WO2017109890A1

WO2017109890A1 - Management computer and batch processing execution method

Info

Publication number: WO2017109890A1
Application number: PCT/JP2015/085993
Authority: WO
Inventors: 貴志爲重; 水野　潤; 大介飯塚
Original assignee: 株式会社日立製作所
Priority date: 2015-12-24
Filing date: 2015-12-24
Publication date: 2017-06-29

Abstract

Provided is a management computer for managing a plurality of computers providing computer resources, wherein: a business system is built on the plurality of computers; the business system executes batch processing composed of a plurality of jobs; the management computer has a batch assignment management unit and a batch execution unit; the batch assignment management unit manages batch processing management information including information relating to a first resource having a fixed usage fee, job management information for managing job composition, policy management information for defining an operation policy of the business system, and history information of a second resource having a fluctuating usage fee; and the batch assignment management unit generates prediction information from the history information, determines whether the second resource can be used on the basis of the prediction information and the policy management information, and if the second resource can be used, generates execution plan information by calculating a resource amount of the second resource necessary for each of the plurality of jobs.

Description

Management computer and batch processing execution method

The present invention relates to a batch processing execution method executed by a system on a cloud service.

A user system (tenant) is constructed on a computer system that provides a cloud service, and various processes are performed on the tenant. The tenant includes a virtual computer and a virtual switch. The user uses the computer resources of the computer system according to the tenant's requirements and processing contents. Hereinafter, in this specification, computer resources are also simply referred to as resources.

Various methods are known for optimizing the resources allocated to tenants (for example, see Patent Document 1). Patent Document 1 discloses a resource allocation method for maximizing an application level performance index.

International Publication No. 2008/132924

Currently, a cloud service is provided in which the resource usage fee varies depending on the amount of free resources in the computer system or the usage status of the computer system.

The provider of the cloud service can grasp the free resource amount of the computer system and the usage status of the computer system, but the user cannot grasp them. For this reason, a fixed amount of resources is allocated to a standard process such as a batch process.

For the reasons described above, conventionally, the amount of resources allocated to batch processing has not been optimized according to resource usage fees and user requests. Therefore, a cloud service that satisfies user requirements such as cost reduction, batch processing execution time reduction, and batch processing multiple execution has not been provided.

A typical example of the invention disclosed in the present application is as follows. That is, a management computer that manages a plurality of computers that provide computer resources, wherein the management computer includes a first processor, a first memory connected to the first processor, and the first processor. A first interface for connecting to a device, wherein each of the plurality of computers includes a second processor, a second memory connected to the second processor, and the second processor; And having a second interface for connecting to a device, on the plurality of computers, at least one business system for providing a predetermined business is constructed using computer resources of the plurality of computers The at least one business system executes a batch process including a plurality of jobs, and the management computer performs the batch process. A batch allocation management unit that manages allocation of computer resources, and a batch execution unit that controls execution of the batch processing, wherein the batch allocation management unit includes a start time of the batch processing and the batch processing Batch processing management information including information related to the first resource, which is a computer resource with a fixed usage charge regardless of the usage time, and job management information for managing the configuration of the plurality of jobs of the batch processing, Policy management information for managing a computer resource allocation policy for the batch processing, and free resource amount history information for managing a history of the free capacity of the second resource, which is a computer resource whose usage fee varies depending on the usage time, Charge history information for managing a history of usage charges of the second resource, and the batch allocation management unit Generating first resource amount prediction information for predicting a free capacity of the second resource based on the resource amount history information, and predicting a usage fee of the second resource based on the fee history information Whether or not the second resource can be used for the batch processing based on the first resource amount prediction information, the first fee prediction information, and the policy management information is generated. And when it is determined that the second resource can be used for the batch processing, the second resource necessary for each of the plurality of jobs is determined based on the batch processing management information and the job management information. Calculating a resource amount, generating first execution plan information for reserving a resource amount of the second resource necessary for each of the plurality of jobs, and the batch execution unit, The execution of the batch process is controlled based on the first execution plan information.

According to the present invention, it is possible to execute a batch process that satisfies a user's request using a resource whose usage fee varies. That is, it is possible to provide a cloud service that satisfies the user's request by shortening the completion time of batch processing and reducing operation costs. Problems, configurations, and effects other than those described above will become apparent from the description of the following examples.

1 is a diagram illustrating an example of a system configuration of Embodiment 1. FIG. FIG. 3 is a block diagram illustrating a configuration of a management server according to the first embodiment. FIG. 2 is a block diagram illustrating a hardware configuration and a software configuration example of a server that implements IaaS of the first embodiment. FIG. 2 is a block diagram illustrating a hardware configuration and a software configuration example of a server that implements PaaS according to the first embodiment. It is a figure which shows an example of the system configuration information of Example 1. It is a figure which shows an example of the system configuration information of Example 1. FIG. 6 is a diagram illustrating an example of batch processing management information according to the first embodiment. 6 is a diagram illustrating an example of job management information according to the first exemplary embodiment. FIG. It is a figure which shows an example of the budget information of Example 1. FIG. 6 is a diagram illustrating an example of free resource amount history information according to the first embodiment. FIG. 6 is a diagram illustrating an example of free resource amount history information according to the first embodiment. It is a figure which shows an example of the charge history information of Example 1. FIG. It is a figure which shows an example of the charge history information of Example 1. FIG. It is a figure which shows an example of the policy management information of Example 1. It is a figure which shows an example of the change method definition information of Example 1. FIG. It is a figure which shows an example of the execution plan management information of Example 1. FIG. 6 is a flowchart illustrating an example of an execution plan generation process for batch processing according to the first exemplary embodiment. 6 is a flowchart illustrating an example of an execution plan generation process for batch processing according to the first exemplary embodiment. 6 is a flowchart illustrating an example of an execution plan change process according to the first embodiment. 6 is a flowchart illustrating an example of a resource amount calculation process (scale out) according to the first embodiment. 6 is a flowchart illustrating an example of a resource amount calculation process (scale-up) according to the first embodiment. 3 is a flowchart illustrating an example of execution processing of batch processing according to the first exemplary embodiment. 10 is a flowchart illustrating an example of an execution plan update process for a job according to the first exemplary embodiment. 10 is a flowchart illustrating an example of an execution plan update process for a job according to the first exemplary embodiment. 6 is a diagram illustrating an example of a user interface displayed by a display unit according to the first embodiment. FIG. It is a figure explaining an example of the effect of the process which the management server of Example 1 performs.

FIG. 1 is a diagram illustrating an example of a system configuration of the first embodiment.

The system of the first embodiment includes a plurality of computer systems operated by a cloud service provider. The computer system includes a management server 101, a server 102, a virtual server management server 151, a storage subsystem 105, NW-SWs 103 and 104, and an FC-SW 108.

The management server 101 manages a computer system group included in the management target system. The management server 101 is connected to the management interface (management I / F) 113 of the NW-SW 103 and the management interface 114 of the NW-SW (business network switch) 104 via the NW-SW (management network switch) 103. The management server 101 can set a VLAN (Virtual LAN) for each NW-

SW

103, 104.

Further, in addition to the management server 101 and the server 102, a virtual server management server 151 that manages a virtual server (virtual machine) running on the server 102 is connected to the NW-SW 103.

The NW-SW 103 constitutes a management network. The management network is a network for managing the OS and applications distributed on the plurality of physical servers 102 and the power control, etc. by the management server 101.

NW-SW 104 constitutes a business network. The business network is a network used by an application executed by a virtual server on the server 102. The NW-SW 104 is connected to a WAN or the like and communicates with a client computer outside the virtual computer system.

The management server 101 is connected to the storage subsystem 105 via an FC-SW (Fibre Channel switch) 108. The management server 101 manages LU (Logical Unit) in the storage subsystem 105. In the example shown in FIG. 1, the management server 101 manages N LU1 to LUn.

On the management server 101, a control unit 110 that manages a computer system such as the server 102 is executed. The control unit 110 refers to and updates the management information group 111.

The server 102 to be managed provides a virtual server as described later. The server 102 connects to the NW-

SWs

103 and 104 via the PCIex-SW 107 and the I / O device.

The PCI Express-SW 107 is connected to a PCI Express standard I / O device. The I / O device has an I / O adapter such as NIC (Network Interface Card), HBA (Host Bus Adapter), and CNA (Converged Network Adapter).

Generally, the PCIex-SW 107 is an I / O switch that extends a PCI Express bus from a motherboard (or server blade) to the outside and connects a number of PCI-Express devices. However, a system configuration in which the server 102 is directly connected to the NW-SW 103 and the NW-SW 104 without using the PCIex-SW 107 may be used.

Also, the management server 101 is connected to the management interface 117 of the PCIex-SW 107, and manages the connection of the plurality of servers 102 and I / O devices. Further, the server 102 accesses LU1 to LUn of the storage subsystem 105 via an I / O device (HBA in FIG. 1) connected to the PCIex-SW 107.

The virtual server management server 151 manages the first virtualization unit 311 (see FIG. 3) and the second virtual server 314 (see FIG. 3) executed by the server 102. Specifically, the virtual server management unit 161 instructs the first virtualization unit 311.

For example, the virtual server management unit 161 instructs power supply control of the second virtual server 314 and migration of the second virtual server 314 and the first virtualization unit 311. The management server 101 may include the virtual server management unit 161.

FIG. 2 is a block diagram illustrating the configuration of the management server 101 according to the first embodiment.

The management server 101 includes a processor 201, a memory 202, a disk interface 203, and a network interface 204.

The processor 201 executes a program stored in the memory 202. The memory 202 stores a program executed by the processor 201 and information necessary for executing the program. The program and information stored in the memory 202 will be described later.

The disk interface 203 is an interface for accessing the storage subsystem 105. The network interface 204 is an interface for communicating with other devices via the IP network.

In FIG. 2, one disk interface 203 and one network interface 204 are shown, but a plurality of interfaces may be provided. For example, the management server 101 may have a network interface connected to the NW-SW 103 and a network interface connected to the NW-SW 104.

