that constitutes the job net obtained in step S111 (S113).

[0157] Next, the volume copy start time calculation unit 84 calculates, for each job net to be migrated, the estimated time (volume copy estimated time) for volume copying of each volume VOL used by each job 12 constituting that job net to the destination cloud 3 (S114).

[0158] Specifically, the volume copy start time calculation unit 84 first obtains the data transfer rate per unit time for the time period including the start time of each job 12 calculated in step S113 from the time period data transfer rate retention table 81 (Figure 22).

[0159] The volume copy start time calculation unit 84 then calculates the predicted volume copy time (volume copy prediction time) for each job net to be migrated, based on the volume size of each volume VOL used by each job 12 that constitutes the job net, and the amount of data transferred per unit time for each time period that includes the start time of each job 12 obtained in step S113.

[0160] The volume copy start time calculation unit 84 then stores the calculated predicted time for each volume (volume copy predicted time) in the volume copy predicted time retention table 33 (Figure 6).

[0161] Next, the volume copy start time calculation unit 84 calculates the start time for volume copying each volume VOL from the source cloud 2 to the destination cloud 3, based on the start time of each job 12 calculated in step S113 and the volume copy prediction time of the volume VOL used by each job 12 obtained in step S114, so that the volume copy of the volume VOL used by each job 12 is completed by the start time of each job 12 (S115).

[0162] Specifically, the volume copy start time calculation unit 84 calculates the start time of the volume copy of volume VOL used by job 12 by subtracting the predicted volume copy time of volume VOL used by job 12 from the start time of job 12.

[0163] Next, the volume copy start time calculation unit 84 stores the volume copy start times for each volume VOL calculated as described above in the volume copy start time holding table 34 (Figure 7) (S116), and then terminates the volume copy start time calculation process.

[0164] With the operation management server 80 of this embodiment having the above configuration, it is possible to calculate the volume copy time and volume copy start time for each volume VOL to be copied more accurately, taking into account that the data transfer speed between the source cloud 2 and the destination cloud 3 differs depending on the time of day. In turn, it is possible to suppress system downtime associated with the migration of job nets between cloud services with greater reliability.

[0165] (5) Other embodiments In the first to fourth embodiments described above, various programs for realizing various functions such as data migration (such as the business schedule acquisition unit 36, the job volume mapping unit 37, and the job execution time prediction unit 38) were installed on an operation management server 4, 50, 61, 80 consisting of a single computer device. However, the present invention is not limited to this, and such various programs may be distributed and installed on multiple computer devices constituting a distributed computing system, thereby realizing various functions such as data migration across the entire distributed computing system.

[0166] Furthermore, in the fourth embodiment described above, the volume copy start time calculation unit 84 calculates the predicted time for volume copying of volume VOL (volume copy prediction time). However, the present invention is not limited to this, and for example, in Figure 24, the volume copy time prediction unit 83 may perform the processes from steps S110 to S114 of the volume copy start time calculation process described above. [Industrial applicability]

[0167] The present invention can be widely applied to data migration devices with various configurations that perform data migration associated with system migration between cloud services. [Explanation of symbols]

[0168] 1, 60... Information processing system, 2... Source cloud, 3... Destination cloud, 4, 50, 61, 80... Operation management server, 11... Job net management file, 12... Job, 20... CPU, 30... Business schedule retention table, 31... Job volume mapping table, 32... Job predicted execution time retention table, 33... Volume copy predicted time retention table, 34... Volume copy start time retention table, 36... Business schedule acquisition unit, 37... Job volume mapping unit, 38, 72... Job execution time prediction unit, 39, 83... Volume copy time prediction unit, 40, 84... Volume copy start time calculation unit, 41, 51... Volume copy execution unit, 42... Job monitoring unit, 43... Volume copy start time change unit, 62... Business server, 70... Transaction count retention table, 81... Data transfer volume retention table per time period, VOL... Volume.