Although not shown, the management server 101 may have a BMC (Basement Management Controller) that performs power control and control of each interface, and a PCI-Express interface for connection to the PCIex-SW 107.

The memory 202 stores a program for realizing the control unit 110 and a management information group 111. The control unit 110 includes a plurality of program modules and provides functions for performing various controls. Specifically, the control unit 110 includes a batch assignment management unit 211, a batch execution unit 212, and a display unit 213. The control unit 110 includes program modules other than those described above, but omits program modules not directly related to the present embodiment.

Although it is divided into three functional parts for explanation, it is not limited to this. One functional unit may be divided into a plurality of functional units, or a plurality of functional units may be integrated into one functional unit. For example, the batch assignment management unit 211 and the batch execution unit 212 may be a single functional unit.

The processor 201 loads the function units of the batch assignment management unit 211, the batch execution unit 212, and the display unit 213 as programs into the memory 202, and executes the loaded programs.

The processor 201 operates as a functional unit that realizes a predetermined function by operating according to a program of each functional unit. For example, the processor functions as the batch assignment management unit 211 by operating according to a program that implements the batch assignment management unit 211. The same applies to other programs. Furthermore, the processor 201 also operates as a functional unit that implements each of a plurality of processes executed by each program.

The batch allocation management unit 211 manages resource allocation for batch processing. Specifically, the batch allocation management unit 211 generates an execution plan for batch processing based on the usage status of the computer system and the policy for batch processing. The batch execution unit 212 controls the execution of batch processing based on the execution plan generated by the batch assignment management unit 211. The display unit 213 displays various information.

The management information group 111 includes system configuration information 221, batch processing management information 222, job management information 223, budget information 224, free resource amount history information 225, fee history information 226, policy management information 227, change method definition information 228, and Execution plan management information 229 is included. The management information group 111 includes other information such as business setting information provided using a tenant, but information that is not directly used in the processing of this embodiment is omitted.

System configuration information 221 is information for managing the tenant configuration. Details of the system configuration information 221 will be described with reference to FIGS. 5A and 5B. The batch processing management information 222 is information for managing batch processing executed in each tenant. Details of the batch processing management information 222 will be described with reference to FIG. The job management information 223 is information for managing jobs included in one batch process. Details of the job management information 223 will be described with reference to FIG.

Budget information 224 manages the budget of users who use one or more tenants. Details of the budget information 224 will be described with reference to FIG. The free resource amount history information 225 is history information related to the resource amount of free resources. Details of the free resource amount history information 225 will be described with reference to FIGS. 9A and 9B. The fee history information 226 is history information related to the usage fee of the free resource. Details of the charge history information 226 will be described with reference to FIGS. 10A and 10B.

Policy management information 227 is information indicating a resource allocation policy for each batch process. Details of the policy management information 227 will be described with reference to FIG. The change method definition information 228 is definition information of a resource allocation change method for batch processing. Details of the change method definition information 228 will be described with reference to FIG.

The execution plan management information 229 is information for managing an execution plan for batch processing generated by the batch assignment management unit 211. The execution plan includes information related to the start time of batch processing, the amount of resources to be reserved, the type of resources to be reserved, and the like. The batch execution unit 212 executes batch processing based on the execution plan management information 229. Details of the execution plan management information 229 will be described with reference to FIG.

FIG. 3 is a block diagram illustrating a hardware configuration and a software configuration example of the server 102 that realizes IaaS (Infrastructure as a Services) of the first embodiment. FIG. 3 shows an example of a software group that can configure a multistage virtual machine. Note that the virtualization infrastructure may or may not be nested.

First, the hardware configuration of the server 102 will be described. The server 102 includes a processor 301, a memory 302, a network interface 303, a disk interface 304, a BMC 305, and a PCI-Express interface 306.

The processor 301 executes a program stored in the memory 302. The memory 302 stores a program executed by the processor 301 and information necessary for executing the program. The program and information stored in the memory 302 will be described later.

The network interface 303 is an interface for communicating with other devices via the IP network. The disk interface 304 is an interface for accessing the storage subsystem 105. The BMC 305 performs power supply control and control of each interface. The PCI-Express interface 306 is an interface for connecting to the PCIex-SW 107.

In FIG. 3, one network interface 303, one disk interface 304, and one PCI-Express interface 306 are shown, but a plurality of interfaces may be provided. For example, the server 102 may have a network interface connected to the NW-SW 103 and a network interface connected to the NW-SW 104.

The memory 302 stores a program for realizing the first virtualization unit 311 and the second virtualization unit 313, and stores a data group that is an entity of the first virtual server 312 (or logical partition) and the second virtual server 314. Store.

The first virtualization unit 311 allocates physical resources to a plurality of first virtual servers 312 (or logical partitions). The second virtualization unit 313 allocates resources of the first virtual server 312 to the plurality of second virtual servers 314.

The first virtualization unit 311 that virtualizes the resources of the server 102 is deployed as a lower layer virtualization unit in the memory 302, and the resource (first cluster) is provided to the plurality of second virtualization units 313 that are upper layer virtualization units. 1 virtual server 312). The second virtualization unit 313 generates a plurality of second virtual servers 314 and stores them in the memory 302. The first virtualization unit 311 includes a virtualization unit management interface 351 as a control interface. Although not shown, the second virtualization unit 313 also includes a virtualization unit management interface as a control interface.

The first virtualization unit 311 generates a plurality of first virtual servers 312 by virtualizing resources of the server 102 (or blade server) and allocating the virtualized resources to the plurality of first virtual servers 312. The first virtualization unit 311 can be realized using, for example, a hypervisor or a VMM (Virtual Machine Monitor).

The second virtualization unit 313 further virtualizes the resource (first virtual server 312) provided by the first virtualization unit 311, and uses the resources of the first virtual server 312 virtualized by the plurality of second virtual servers 314. By assigning, a plurality of second virtual servers 314 are generated. The second virtualization unit 313 can be realized using, for example, a hypervisor or a VMM.

The first virtual server 312 includes a virtual device (or logical device) provided by the first virtualization unit 311. The virtual device of the present embodiment includes a virtual processor 321, a virtual memory 322, a virtual network interface 323, a virtual disk interface 324, a virtual BMC 325, and a virtual PCIex interface 326. The virtual device described above is one that is assigned to each first virtual server 312 by the first virtualization unit 311.

Further, the second virtual server 314 is also composed of similar virtual devices. Note that the virtual device provided to the second virtual server 314 is the resource (first virtual server 312) allocated by the first virtualization unit 311 to the plurality of second virtualization units 313 and the second virtualization unit 313. This is realized by allocating to each second virtual server 314.

The OS 331 is stored in the virtual memory 322. The OS 331 manages virtual devices provided to the first virtual server 312 or the second virtual server 314. An application 341 is executed on the OS 331. The management program 342 running on the OS 331 realizes functions such as failure detection, virtual server power control, and inventory management.

The first virtualization unit 311 manages the association between the physical resources of the server 102 and the resources allocated to the second virtualization unit 313. In the present embodiment, the example in which the first virtualization unit 311 allocates the resources of the first virtual server 312 to the second virtualization unit 313 has been described. However, the first virtualization unit 311 allocates the resources of the physical server 102 to the second virtual unit 313. You may make it allocate directly to the conversion part 313. In this case, the first virtual server 312 can be omitted.

The first virtualization unit 311 can dynamically change the resource of the server 102 to be allocated to the plurality of second virtualization units 313 and can cancel the allocation of the resource. The first virtualization unit 311 holds the resource amount, configuration information, and operation history of the resources allocated to the second virtualization unit 313.

The second virtualization unit 313 manages the association between the resources of the first virtual server 312 and the resources allocated to the second virtual server 314. The second virtualization unit 313 can dynamically change the resources of the first virtual server 312 to be allocated to the plurality of second virtual servers 314 and can cancel the allocation of the resources. Further, the second virtualization unit 313 holds the resource amount, configuration information, and operation history of the resources allocated to the second virtual server 314.

In this embodiment, the first virtualization unit 311 is the first layer, the second virtualization unit 313 is the second layer, and the OS 331 is the third layer. The third layer side is the upper layer, and the first layer side is the lower layer. However, if the multi-stage configuration is not used, the first virtualization unit 311 is the first layer, and the OS 331 operates in the upper layer.

FIG. 4 is a block diagram illustrating a hardware configuration and a software configuration example of the server 102 that realizes PaaS (Platform as a Services) of the first embodiment.

The hardware configuration of the server 102 is the same as the hardware configuration of the server 102 shown in FIG.

The memory 302 stores a program for realizing the OS 401, the application 402, and the monitoring unit 403. The application 402 and the monitoring unit 403 operate on the OS 401. Note that the memory 302 may store a program for realizing the first virtualization unit 311 and the like.

In this embodiment, the user constructs a virtual board (tenant) for executing a predetermined task using one or more servers 102 as shown in FIG. 3 or FIG. The user executes a plurality of batch processes using a tenant. In addition, the user pays the cloud service provider according to the amount of resources used by the tenant.

Conventionally, in a computer system in which the resource amount and usage fee of a free resource fluctuate due to various factors, a user cannot predict the resource amount and usage fee of a free resource. Batch processing is set to be assigned. Therefore, the computer resource usage fee for batch processing is a fixed fee.

In the following explanation, a resource with a fixed usage charge is described as a fixed price resource regardless of the usage time. Also, a resource whose usage fee varies according to usage time is also referred to as a price variation resource.

The management server 101 according to the present embodiment generates an execution plan that defines resources to be allocated to batch processing so as to satisfy the user policy based on the resource amount of the price fluctuation resource and the usage charge prediction information. As a result, resource allocation for batch processing is dynamically changed. Furthermore, the management server 101 presents an additional batch process related to the batch process, and also generates an execution plan for the additional batch process.

5A and 5B are diagrams illustrating an example of the system configuration information 221 according to the first embodiment.

The system configuration information 221 includes an identifier 501, a tenant ID 502, a system ID 503, a hardware characteristic 504, and a software characteristic 505.

The identifier 501 is an identifier for uniquely identifying an entry of the system configuration information 221. In this embodiment, it is assumed that an ascending value is automatically set in the identifier 501. A value of any column of the system configuration information 221 or a value obtained by combining values of a plurality of columns may be used as the identifier of the entry. In this case, the identifier 501 may be omitted.

The tenant ID 502 is an identifier for uniquely identifying a tenant. In this embodiment, it is assumed that one or more tenants exist for one user.

The system ID 503 is an identifier for identifying devices that constitute a tenant. When the server 102 is assigned as a device to the tenant, the identifier of the server 102 is stored in the system ID 503, and when the first virtual server 312 or the second virtual server 314 is assigned as a device to the tenant, The identifier of the virtual server is set in the system ID 503.

Hardware characteristics 504 indicate the hardware characteristics of the device corresponding to the system ID 503. For example, HBA, NIC, CNA, etc., HBA identifier WWN, NIC identifier MAC address, performance information, architecture information, generation information, model number, support function, vendor type, firmware information, driver information, I / F information, switch information, RAID information, virtualization type, virtualization correspondence information, and the like are stored.

Software characteristic 505 indicates the software characteristic of the device corresponding to the system ID 503. For example, OS type, OS version, middleware type, middleware version, application information, management program information, and the like are stored.

The information stored in each column of the system configuration information 221 is an example and does not limit the present invention.

FIG. 6 is a diagram illustrating an example of the batch processing management information 222 according to the first embodiment.

The batch processing management information 222 includes a tenant ID 601, job net ID 602, type 603, start time 604, target time 605, instance number 606, instance type 607, business type 608, execution time 609, and fee 610.

The tenant ID 601 is the same as the tenant ID 502. The job net ID 602 is an identifier for uniquely identifying batch processing executed in the tenant. The batch processing includes a plurality of jobs that are the minimum unit of processing. A job net includes a plurality of jobs, and each job is connected to at least one job. In this embodiment, batch processing is managed as a job net.

Type 603 is a type of batch processing. The type 603 stores at least one of “normal” and “addition”. “Normal” indicates a batch process planned by the user. “Addition” indicates a batch process newly added by an execution plan generation process of a batch process to be described later. In the following description, a batch process planned by the user is described as a main batch process, and a batch process added in an execution plan generation process and an execution plan update process described later is described as an additional batch process. The start time 604 is the start time of batch processing. The target time 605 is a batch processing completion time when only the fixed price resource is used.

The instance number 606 indicates the number of instances allocated to batch processing. Here, the instance indicates a device that executes jobs such as the server 102, the first virtual server 312, and the second virtual server 314. Note that the number of instances of the fixed price resource is set in the instance number 606 of the main batch process. That is, in the initial setting, a value is set in the main batch process so that a fixed price resource is reserved. The instance type 607 is an instance type. In this embodiment, it is assumed that a usage fee is set for each instance type.

The business type 608 is a type of batch processing. In this embodiment, the business type 608 stores the name of an application executed in batch processing. The execution time 609 is the execution time of batch processing. The fee 610 is a usage fee for resources used by batch processing. The fee 610 stores a usage fee when the number of instances indicated by the instance number 606 is used for the time indicated by the execution time 609.

FIG. 7 is a diagram illustrating an example of the job management information 223 according to the first embodiment. One job management information 223 exists for one batch process (job net).

The job management information 223 includes a job ID 701, a parallel degree 702, a maximum parallel degree 703, a parallelization time 704, a required performance 705, an execution time 706, and an output job ID 707.

The job ID 701 is an identifier that uniquely identifies a job included in a batch process (job net).

The degree of parallelism 702 is the number of jobs executed in parallel. An instance (resource) matching the value of the degree of parallelism 702 is assigned to the job executed in parallel. For example, the number of parallel executions of a job whose job ID 701 is “job 1” in FIG. 7 is “2”. Therefore, two instances are assigned to the job, and the two instances execute the same job in parallel. The maximum parallelism 703 is the maximum value of the number of jobs executed in parallel. As will be described later, when the value of the degree of parallelism 702 is smaller than the value of the maximum degree of parallelism 703, it indicates that the job can add resources.

The parallelization time 704 is a processing time required for executing jobs in parallel. The required performance 705 is performance required for an instance executing a job, that is, a resource amount. In this embodiment, there are two methods for changing the resource allocation: a method for increasing / decreasing the number of instances, and a method for increasing / decreasing the resources allocated to the instances. The execution time 706 is a job execution time.

The output job ID 707 is an identifier of a job to which a job processing result corresponding to the job ID 701 is output. Based on the output job ID 707, the management server 101 can grasp the connection relationship of jobs in the job net. In the case of a job that does not output the processing result to another job, the output job ID 707 is blank. For example, the output job ID 707 of the job executed at the end of the job net is blank.

In this embodiment, the management server 101 holds the job management information 223 as job initial setting information. This makes it possible to compare the execution time before and after the resource allocation change, performance, price, and the like.

The batch process management information 222 and the job management information 223 are information generated from information input when the user constructs a tenant. The management server 101 may generate the batch processing management information 222 and the job management information 223 based on the current tenant setting information and the like.

FIG. 8 is a diagram illustrating an example of the budget information 224 according to the first embodiment.

Budget information 224 includes tenant ID 801, date 802, and budget 803.

The tenant ID 801 is the same as the tenant ID 502. One tenant entry includes budget rows with different dates. A date 802 is a date on which the set budget is applied. A budget 803 is a budget applied on the date designated by the date 802. The budget may be managed for each user.

In the budget information 224 shown in FIG. 8, a monthly budget for each tenant is set. Therefore, year and month are set for the date 802. In the budget 803, an operation budget for one month is set. Note that the budget setting period can be changed as appropriate.

The budget information 224 is information generated from information input when a user constructs a tenant or the like. Note that the management server 101 can appropriately update the budget information 224 based on a request from the user.

9A and 9B are diagrams illustrating an example of the free resource amount history information 225 according to the first embodiment. The free resource amount history information 225 of the present embodiment includes first resource amount history information 900 and second resource amount history information 910.

The first resource amount history information 900 is history information on the number of unused instances provided as price fluctuation resources. The first resource amount history information 900 includes a date 901 and columns of a plurality of instance types.

The date 901 is the date when the history is acquired. In this embodiment, there are six types of instances, and the first resource amount history information 900 includes columns corresponding to the types of instances. Specifically, the first resource amount history information 900 includes Physical Server 902, VM xlarge 903, VM large 904, VM medium 905, VM small 906, and VM tiny 907. Each column stores the number of unused instances corresponding to the instance type.

The second resource amount history information 910 is history information on the amount of free resources of a device that allocates resources to instances provided as price fluctuation resources. The second resource amount history information 910 includes a date 911, an instance type 912, a system ID 913, a processor 914, a memory 915, and an IOPS 916.

Date 911 is the same as date 901. The instance type 912 is an instance type. The system ID 913 is an identifier of an instance included in the price fluctuation resource. The system ID 913 stores an identifier similar to the system ID 503.

The processor 914, the memory 915, and the IOPS 916 are free resource amounts of the instance. The processor 914 stores the number of processors that are not used. The memory 915 stores the capacity of the memory that is not used. The IOPS 916 is an IOPS (I / O per Second) indicating the I / O performance of the instance. The IOPS is not a free resource, but is used as one of resource allocation indexes. Therefore, the second resource amount history information 910 includes the IOPS of each instance.

10A and 10B are diagrams illustrating an example of the charge history information 226 according to the first embodiment. The fee history information 226 of the present embodiment includes first fee history information 1000 and second fee history information 1010.

The first fee history information 1000 is usage fee history information for each instance type provided as a price fluctuation resource. The first fee history information 1000 includes a date 1001 and a column of a plurality of instance types.

The date 1001 is the same as the date 901. The first fee history information 1000 includes a physical server 1002, a VM xlarge 1003, a VM large 1004, a VM medium 1005, a VM small 1006, and a VM tiny 1007 as columns of instance types. Each column stores a usage fee for one instance corresponding to the instance type.

The second fee history information 1010 is the history information of the usage fee of the free resource of the device that allocates the resource to the instance provided as the price fluctuation resource. The second fee history information 1010 includes a date 1011, an instance type 1012, a fee (processor) 1013, a fee (memory) 1014, and a fee (IOPS) 1015.

The date 1011 is the same as the date 1001. The instance type 1012 is the same as the instance type 912. In the second fee history information 1010, a usage fee history is managed for each instance type. The fee (processor) 1013 stores the usage fee of the processor of the instance corresponding to the instance type 1012. The fee (memory) 1014 stores a memory usage fee of the instance corresponding to the instance type 1012. The charge (IOPS) 1015 stores the IOPS usage charge of the instance corresponding to the instance type 1012.

FIG. 11 is a diagram illustrating an example of the policy management information 227 according to the first embodiment.

Policy management information 227 manages a policy preset by the user. For example, when registering a batch process, the user sets a policy for the batch process (job net). The timing for setting the policy is not limited to this. The policy management information 227 includes a job net ID 1101, a policy ID 1102, a policy type 1103, a change method 1104, a batch processing addition condition 1105, and a selection flag 1106.

The job net ID 1101 is the same as the job net ID 602. In this embodiment, one or more policies are set for one job net. The policy ID 1102 is an identifier for uniquely identifying the policy set for the job net. The policy type 1103 is a policy type. The change method 1104 is definition information of a resource allocation change method. Note that the change method 1104 may store a program, a processing code, and the like.

The batch process addition condition 1105 is a condition for adding a new batch process. The batch processing addition condition 1105 of “Policy 2” indicates that the addition of the batch processing is permitted if it can be added within a fee exceeding 10% of the budget. The batch processing addition condition 1105 of “Policy 3” indicates that the addition of the batch processing is permitted if it can be added within the budget.

The selection flag 1106 is a flag indicating whether or not the policy is selected by the user. In this embodiment, “ON” is set in the policy selection flag 1106 selected by the user. In this embodiment, when one policy is set for one job, “ON” is automatically set in the selection flag 1106. In addition, when a plurality of policies are set for one job, the policy to be applied by the user is selected.