Claims

1. In a data migration device that migrates data from each volume used by each job constituting a jobnet executed in the first source cloud service to the second destination cloud service, A schedule acquisition unit that acquires the execution schedule of the aforementioned job net, A job execution time prediction unit that predicts the execution time of each of the aforementioned jobs, A volume copy time prediction unit predicts the volume copy time for each volume used by each of the jobs, which is the time required to copy the volume from the first cloud service to the second cloud service. A volume copy start time calculation unit calculates the start time of each job based on the execution schedule of the job net and the execution time of each job, and calculates the start time of volume copying for each volume based on the calculated start time of each job and the volume copy time of each volume, so that volume copying is completed by the start time of the corresponding job. A volume copy execution unit that starts a volume copy of each volume from the first cloud service to the second cloud service at the volume copy start time calculated for each volume. A data migration device characterized by comprising the following features.

2. A job monitoring unit that monitors the execution status of the job in the second cloud service, If the execution status of the currently running job deviates from the execution time of the job predicted by the job execution time prediction unit, the volume copy start time change unit calculates the start time of a subsequent job to be executed after the currently running job, and changes the start time of the volume copy of the volume used by the subsequent job based on the calculated start time of the subsequent job. The data migration device according to claim 1, further comprising the features described above.

3. The job execution time prediction unit, Based on logs from previous executions of the job, the average execution time of past executions of the job is used to predict the current job's execution time. The data migration device according to feature 1.

4. The volume copy execution unit, After the volume copy of the aforementioned volume is completed, delete the volume in the first cloud service. The data migration device according to feature 1.

5. The volume copy execution unit, After the job in the second cloud service has finished executing, the volume used by the job will be deleted. The data migration device according to feature 4.

6. The job execution time prediction unit, Based on the number of business transactions for the volume, predict the execution time of the corresponding job. The data migration device according to feature 1.

7. The volume copy time prediction unit is, The amount of data transferred per unit time for each time period is obtained, and the volume copy time for each volume is predicted based on the obtained amount of data transferred per unit time for each time period, the start time of each job, and the volume size of each volume used by each job. The volume copy start time calculation unit, Based on the volume copy time predicted by the volume copy time prediction unit for each volume and the start time of each job, the start time for volume copying of each volume is calculated. The data migration device according to feature 1.

8. A data migration method performed by a data migration device that migrates data from each volume used by each job constituting a jobnet executed in a first cloud service (source) to a second cloud service (destination), The first step involves obtaining the execution schedule of the job net, predicting the execution time of each job, and further predicting the volume copy time for each volume used by each job, which is the time required to copy the volume from the first cloud service to the second cloud service. A second step involves calculating the start time for each job based on the execution schedule of the job net and the execution time of each job, and then calculating the start time for the volume copy of each volume based on the calculated start time for each job and the volume copy time of each volume, such that the volume copy is completed by the start time of the corresponding job. A third step is to start the volume copy of each volume from the first cloud service to the second cloud service at the start time of the volume copy calculated for each volume. A data migration method characterized by comprising the following features.

9. A fourth step involves monitoring the execution status of the job in the second cloud service, and if the execution status of the running job deviates from the predicted execution time of the job, calculating the start time of a subsequent job to be executed after the running job, and changing the start time of the volume copy of the volume used by the subsequent job based on the calculated start time of the subsequent job. The data migration method according to claim 8, further characterized by recording.

10. In the first step described above, the data migration device is Based on logs from previous executions of the job, the average execution time of past executions of the job is used to predict the current job's execution time. The data migration method according to feature 8.

11. In the third step described above, the data migration device, After the volume copy of the aforementioned volume is completed, delete the volume in the first cloud service. The data migration method according to feature 8.

12. In the third step described above, the data migration device, After the job in the second cloud service has finished executing, the volume used by the job will be deleted. The data migration method according to feature 11.

13. In the first step described above, the data migration device is Based on the number of business transactions for the volume, predict the execution time of the corresponding job. The data migration method according to feature 8.

14. In the first step described above, the data migration device is Obtain the amount of data transferred per unit of time for each time period, Based on the amount of data transferred per unit time for each acquired time period, the start time of each job, and the volume size of each volume used by each job, the volume copy time for each volume is predicted. In the second step described above, the data migration device, Based on the predicted volume copy time for each volume and the start time of each job, the start time for volume copying of each volume is calculated. The data migration method according to feature 8.

Images