FIG. 12 is a diagram illustrating an example of the change method definition information 228 according to the first embodiment.

The change method definition information 228 manages the resource allocation change method preset by the administrator. The change method definition information 228 includes a method ID 1201, a process type 1202, a target 1203, and an additional amount 1204.

Method ID 1201 is an identifier for uniquely identifying the change method registered in the change method definition information 228. The process type 1202 is a type of change method. In this embodiment, either “scale out” or “scale up” is stored in the processing type 1202.

“Scaling out” is a method of changing resources allocated to a job by adding instances. “Scale-up” is a method of changing the resource allocated to a job by adding the performance of the instance, that is, the resource allocated to the instance.

Target 1203 is a resource to be added. The addition amount 1204 is a change amount of the resource to be added.

FIG. 13 is a diagram illustrating an example of the execution plan management information 229 according to the first embodiment.

In the execution plan management information 229, one entry corresponds to one batch processing execution plan. One entry includes a job net ID 1301, method ID 1302, job ID 1303, parallelism 1304, parallelization time 1305, required performance 1306, execution time 1307, fee 1308, efficiency 1309, confirmation flag 1310, and execution flag 1311.

The job net ID 1301 is the same as the job net ID 602. The method ID 1302 is the same as the method ID 1201. The job ID 1303 is the same as the job ID 701. The parallelism 1304, the parallelization time 1305, the required performance 1306, and the execution time 1307 are the same as the parallelism 702, the parallelization time 704, the required performance 705, and the execution time 706. When the resource allocation is changed, values after the resource allocation is changed are stored in the degree of parallelism 1304, the parallelization time 1305, the required performance 1306, and the execution time 1307.

The fee 1308 is a fee for using resources used when executing a job. Efficiency 1309 is an index indicating cost performance due to a change in resource allocation.

The confirmation flag 1310 is a flag indicating that the execution plan is actually used. That is, the confirmation flag 1310 indicates that the job is executed based on the generated execution plan. The value of “0” or “1” is stored in the confirmation flag 1310 of this embodiment. “0” indicates that the execution plan is not adopted, and “1” indicates that the execution plan is adopted.

The execution flag 1311 is a flag indicating whether or not the job has been executed. The execution flag 1311 of this embodiment stores either “0” or “1”. “0” indicates that the job has been executed, and “1” indicates that the job has not been executed.

Next, an execution plan generation process for batch processing will be described. 14A and 14B are flowcharts illustrating an example of an execution plan generation process for batch processing according to the first embodiment.

The batch allocation management unit 211 starts processing at a predetermined opportunity. For example, it is executed when a batch process is newly registered, when the free resource amount history information 225 or the like is updated, or when the change rate of the free resource amount of the price fluctuation resource is larger than a predetermined threshold. Also, the batch allocation management unit 211 periodically refers to the start time 604 of the batch processing management information 222 and determines whether there is a batch process that is executed within one hour from the current time. Good.

In the following explanation, a case where a new batch process is registered will be described as an example. At this time, a list including a plurality of entries including the identifier of the newly registered batch process (job net) and the identifier of the tenant on which the batch process is executed is input to the batch assignment management unit 211 as input information. .

First, the batch allocation management unit 211 predicts the state of the computer system (step S1401). Specifically, the following processing is executed.

The batch allocation management unit 211 performs machine learning using the free resource amount history information 225, and generates a free resource amount prediction function with time as a variable as resource amount prediction information.

In the following description, the prediction function generated based on the first resource amount history information 900 is described as a first resource amount prediction function, and the prediction function generated based on the second resource amount history information 910 is referred to as a second resource. Also described as quantity prediction function. In the present embodiment, the batch allocation management unit 211 can calculate the number of instances using the first resource amount prediction function, and calculate the resource amount allocated to one instance using the second resource amount prediction function. Can do. Note that there are as many first resource amount prediction functions and second resource amount prediction functions as the number of instance types.

The batch allocation management unit 211 performs machine learning using the fee history information 226, and generates a function for predicting the usage fee of the price fluctuation resource with time as a variable, as usage fee prediction information.

In the following description, the prediction function generated based on the first fee history information 1000 is described as a first fee prediction function, and the prediction function generated based on the second fee history information 1010 is referred to as a second fee prediction function. Describe. Note that there are as many first charge prediction functions and second charge prediction functions as the number of instance types.

For example, the batch allocation management unit 211 executes machine learning using history information in units of one week, and generates each prediction function from the present to one week later. Since machine learning using history information is a well-known technique, description thereof is omitted. The above is the description of the processing in step S1401.

Next, the batch assignment management unit 211 starts a loop process for batch processing (step S1402). Specifically, one target batch process is selected from a plurality of batch processes. In the following description, the batch process selected in step S1402 is also referred to as a selected batch process.

As a selection method of batch processing, a method of selecting in order from an entry on the list, a method of selecting in order from an important batch processing, and the like can be considered. Further, the batch allocation management unit 211 may select a batch processing group of a tenant with a large operation budget and select a plurality of batch processing of the tenant using the method described above.

Next, the batch assignment management unit 211 determines a policy for the selected batch process (step S1403). Specifically, the following processing is executed.

The batch assignment management unit 211 refers to the policy management information 227 and searches for an entry in which the job net identifier corresponding to the selected batch process is stored in the job net ID 1101.

When a plurality of entries are searched, the batch assignment management unit 211 presents a selectable policy to the user via the display unit 213, and accepts input from the user. In this case, the batch assignment management unit 211 sets “ON” in the selection flag 1106 of the entry corresponding to the policy selected by the user.

If there is no policy applicable to the selected batch process, or if the user does not select a policy, the batch assignment management unit 211 proceeds to step S1414 without executing any particular process. The above is the description of the process in step S1403.

Next, the batch assignment management unit 211 starts a loop process of the changing method (step S1404). Specifically, the following processing is executed.

The batch allocation management unit 211 refers to the change method definition information 228 and selects one change method to be used. For example, the batch assignment management unit 211 sequentially selects entries from the change method definition information 228 in order. Further, based on the policy, applicable change methods may be listed, and one change method to be used may be selected from the list.

The batch allocation management unit 211 adds an entry to the execution plan management information 229, and sets the job net identifier corresponding to the selected batch process in the job net ID 1301. Further, the batch assignment management unit 211 sets the identifier of the selected change method in the method ID 1302 of the added entry.

Furthermore, the batch assignment management unit 211 refers to the job management information 223 associated with the identifier of the job net corresponding to the selected batch process, and adds as many lines as the number of jobs to the added entry. The batch allocation management unit 211 includes a job ID 701, a degree of parallelism 702, a parallelization time 704, a required performance 705, a job ID 1303, a parallelism 1304, a parallelization time 1305, a required performance 1306, and an execution time 1307 of each added row. And the value of the execution time 706 is set. That is, the batch assignment management unit 211 copies the contents of the job management information 223 to the added entry.

Also, the batch assignment management unit 211 acquires the number of instances 606 of the selected batch process, the execution time 609, and the charge 610 from the batch process management information 222. The batch allocation management unit 211 calculates a charge per unit time (first charge) of one instance using the number of instances 606, the execution time 609, and the charge 610. For example, the batch allocation management unit 211 can calculate the first fee by dividing the value of the fee 610 by the value of the number of instances 606 and further dividing the calculated value by the execution time 609.

Further, the batch allocation management unit 211 calculates a fee required for job execution based on the parallelism 702, the parallelization time 704, and the execution time 706 of each job and the first fee. For example, the batch allocation management unit 211 can calculate the above-described fee by multiplying the first fee by the degree of parallelism 702 and multiplying the calculated value by a value obtained by adding the parallelization time 704 and the execution time 706.

The batch allocation management unit 211 sets the calculated value in the charge 1308. Further, the batch allocation management unit 211 sets “1” for the efficiency 1309 of the added entry, and sets “0” for the confirmation flag 1310 and the execution flag 1311. The above is the description of step S1404.

Next, the batch assignment management unit 211 starts a loop process of the selected batch process job (step S1405). In the following description, the job selected in step S1405 is also referred to as a selected job. Specifically, the following processing is executed.

The batch assignment management unit 211 grasps the connection relation of a plurality of jobs included in the selected batch process based on the output job ID 707 of the job management information 223 for the selected batch process.

The batch allocation management unit 211 selects jobs in the execution order based on the connection relation of jobs. Note that when there are a plurality of jobs having a connection relationship to be output to the same job, the batch assignment management unit 211 selects one job from the plurality of jobs based on the job ID 701. The above is the description of the processing in step S1405.

Next, based on the determined policy, the batch assignment management unit 211 determines whether or not a price fluctuation resource is available for the selected job (step S1406). In this embodiment, as described below, the determination method differs depending on the policy.

(1) When the policy type 1103 of the determined policy is “Cost”, the following processing is executed.

The batch allocation management unit 211 refers to the batch process management information 222 and acquires the time from the start time 604 of the entry corresponding to the selected batch process.

When the change method target 1203 selected in step S1404 is “instance”, the batch allocation management unit 211 substitutes the time into the first fee prediction function corresponding to the type of the instance, and starts the selected batch process. The price (second charge) of the price fluctuation resource (instance) is calculated.

The batch allocation management unit 211 determines whether or not the second fee is lower than the first fee. That is, it is determined whether the price of the price fluctuation resource is lower than the price of the fixed price resource.

When it is determined that the second fee is lower than the first fee, the batch allocation management unit 211 determines that the price fluctuation resource is available because there is a possibility of reducing the fee.

(2) When the policy type 1103 of the determined policy is “target time”, the following processing is executed.

The batch allocation management unit 211 calculates the free resource amount (number of instances) of the price fluctuation resource when executing the selected job by substituting the time into the first resource amount prediction function corresponding to the instance type.

The batch allocation management unit 211 refers to the entry corresponding to the selected job in the job management information 223. The batch assignment management unit 211 determines whether or not the value of the parallel degree 702 of the referenced entry is smaller than the calculated number of instances. That is, it is determined whether or not the number of unused instances of the price fluctuation resource is larger than the number of instances assigned to the selected job.

If the value of the degree of parallelism 702 of the entry is smaller than the calculated number of instances, the batch allocation management unit 211 determines that the price fluctuation resource is available because there is a possibility that the execution time of the batch processing may be shortened. To do.

(3) When the policy type 1103 of the determined policy is “cost performance”, the batch allocation management unit 211 uses a determination method in which (1) and (2) are combined. That is, the batch allocation management unit 211 determines whether or not the second fee is lower than the first fee and the value of the degree of parallelism 702 is smaller than the calculated number of instances. If any determination result is YES, the batch allocation management unit 211 determines that the price fluctuation resource is available.

Note that the determination methods (1) to (3) described above are merely examples and are not limited thereto. For example, when the second fee is lower than the first fee regardless of the policy type 1103, it may be determined that the price fluctuation resource is available. The above is the description of the process in step S1406.

When it is determined that the price fluctuation resource is available, the batch allocation management unit 211 executes an execution plan change process (step S1407). Details of the execution plan change process will be described with reference to FIG.

Next, the batch assignment management unit 211 determines whether or not the processing has been completed for all jobs included in the selected batch processing (step S1408).

If it is determined that the processing has not been completed for all jobs included in the selected batch process, the batch assignment management unit 211 returns to step S1405, selects a new job, and executes the same process.

When it is determined that processing has been completed for all jobs included in the selected batch processing, the batch assignment management unit 211 determines whether processing has been completed for all change methods (step S1409). .

If it is determined that the processing has not been completed for all the change methods, the batch assignment management unit 211 returns to step S1404, selects a new change method, and executes the same processing.

When it is determined that the processing has been completed for all the change methods, the batch allocation management unit 211 determines whether there is an execution plan that satisfies the policy based on the entry group added to the execution plan management information 229. Is determined (step S1410). In this embodiment, as described below, the determination method differs depending on the policy.

The batch allocation management unit 211 calculates a fee for each execution plan. Specifically, the batch allocation management unit 211 calculates the charge for one execution plan by summing the charges 1308 for each row included in one execution plan entry. The batch allocation management unit 211 determines whether there is an execution plan for which the calculated fee is lower than the fee 610 for the selected batch process. If there is an execution plan whose calculated charge is lower than the charge 610 for the selected batch process, the batch allocation management unit 211 determines that there is an execution plan that satisfies the policy.

The batch allocation management unit 211 calculates the execution time of each execution plan. Specifically, the batch allocation management unit 211 calculates the execution time of one execution plan by summing the parallelization time 1305 and the execution time 1307 of each row included in one execution plan entry. The batch assignment management unit 211 determines whether there is an execution plan whose calculated execution time is shorter than the execution time 609 of the selected batch process. If there is an execution plan whose calculated execution time is smaller than the execution time 609 of the selected batch process, the batch assignment management unit 211 determines that there is an execution plan that satisfies the policy.

When there is an additional batch process, the batch allocation management unit 211 determines whether the total execution time of the main batch process and the execution time of the additional batch process is smaller than the target time 605. In addition, the batch allocation management unit 211 refers to the budget information 224, stores the tenant identifier for which the selected batch process is executed in the tenant ID 801, and sets the date and time corresponding to the time when the selected batch process is executed as the date 802. Search for entries. The batch assignment management unit 211 acquires the budget 803 of the searched entry. The batch allocation management unit 211 determines whether or not the batch processing additional condition 1105 is satisfied based on the sum of the main batch processing fee and the additional batch processing fee, and the budget 803.

If any of the conditions is not satisfied, the batch allocation management unit 211 determines that there is no execution plan that satisfies the policy.

(3) When the policy type 1103 of the determined policy is “cost performance”, the following processing is executed.

The batch allocation management unit 211 calculates the execution efficiency of each execution plan. Specifically, the batch allocation management unit 211 calculates an average value of efficiency using the efficiency 1309 of each row included in one execution plan entry, and sets the average value as the execution efficiency of one execution plan. . The batch allocation management unit 211 determines whether there is an execution plan with the calculated execution efficiency greater than “1”. When there is an execution plan with the calculated execution efficiency greater than “1”, the batch allocation management unit 211 determines that there is an execution plan that satisfies the policy.

When there is an additional batch process, the batch allocation management unit 211 determines whether the execution efficiency of the main batch process is greater than “1”, and the execution efficiency of the additional batch process is greater than “1”. It is determined whether or not. Further, the batch allocation management unit 211 determines whether or not the batch processing additional condition 1105 is satisfied based on the total of the main batch processing fee and the additional batch processing fee and the budget 803. If any one of the conditions is not satisfied, the batch allocation management unit 211 determines that there is no execution plan that satisfies the policy.

Note that the determination methods (1) to (3) described above are merely examples and are not limited thereto. The above is the description of the process in step S1410.

If it is determined that there is an execution plan that satisfies the policy, the batch allocation management unit 211 selects one execution plan to be adopted (step S1411).

Specifically, the batch allocation management unit 211 selects an execution plan that is most effective. For example, in the case of a policy whose policy type 1103 is “cost”, the batch allocation management unit 211 selects an execution plan with the lowest charge. When the policy type 1103 is a policy having “target time”, the batch allocation management unit 211 selects an execution plan that has the shortest execution time. When the policy type 1103 is a policy having “cost performance”, the batch allocation management unit 211 selects an execution plan having the highest execution efficiency. Note that the batch allocation management unit 211 may delete the execution plan that has not been selected, or may use the execution plan that has not been selected as new knowledge.

The batch allocation management unit 211 reserves resources based on the selected execution plan (step S1412).

Specifically, the batch assignment management unit 211 secures resources so that they are not assigned to other batch processes. Since the resource reservation method is a known technique, the description thereof is omitted. At this time, the batch assignment management unit 211 sets “1” to the confirmation flag 1310 of each line of the selected execution plan.

Next, the batch assignment management unit 211 determines whether addition of batch processing is permitted (step S1413).

Specifically, the batch assignment management unit 211 refers to the policy batch processing addition condition 1105 and determines whether or not information indicating that the addition of batch processing is permitted is set.

When it is determined that the addition of the batch process is not permitted, the batch assignment management unit 211 determines whether the process has been completed for all the batch processes included in the list (step S1414).

If it is determined that the processing has not been completed for all the batch processes included in the list, the batch assignment management unit 211 returns to step S1402, selects a new batch process, and executes the same process. At this time, the batch assignment management unit 211 presents the processing result for the batch processing to the user via the display unit 213.

When it is determined that the processing has been completed for all the batch processes included in the list, the batch assignment management unit 211 ends the execution plan generation process of the batch process.

If it is determined in step S1413 that the addition of batch processing is permitted, the batch assignment management unit 211 presents a list of batch processes that can be added to the user via the display unit 213 (step S1420). Settings for additional batch processing are accepted (step S1421). Thereafter, the batch assignment management unit 211 returns to step S1403, and executes the same process with the additional batch process as the selected batch process. In this case, the loop process of the main batch process is continued.

For example, the batch allocation management unit 211 may be considered to present a batch process for which an execution plan has been generated as an addable batch process. In addition, the batch allocation management unit 211 may be presented as a batch process that can add another batch process executed in the tenant.

User sets parameters, number of instances, etc. for the presented batch process. When the batch assignment management unit 211 receives the setting, the batch assignment management unit 211 updates the batch processing management information 222 and the job management information 223. Note that “addition” is set in the additional batch processing type 603.

Further, a batch process added in advance by the user may be defined in the batch process management information 222. In this case, the batch assignment management unit 211 selects a batch process to be automatically added based on the batch process management information 222. In this case, the process of step S1422 can be omitted.

If it is determined in step S1410 that there is no execution plan that satisfies the policy, the batch assignment management unit 211 determines whether the selected batch process is an additional batch process (step S1417).

Specifically, the batch assignment management unit 211 determines whether the type 603 of the row corresponding to the selected batch process in the batch process management information 222 is “addition”. When the type 603 is “addition”, the batch assignment management unit 211 determines that the selected batch process is an additional batch process.

If it is determined that the selected batch process is not an additional batch process, the batch assignment management unit 211 reserves a resource for the batch process based on the job management information 223 (step S1418). Thereafter, the batch assignment management unit 211 proceeds to step S1414.

Specifically, the batch allocation management unit 211 deletes all execution plans for the selected batch process registered in the execution plan management information 229 and generates an execution plan for the selected batch process based on the job management information 223. That is, an execution plan for the selected batch process is generated according to the initial setting. In this case, an entry having the same content as the entry generated in step S1404 is generated in the execution plan management information 229. That is, the price fluctuation resource is not used.

When it is determined that the selected batch process is an additional batch process, the batch allocation management unit 211 deletes all execution plans for the additional batch process from the execution plan management information 229 (step S1419). Thereafter, the batch assignment management unit 211 proceeds to step S1414.

That is, the execution plan of the additional batch process that does not satisfy the policy is not generated, and the loop process of the main batch process ends. At this time, the batch assignment management unit 211 presents the processing result for the main batch process to the user via the display unit 213.

If it is determined in step S1406 that the price fluctuation resource is not available, the batch assignment management unit 211 determines whether the selected batch process is an additional batch process (step S1415). The process of step S1415 is the same as that of step S1417.

If it is determined that the selected batch process is not an additional batch process, the batch assignment management unit 211 proceeds to step S1408. That is, for the selected batch process, the execution plan change process is not executed, and the loop process of the selected batch process ends.

When it is determined that the selected batch process is an additional batch process, the batch assignment management unit 211 deletes the execution plan of the additional batch process from the execution plan management information 229 (step S1416). Thereafter, the batch assignment management unit 211 proceeds to step S1414. At this time, the batch assignment management unit 211 presents the processing result for the main batch process to the user via the display unit 213.

FIG. 15 is a flowchart illustrating an example of an execution plan change process according to the first embodiment.

The batch allocation management unit 211 determines whether or not the processing type 1202 of the change method selected in step S1404 is “scale-out” (step S1501).

If it is determined that the processing type 1202 of the selected change method is “scale-out”, the batch allocation management unit 211 executes a resource amount calculation process (scale-out) (step S1502), and then changes the execution plan. The process ends. Details of the resource amount calculation process (scale-out) will be described with reference to FIG.

If it is determined that the process type 1202 of the selected change method is “scale up”, the batch allocation management unit 211 executes a resource amount calculation process (scale up) (step S1503), and then changes the execution plan. The process ends. Details of the resource amount calculation process (scale-up) will be described with reference to FIG.

FIG. 16 is a flowchart illustrating an example of a resource amount calculation process (scale out) according to the first embodiment.

The batch assignment management unit 211 determines whether or not the parallelism of the selected job can be increased (step S1601). Specifically, the following processing is executed.

Specifically, the batch assignment management unit 211 refers to the entry of the selected job in the job management information 223 and determines whether the value of the parallel degree 702 is smaller than the value of the maximum parallel degree 703.

If the value of the degree of parallelism 702 is equal to or greater than the value of the maximum degree of parallelism 703, the batch assignment management unit 211 determines that the degree of parallelism of the selected job cannot be increased.

When the value of the degree of parallelism 702 is smaller than the value of the maximum degree of parallelism 703, the batch allocation management unit 211 calculates the additional degree of parallelism by subtracting the value of the degree of parallelism 702 from the value of the maximum degree of parallelism 703. The batch allocation management unit 211 determines whether or not the value of the additional amount 1204 of the change process selected in step S1404 is smaller than the additional parallelism.

If the value of the selected additional amount 1204 of change processing is smaller than the additional parallelism, the batch assignment management unit 211 determines that the parallelism of the selected job cannot be increased.

If the value of the added amount 1204 of the selected change process is equal to or greater than the additional parallelism, the batch assignment management unit 211 determines that the parallelism of the selected job can be increased. The above is the description of the processing in step S1601.

If it is determined that the parallelism of the selected job cannot be increased, the batch allocation management unit 211 ends the resource amount calculation process (scale-out).

When it is determined that the degree of parallelism of the selected job can be increased, the batch allocation management unit 211 calculates a resource amount necessary for increasing the degree of parallelism (step S1602).

Specifically, the batch allocation management unit 211 calculates the number of instances (resource amount) required when the degree of parallelism is increased using the selected addition amount 1204 of the change process and the degree of parallelism 702. For example, when the additional amount 1204 is “twice”, the value obtained by multiplying the value of the degree of parallelism 702 by “2” is the number of instances (resource amount) required when the degree of parallelism is increased.

Next, the batch allocation management unit 211 determines whether there is a resource amount necessary for the price fluctuation resource (step S1603).

The method for determining whether or not the necessary resource amount exists in the resource pool (price fluctuation resource) is a known method, and therefore detailed description thereof is omitted. For example, the following method can be used.

The batch allocation management unit 211 refers to the second resource amount history information 910 and determines whether or not the number of instances of a predetermined type having a resource with the required performance 1306 or higher is greater than or equal to the degree of parallelism after the change. When the above-described conditions are satisfied, the batch allocation management unit 211 determines that there is a necessary resource amount for the price fluctuation resource.

When it is determined that the necessary resource amount does not exist for the price fluctuation resource, the batch allocation management unit 211 ends the resource amount calculation process (scale-out).

When it is determined that there is a necessary resource amount for the price fluctuation resource, the batch allocation management unit 211 calculates the execution time and usage fee of the selected job after the change of the parallelism (step S1604). Further, the batch allocation management unit 211 updates the value of the row corresponding to the selected job in the execution plan management information 229 (step S1605), and then ends the resource amount calculation process (scale out). Specifically, the following processing is executed.

The batch allocation management unit 211 calculates the change ratio by dividing the parallel degree after the change by the parallel degree before the change. The batch assignment management unit 211 calculates a first evaluation value by multiplying the change ratio by a predetermined coefficient. The batch allocation management unit 211 calculates a new execution time by dividing the value of the execution time 1307 by the first evaluation value. For example, when the first evaluation value is “2”, the execution time of the selected job is half the time before the change.

Also, the batch assignment management unit 211 calculates a second evaluation value by multiplying the change ratio by a predetermined coefficient. The batch allocation management unit 211 calculates a new fee by multiplying the value of the fee 1308 by the second evaluation value. For example, when the second evaluation value is “2”, the charge for the selected job is double the charge before the change.

The batch allocation management unit 211 sets the calculated execution time and fee in the execution time 1307 and fee 1308 of the row corresponding to the selected job. The above is the description of the processing in steps S1604 and S1605.

FIG. 17 is a flowchart illustrating an example of a resource amount calculation process (scale-up) according to the first embodiment.

The batch allocation management unit 211 calculates a resource amount necessary when the required performance of the instance is changed (step S1701).

The batch allocation management unit 211 calculates the required performance of the instance based on the selected addition amount 1204 of the change process and the required performance 1306. For example, when the additional amount 1204 is “double”, the batch assignment management unit 211 multiplies each value of the required performance 1306 in the row of the selected job by “2”.

Next, the batch allocation management unit 211 determines whether there is a necessary resource amount for the price fluctuation resource (step S1702).

The batch allocation management unit 211 refers to the second resource amount history information 910 and determines whether or not the number of instances of a predetermined type having a resource equal to or higher than the calculated required performance is greater than or equal to the value of the parallelism 1304. . When the above-described conditions are satisfied, the batch allocation management unit 211 determines that there is a necessary resource amount for the price fluctuation resource.

When it is determined that the required resource amount does not exist for the price fluctuation resource, the batch allocation management unit 211 ends the resource amount calculation process (scale-up).

When it is determined that there is a necessary resource amount for the price fluctuation resource, the batch allocation management unit 211 calculates the execution time and usage fee of the selected job after changing the requested performance of the instance (step S1703). Furthermore, the batch allocation management unit 211 updates the value of the row corresponding to the selected job in the execution plan management information 229 (step S1704), and then ends the resource amount calculation process (scale-up). Specifically, the following processing is executed.

The batch allocation management unit 211 calculates the change ratio by dividing the required performance after the change by the required performance before the change. When conditions are set for a plurality of resources, the batch allocation management unit 211 calculates a change ratio for each resource, and uses an average value of the change ratios for each resource.

The batch allocation management unit 211 calculates a third evaluation value by multiplying the change ratio by a predetermined coefficient. The batch allocation management unit 211 calculates a new execution time by dividing the value of the execution time 1307 by the third evaluation value. For example, when the third evaluation value is “2”, the execution time of the selected job is half the time before the change.

Also, the batch allocation management unit 211 calculates the fourth evaluation value by multiplying the change ratio by a predetermined coefficient. The batch allocation management unit 211 calculates a new fee by multiplying the value of the fee 1308 by the fourth evaluation value. For example, when the fourth evaluation value is “2”, the charge for the selected job is double the charge before the change.

The batch allocation management unit 211 sets the calculated execution time and fee in the execution time 1307 and fee 1308 of the row corresponding to the selected job. The above is the description of the processing in steps S1703 and S1704.

FIG. 18 is a flowchart illustrating an example of the execution process of the batch process according to the first embodiment.

The batch execution unit 212 periodically refers to the start time 604 of the batch process management information 222 and identifies the target batch process (step S1801). For example, the batch execution unit 212 identifies a batch process executed within one hour from the current time as the target batch process.

Next, the batch execution unit 212 refers to the execution plan management information 229 and specifies an execution plan corresponding to the specified batch process (step S1802).

Specifically, the batch execution unit 212 searches for an entry in which the job net ID 1301 matches the identified identifier of the batch process, and the confirmation flag 1310 is set to “1”.

Next, the batch execution unit 212 deploys a job to be executed first based on the specified execution plan, and instructs the start of the job (step S1803).

Specifically, the batch execution unit 212 allocates a resource to the job to be executed first, and deploys necessary software to the resource (instance). The batch execution unit 212 instructs the job to be executed after the start time 604 has elapsed. At this time, the batch execution unit 212 sets “0” in the determination flag 1310 of the line of the job instructed to execute, and sets “1” in the execution flag 1311.

Next, the batch execution unit 212 determines whether there is a job to be executed next (step S1804).

Specifically, the batch execution unit 212 refers to the job management information 223 based on the value of the job net ID 1301. The batch execution unit 212 determines whether there is a job to be executed next to the started job based on the output job ID 707 of the job management information 223. In the following description, the job specified in step S1804 is also referred to as a target job.

When it is determined that the target job does not exist, the batch execution unit 212 ends the execution process of the batch process.

When it is determined that the target job exists, the batch execution unit 212 sets the start time of the execution plan update process for the job (step S1805). For example, a method is conceivable in which 10 minutes before the target job is started is set as the start time of the execution plan update process. The start time of the target job can be calculated using the start time of the job that is the output source of the target job and the execution time 1307 of the job.

Next, the batch execution unit 212 determines whether or not the set start time has elapsed (step S1806).

If it is determined that the set start time has not elapsed, the batch execution unit 212 returns to step S1806 and waits until the start time elapses.

If it is determined that the set start time has elapsed, the batch execution unit 212 outputs an execution instruction for execution plan update processing to the batch assignment management unit 211 (step S1807). The execution instruction includes an identifier for batch processing and an identifier for the target job.

Next, the batch execution unit 212 determines whether or not a processing completion notification has been received from the batch assignment management unit 211 (step S1808).

If it is determined that the process completion notification has not been received from the batch allocation management unit 211, the batch execution unit 212 returns to step S1808 and waits until a completion notification is received.

If it is determined that the process completion notification has been received from the batch assignment management unit 211, the batch execution unit 212 allocates resources to the target job based on the execution plan, and the target job is output after completion of the output source job. The job execution is instructed (step S1809). Thereafter, the batch execution unit 212 returns to step S1804 and executes the same processing.

19A and 19B are flowcharts for explaining an example of an execution plan update process for a job according to the first embodiment.

When the batch allocation management unit 211 receives an execution instruction for job execution plan update processing from the batch execution unit 212, the batch allocation management unit 211 starts processing described below.

The batch allocation management unit 211 predicts the state of the computer system (step S1901). The process of step S1901 is the same as the process of step S1401. Since each prediction function is generated from the history information at the time of execution of step S1901, it differs from the prediction function generated in step S1401.

Next, the batch assignment management unit 211 determines a policy for batch processing including the target job (step S1902). In the following description, batch processing including a target job is also referred to as target batch processing.

Specifically, the batch allocation management unit 211 refers to the policy management information 227, stores the identifier of the job net corresponding to the target batch process in the job net ID 1101, and stores “ON” in the selection flag 1106. Search for entries. If a batch processing policy is selected during execution of a batch processing execution plan change process, the target batch processing policy is uniquely determined.

Next, the batch assignment management unit 211 starts a loop process of the change method (step S1903). The process in step S1903 is the same as the process in step S1404.

Next, whether the batch allocation management unit 211 is in a state where the price fluctuation resource is available for the target job based on the determined policy (step S1904). The process in step S1904 is the same as the process in step S1406.

If it is determined that the price fluctuation resource is not available, the batch allocation management unit 211 proceeds to step S1906.

If it is determined that the price fluctuation resource is available, the batch allocation management unit 211 executes an execution plan change process (step S1905). The process in step S1905 is the same as the process in step S1407. By executing the execution plan change process, the resource reservation for the target job is updated. It does not change the reservation of resources for other jobs.

Next, the batch assignment management unit 211 determines whether or not processing has been completed for all the change methods (step S1906). The process in step S1906 is the same as the process in step S1409.

If it is determined that the processing has not been completed for all the change methods, the batch assignment management unit 211 returns to step S1903, selects a new change method, and executes the same processing.

When it is determined that the processing has been completed for all the change methods, the batch allocation management unit 211 determines whether there is an execution plan that satisfies the policy based on the entry group added to the execution plan management information 229. Is determined (step S1907). The process in step S1907 is the same as the process in step S1410.

When it is determined that there is no execution plan that satisfies the policy, the batch assignment management unit 211 ends the execution plan update process.

If it is determined that there is an execution plan that satisfies the policy, the batch allocation management unit 211 selects one execution plan to be adopted (step S1908). The process in step S1908 is the same as the process in step S1411.

Next, the batch assignment management unit 211 reserves the resource of the target job based on the selected execution plan (step S1909).

Specifically, the batch assignment management unit 211 secures the resource of the target job so that it is not assigned to other batch processing. Note that the reservation of resources for other jobs is not performed.

Next, the batch assignment management unit 211 determines whether or not addition of batch processing is permitted (step S1910). The process in step S1910 is the same as the process in step S1413.

If it is determined that the addition of batch processing is not permitted, the batch assignment management unit 211 ends the execution plan update processing.

If it is determined that the addition of the batch process is permitted, the batch assignment management unit 211 presents a list of batch processes that can be added to the user via the display unit 213 (step S1911), and The setting is accepted (step S1912). Note that the processing in step S1911 and step S1912 is the same as the processing in step S1420 and step S1421. The processes after step S1912 are processes related to the additional batch process.

Next, the batch assignment management unit 211 determines a policy for additional batch processing (step S1913). The process in step S1913 is the same as the process in step S1403.

Next, the batch assignment management unit 211 starts a loop process of the change method (step S1914). In addition, the batch assignment management unit 211 starts a loop process of the job of the selected batch process (step S1915). The processes in steps S1914 and S1915 are the same as the processes in steps S1404 and S1405.

Next, whether the batch allocation management unit 211 is in a state where the price fluctuation resource is available based on the determined policy (step S1916). The process in step S1916 is the same as the process in step S1406.

When it is determined that the price fluctuation resource is not available, the batch allocation management unit 211 deletes all execution plans for the additional batch processing from the execution plan management information 229 (step S1923). Thereafter, the batch assignment management unit 211 ends the execution plan update process. The process of step S1923 is the same as the process of step S1416.

When it is determined that the price fluctuation resource is available, the batch allocation management unit 211 executes an execution plan change process (step S1917). The process of step S1917 is the same as the process of step S1407.

Next, the batch assignment management unit 211 determines whether or not processing has been completed for all jobs included in the additional batch processing (step S1918). The process of step S1918 is the same as the process of step S1408.

If it is determined that processing has not been completed for all jobs included in the additional batch processing, the batch allocation management unit 211 returns to step S1915, selects a new job, and executes similar processing.

When it is determined that the processing has been completed for all jobs included in the additional batch processing, the batch assignment management unit 211 determines whether the processing has been completed for all the change methods (step S1919). . The process in step S1919 is the same as the process in step S1409.

If it is determined that the processing has not been completed for all the change methods, the batch assignment management unit 211 returns to step S1914, selects a new job, and executes the same processing.

When it is determined that the processing has been completed for all the change methods, the batch allocation management unit 211 determines whether there is an execution plan that satisfies the policy based on the entry group added to the execution plan management information 229. Is determined (step S1920). The process of step S1920 is the same as the process of step S1410.

If it is determined that there is no execution plan that satisfies the policy, the batch allocation management unit 211 deletes all execution plans for the additional batch processing from the execution plan management information 229 (step S1923). Thereafter, the batch assignment management unit 211 ends the execution plan update process.

If it is determined that there is an execution plan that satisfies the policy, the batch allocation management unit 211 selects one execution plan to be adopted (step S1921). Further, the batch allocation management unit 211 reserves resources based on the selected execution plan (step S1922). Thereafter, in the batch assignment management unit 211, the processes in steps S1921 and S1922 are the same as the processes in steps S1411 and S1412.

As described with reference to FIGS. 19A and 19B, during the execution of batch processing, the execution plan is updated so as to satisfy the policy based on the latest prediction information. As a result, finer control can be realized.

FIG. 20 is a diagram illustrating an example of a user interface displayed by the display unit 213 according to the first embodiment.

The user interface 2000 displays the processing results of the execution plan generation process and the execution plan update process.

The user interface 2000 shown in FIG. 20 is a display format for displaying the processing result of one batch process using one tab screen 2001.

The tab screen 2001 includes a start time column 2011, a target time column 2012, a scheduled completion time column 2013, and an additional batch process candidate column 2014.

The start time column 2011 is the start time of batch processing set by the user. Note that the value set in the start time column 2011 may be automatically calculated based on the execution history of similar batch processing. The target time column 2012 is a batch processing completion time set by the user. In the start time column 2011 and the target time column 2012, an allowable time gap may be input.

The scheduled completion time column 2013 is the completion time of the batch process when the resource allocation is changed by the execution plan generation process and the execution plan update process. Note that the scheduled completion time column 2013 may include items for displaying a policy type, a fee, a reserved resource amount, and the like.

The additional batch processing candidate column 2014 is a column for displaying additional batch processing candidates. The additional batch processing candidate column 2014 includes an additional batch processing display column 2020 for each policy type. The additional batch processing display field 2020 includes one or more detailed information fields 2021.

The detailed information column 2021 includes parameters, start time, and completion time set for the additional batch processing. The detailed information column 2021 includes a cancel button 2022. A cancel button 2022 is an operation button for canceling the addition of the additional batch process. When the operation button is operated, the batch assignment management unit 211 deletes the execution plan for the additional batch process from the execution plan management information 229.

Note that the detailed information column 2021 may include an operation button that permits addition of additional batch processing. In this case, when the operation button is operated, the batch assignment management unit 211 sets “1” to the execution plan confirmation flag 1310 of the additional batch process in the execution plan management information 229.

FIG. 21 is a diagram illustrating an example of the effect of the process executed by the management server 101 according to the first embodiment.

Here, explanation will be made using a graph in which the horizontal axis represents time and the vertical axis represents the resource amount of the resource. A curve 2101 is a graph showing the free resource amount of the price fluctuation resource. A rectangle 2102 with a pattern indicates one instance of a fixed price resource. A rectangle 2103 indicates one instance using a price fluctuation resource. The heights of the

rectangles

2102 and 2103 correspond to the required performance of the instance. A resource group 2104 surrounded by a thick frame indicates resources for additional batch processing.

Pattern 1 shows an image of an execution plan for batch processing that does not use price fluctuation resources. In this case, batch processing is executed using instances having the same required performance.

Pattern 2 shows an image of an execution plan for batch processing using only price fluctuation resources. Pattern 3 shows an image of an execution plan for batch processing using a fixed price resource and a price fluctuation resource. By using the price fluctuation resource, the batch processing completion time can be shortened from the conventional target time. Moreover, cost performance can be improved.

Pattern 4 shows an image of an execution plan with additional batch processing added. It is possible to improve the business by executing the additional batch process while shortening the completion time of the main batch process using the price fluctuation resource.

For example, when the main batch process is a cost calculation process, an additional batch process for calculating a more detailed cost or an additional batch process corresponding to a cost calculation process with a changed parameter is executed. The influence of setting can be analyzed from various perspectives.

As shown in FIG. 21, according to the present embodiment, it is possible to reserve resources for batch processing that satisfies a user's request using price fluctuation resources, so that it is possible to improve tenant operation efficiency and reduce operation costs. It becomes.

In addition, this invention is not limited to the above-mentioned Example, Various modifications are included. Further, for example, the above-described embodiments are described in detail for easy understanding of the present invention, and are not necessarily limited to those provided with all the described configurations. Further, a part of the configuration of each embodiment can be added to, deleted from, or replaced with another configuration.

In addition, each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. The present invention can also be realized by software program codes that implement the functions of the embodiments. In this case, a storage medium in which the program code is recorded is provided to the computer, and a CPU included in the computer reads the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the program code itself and the storage medium storing it constitute the present invention. Examples of storage media for supplying such program codes include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs (Solid State Drives), optical disks, magneto-optical disks, CD-Rs, magnetic tapes, Non-temporary storage media such as a nonvolatile memory card and ROM are used.

Further, the program code for realizing the functions described in this embodiment can be implemented by a wide range of programs or script languages such as assembler, C / C ++, Perl, Shell, PHP, Java, and the like.

Furthermore, by distributing the program code of the software that realizes the functions of the embodiments via a network, the program code is stored in a storage means such as a hard disk or memory of a computer or a storage medium such as a CD-RW or CD-R. The CPU included in the computer may read and execute the program code stored in the storage unit or the storage medium.

In the above-described embodiments, the control lines and information lines indicate those that are considered necessary for the explanation, and do not necessarily indicate all the control lines and information lines on the product. All the components may be connected to each other.

Claims

A management computer that manages a plurality of computers that provide computer resources,
The management computer has a first processor, a first memory connected to the first processor, and a first interface connected to the first processor and connected to a device,
Each of the plurality of computers has a second processor, a second memory connected to the second processor, and a second interface connected to the second processor and connected to the device. ,
On the plurality of computers, at least one business system that provides a predetermined business is constructed using computer resources of the plurality of computers,
The at least one business system executes batch processing including a plurality of jobs,
The management computer has a batch allocation management unit that manages allocation of computer resources for the batch processing, and a batch execution unit that controls execution of the batch processing,
The batch allocation management unit
Batch processing management information including information related to a first resource that is a computer resource that is assigned to the batch process and has a fixed usage fee regardless of the usage time, and is assigned to the batch process.
Job management information for managing the configuration of the plurality of jobs of the batch processing;
Policy management information for managing a computer resource allocation policy for the batch processing;
Free resource amount history information for managing the history of the free capacity of the second resource, which is a computer resource whose use fee varies depending on the use time,
Managing charge history information for managing a history of usage charges of the second resource;
The batch allocation management unit
Generating first resource amount prediction information for predicting the free capacity of the second resource based on the free resource amount history information;
Based on the fee history information, generating first fee prediction information for predicting a usage fee of the second resource,
Determining whether the second resource can be used for the batch processing based on the first resource amount prediction information, the first fee prediction information, and the policy management information;
When it is determined that the second resource can be used for the batch processing, the resource amount of the second resource required for each of the plurality of jobs is determined based on the batch processing management information and the job management information. Calculate
Generating first execution plan information for reserving a resource amount of the second resource required for each of the plurality of jobs;
The batch execution unit controls execution of the batch process based on the first execution plan information.
The management computer according to claim 1,
The batch allocation management unit
After the first execution plan information is generated, can the second resource be used for additional batch processing based on the first resource amount prediction information, the first fee prediction information, and the policy management information? Determine whether or not
When it is determined that the second resource can be used for the additional batch processing, the second resource necessary for each of the plurality of jobs of the additional batch processing is determined based on the batch processing management information and the job management information. Calculate the resource amount of 2 resources,
Generating second execution plan information for reserving a resource amount of the second resource necessary for each of the plurality of jobs of the additional batch processing;
The batch execution unit controls execution of the additional batch process based on the second execution plan information.
The management computer according to claim 2,
The batch allocation management unit manages change method definition information that defines a change method of a computer resource amount allocated to the plurality of jobs,
The batch allocation management unit
When generating the first execution plan information and when generating the second execution plan information, select the change method adapted to the batch processing and the additional batch processing based on the policy management information,
Based on the selected change method, the batch processing management information, and the job management information, a resource amount of the second resource required for each of the plurality of jobs is calculated,
Based on the first resource amount prediction information, it is determined whether there is a free resource by the calculated resource amount of the second resource,
When it is determined that there are free resources corresponding to the calculated amount of the second resource, information for reserving the second resource is set in the first execution plan information and the second execution plan information,
When it is determined that there are no free resources for the calculated amount of the second resource, information for reserving the first resource is set in the first execution plan information and the second execution plan information. Management computer characterized by
The management computer according to claim 2,
The batch allocation management unit
When a plurality of the first execution plan information are generated and a plurality of the second execution plan information are generated, one first execution plan information and one second one are generated based on the policy management information. Select execution plan information,
A management computer that reserves the second resource based on the selected first execution plan information and the selected second execution plan information.
The management computer according to claim 2,
The batch execution unit
When processing of the first job included in the batch processing is executed based on the first execution plan information, a second job to be executed after execution of the first job is specified,
Outputting an instruction to update the first execution plan information including the identification information of the second job to the batch allocation management unit;
The batch allocation management unit
Generating second resource amount prediction information for predicting the free capacity of the second resource based on the free resource amount history information;
Generating second fee prediction information for predicting a usage fee of the second resource based on the fee history information;
Determining whether the second resource can be used for the second job based on the second resource amount prediction information, the second fee prediction information, and the policy management information;
When it is determined that the second resource can be used for the second job, a resource amount of the second resource necessary for the second job is calculated based on the batch processing management information and the job management information. And
A management computer that updates the first execution plan information by reflecting the calculated resource amount of the second resource necessary for the second job in the first execution plan information.
An execution method of batch processing executed by a management computer that manages a plurality of computers that provide computer resources,
The management computer has a first processor, a first memory connected to the first processor, a first interface connected to the first processor and connected to a device,
Each of the plurality of computers has a second processor, a second memory connected to the second processor, a second interface connected to the second processor and connected to a device,
On the plurality of computers, at least one business system that provides a predetermined business is constructed using computer resources of the plurality of computers,
The at least one business system executes batch processing including a plurality of jobs,
The management computer has a batch allocation management unit that manages allocation of computer resources for the batch processing, and a batch execution unit that controls execution of the batch processing,
The batch allocation management unit
Batch processing management information including information related to a first resource that is a computer resource that is assigned to the batch process and has a fixed usage fee regardless of the usage time, and is assigned to the batch process.
Job management information for managing the configuration of the plurality of jobs of the batch processing;
Policy management information that defines the operational policy of the business system;
Free resource amount history information for managing the history of the free capacity of the second resource, which is a computer resource whose use fee varies depending on the use time,
Managing charge history information for managing a history of usage charges of the second resource;
The execution method of the batch process is as follows:
A first step in which the batch allocation management unit generates first resource amount prediction information for predicting a free capacity of the second resource based on the free resource amount history information;
A second step in which the batch allocation management unit generates first fee prediction information for predicting a usage fee of the second resource based on the fee history information;
The batch allocation management unit determines whether the second resource can be used for the batch processing based on the first resource amount prediction information, the first fee prediction information, and the policy management information. A third step;
When the batch allocation management unit determines that the second resource can be used for the batch processing, the batch allocation management unit is required for each of the plurality of jobs based on the batch processing management information and the job management information. A fourth step of calculating the resource amount of the second resource;
A fifth step in which the batch allocation management unit generates first execution plan information for reserving a resource amount of the second resource necessary for each of the plurality of jobs;
The batch execution unit includes a sixth step of controlling execution of the batch process based on the first execution plan information.
The execution method of batch processing according to claim 6,
After the first execution plan information is generated, the batch allocation management unit performs the additional batch processing on the basis of the first resource amount prediction information, the first fee prediction information, and the policy management information. A seventh step of determining whether the second resource can be used;
When the batch allocation management unit determines that the second resource can be used for the additional batch processing, the plurality of jobs of the additional batch processing based on the batch processing management information and the job management information An eighth step of calculating a resource amount of the second resource required for each of
A ninth step in which the batch allocation management unit generates second execution plan information for reserving a resource amount of the second resource necessary for each of the plurality of jobs of the additional batch processing;
The batch execution unit includes: a tenth step of controlling execution of the additional batch process based on the second execution plan information.
The execution method of batch processing according to claim 7,
The batch allocation management unit manages change method definition information that defines a change method of a computer resource amount allocated to the plurality of jobs,
The third step and the seventh step include a step in which the batch allocation management unit selects the change method adapted to the batch processing and the additional batch processing based on the policy management information,
The fourth step and the eighth step are required for each of the plurality of jobs by the batch allocation management unit based on the selected change method, the batch processing management information, and the job management information. Calculating a resource amount of the second resource,
The fifth step and the ninth step are:
The batch allocation management unit determining, based on the first resource amount prediction information, whether or not there are free resources corresponding to the calculated resource amount of the second resource;
When the batch allocation management unit determines that there are free resources for the calculated amount of the second resource, information for reserving the second resource is the first execution plan information and the second execution information. Steps to set in the plan information,
When the batch allocation management unit determines that there are no free resources by the calculated amount of the second resource, information for reserving the first resource is the first execution plan information and the second And a step for setting the execution plan information.
The execution method of batch processing according to claim 7,
The fifth step and the ninth step are:
When the batch allocation management unit generates a plurality of the first execution plan information and a plurality of the second execution plan information, one batch of the first execution plan is generated based on the policy management information. Selecting information and one piece of the second execution plan information;
The batch allocation management unit includes a step of reserving the second resource based on the selected first execution plan information and the selected second execution plan information. Execution method.
The execution method of batch processing according to claim 7,
The sixth step, the tenth step,
The batch execution unit specifying a second job to be executed after execution of the first job when processing of the first job included in the batch processing is executed based on the first execution plan information; ,
The batch execution unit, including the identification information of the second job, outputting an update instruction for the first execution plan information to the batch allocation management unit,
The execution method of the batch process is as follows:
When the batch allocation management unit receives an instruction to update the first execution plan information, second resource amount prediction information for predicting the free capacity of the second resource is based on the free resource amount history information. Generating step;
The batch allocation management unit generating second fee prediction information for predicting a usage fee of the second resource based on the fee history information;
The batch allocation management unit determines whether the second resource can be used for the second job based on the second resource amount prediction information, the second fee prediction information, and the policy management information. And steps to
If the batch allocation management unit determines that the second resource can be used for the second job, the second job necessary for the second job is based on the batch processing management information and the job management information. Calculating a resource amount of two resources;
The batch allocation management unit updating the first execution plan information by reflecting the calculated resource amount of the second resource necessary for the second job in the first execution plan information; A method for performing batch processing, comprising: