JP2015060279A - Scale control server, scale control method, and scale control program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce wasteful operation cost of a virtual server by scale-controlling a system according to a load condition.SOLUTION: In a case where a job execution request is received from a client terminal, based on an execution condition of the job which a job management database stores and a load state which a resource monitor part monitors, such number of servers as execute the job is decided, and the job is executed. Further, the execution time is stored in a process history database as an execution result, and based on the execution time of the job made to be stored by a job management part, an execution schedule stored in a schedule management database, and the time managed by a buffer, the execution schedule or the buffer is updated to be the execution time of the job.

Description

  The present invention relates to a scale control server, a scale control method, and a scale control program.

  Recently, cloud computing is prosperous. In general, cloud computing is a model that makes it possible to access sharable computing resources such as a network, a server, and a storage via a network. Scale control refers to expanding or reducing the scale of a cloud constructed by cloud computing and extending or shortening the startup time of computing resources.

  Various operators such as data center operators, telecommunications carriers, providers, etc. that provide the cloud efficiently monitor the usage status of servers assigned to the cloud and scale the cloud according to the load status. It is common to operate resources. For example, in Patent Document 1, a delay is generated from a past performance of a virtual server associated with a load, a schedule is generated, and a virtual server is started according to the schedule, thereby providing a service without delay to a customer. Technology is disclosed.

JP 2013-41397 A

  In the technique described in Patent Document 1 described above, by taking into account the delay in the startup time of the virtual server, the virtual server is started earlier by that time, and the service can be provided by the target date and time. However, in Patent Document 1, when there are a plurality of virtual servers, there is no disclosure about a specific method of how to efficiently provide services by controlling the scale according to the load status of each virtual server.

  In addition, for example, when performing cyclic processing such as monthly or weekly, the system administrator etc. estimates and schedules the execution timing of the processing and the predicted processing time, and manages the information together with the job to manage the cloud It may register on the server and execute the process. In general, the administrator or the like determines the execution timing and the predicted processing time after providing a buffer for a certain amount of time in consideration of the effect of processing in other virtual servers and the load on the network. However, depending on the load that the virtual server actually receives, the set buffer may be exceeded or the buffer may be too large. As a result, the processing cannot be executed according to the schedule, which increases the operating cost of the virtual server. Waste may occur.

  The present invention has been made in view of the above, and has a scale control server, a scale control method, and scale control capable of performing scale control of a system according to a load situation and reducing waste of operating costs of a virtual server. The purpose is to provide a program.

  In order to solve the above-described problems and achieve the object, a scale control server according to the present invention is a scale control server that controls a cloud having a plurality of servers that execute jobs, the job, A job management database that stores the number of servers that execute jobs and the job execution conditions in association with each other, and stores a job execution schedule and a buffer that indicates a margin of processing time for the job in association with each other. A storage unit for storing a schedule management database, a processing history database for storing the job execution history, a resource monitoring unit for monitoring the load status of each of the plurality of servers, and a job execution request from a client terminal When the job is received, the job management database stores the job execution condition and Based on the load status monitored by the resource monitoring unit, a job management unit that determines the number of servers to execute the job, executes the job, and stores the execution time in the processing history database as an execution result And the execution schedule stored in the job management unit, the execution schedule stored in the schedule management database, and the time managed in the buffer. And a scale control server that is updated at the execution time of the scale control server.

  Further, the present invention is configured as a scale control method and a scale control program performed by the scale control server.

  According to the present invention, it is possible to control the scale of a system in accordance with the load status and reduce the waste of operating costs of virtual servers.

It is a figure which shows the structural example of a scale control system. It is a figure which shows the structural example of schedule DB. It is a figure which shows the structural example of job management DB. It is a figure which shows the structural example of process log | history DB. It is a figure which shows the structural example of queue management DB. It is a figure which shows the structural example of calendar DB. It is a figure which shows the structural example of resource management DB. It is a flowchart which shows the process sequence of a scale control process.

  Exemplary embodiments of a scale control server, a scale control method, and a scale control program according to the present invention will be described below in detail with reference to the accompanying drawings.

  FIG. 1 is a diagram showing a configuration example of a scale control system 1000 that is an embodiment of a scale control server, a scale control method, and a scale control program according to the present invention. As shown in FIG. 1, the scale control system 1000 includes a scale control server 100, a cloud 200, and a client terminal 300. Each of these devices is connected to each other by networks N1 and N2 such as a general public line network and a LAN (Local Area Network). First, the scale control server 100 will be described.

  As shown in FIG. 1, the scale control server 100 includes a storage unit 101, a scale control unit 102, a resource control unit 103, and a job management unit 104. The storage unit 101 includes a storage device such as an HDD (Hard Disk Drive), and includes a schedule DB 1011, a job management DB 1012, a processing history DB 1013, a queue management DB 1014, a calendar DB 1015, and a resource management DB 1016. doing.

  The schedule DB 1011 is a database that manages the schedule of jobs executed by the virtual server 210 that constitutes the cloud 200. FIG. 2 is a diagram illustrating a configuration example of the schedule DB 1011. As shown in FIG. 2, the schedule DB 1011 includes a job ID for identifying a job, a job name, a cycle for executing the job, a start time and an end time of the job, a processing delay of the job, and the like. Are stored in association with a buffer indicating a surplus of processing time by a subsequent job ID for identifying a subsequent job following the job. In FIG. 2, for example, the job “date and time sales summary” identified by the job ID “D00001” is a daily process, and the schedule is set to start at 22:00 and end at 23:30. Is shown. Further, it is indicated that there is a job “S00001” as the subsequent job. Furthermore, 30 minutes is set as a buffer for processing time, and if it is a normal execution condition (described later), the job will be finished in 60 minutes (22: 0 to 23:00). It can be seen that the time and end time are set to the above-mentioned times. The schedule DB 1011 is set in advance by a system administrator or the like.

  As will be described later, in this system, when a job scheduled in advance with the processing time registered in the schedule DB 1011 is frequently terminated due to the load on the virtual server 210 or other jobs, By automatically updating the schedule of the job so that the end time and the time managed by the buffer are shortened, the usage time of the virtual server 210 by the job is reduced. A user who uses the cloud 200 is often in a usage form in which a fee is charged according to the usage time of the virtual server 210 by a scheduled job, but it is possible to suppress the burden of operation costs due to such charging. Further, when the number of virtual servers 210 decreases as viewed from the system side, the surplus CPUs of the virtual servers 210 can be used effectively.

  On the other hand, if a job scheduled in advance with the processing time registered in the schedule DB 1011 is frequently terminated due to the load on the virtual server 210 or other jobs, the above-described case may occur. By automatically updating the schedule of the job so that the end time and the time managed by the buffer become longer, the usage time of the virtual server 210 by the job is increased. An administrator or the like who manages the cloud 200 often charges a fee according to the usage time of the virtual server 210 by a scheduled job, but the usage status of the virtual server 210 for such charging is investigated and a schedule is set. The burden of renewal costs due to changes or the like can be suppressed. Further, when the number of virtual servers 210 increases as viewed from the system side, the load on the virtual servers 210 can be reduced and the remaining CPUs of the virtual servers 210 can be effectively utilized.

  The job management DB 1012 is a database that manages a range in which the cloud 200 is scale-controlled during the above-described job execution. FIG. 3 is a diagram illustrating a configuration example of the job management DB 1012. As shown in FIG. 3, the job management DB 1012 indicates the above-described job ID, job name, and cycle, the above-described buffer, the priority for executing the job, and the maximum number of virtual servers 210 that execute the job. The number of servers and the execution conditions indicating the conditions of the virtual server 210 when the job is executed are stored in association with each other. In FIG. 3, for example, a job identified by the job ID “D000001” is executed using a maximum of three virtual servers 210. In addition, as an execution condition, the job is executed using a virtual server 210 having a CPU usage rate of 70% or less. Furthermore, the job is executed with priority over other jobs. The job management DB 1012 is set in advance by a system administrator or the like.

  The processing history DB 1013 is a database that manages job processing results in a history format. FIG. 4 is a diagram illustrating a configuration example of the processing history DB 1013. As shown in FIG. 4, the processing history DB 1013 includes the job ID, the actual start time indicating the actual processing start time of the job, the actual end time indicating the actual processing end time of the job, and the job. Is stored in association with the processing server that indicates the virtual server 210 that executed. Further, these pieces of information are stored cyclically in one record (in the example shown in FIG. 4, they are sequentially added to the right side of one record). In FIG. 4, for example, it is indicated that the time when the job identified by the job ID “D000001” was most recently started is “22:00” and the end time is “23:10”. Yes. Further, it is indicated that there are three virtual servers 210 “S00001, S00005, S00006” that have executed the job at this time. Further, in the subsequent cycle, the job started at “22:05”, ended at “23:40”, and executed by the three virtual servers 210 of “S00001, S00002, and S00003”. . The processing history DB 1013 is updated when the job execution unit 105 actually executes a job.

  The queue management DB 1014 is a database that manages job queuing. FIG. 5 is a diagram illustrating a configuration example of the queue management DB 1014. As shown in FIG. 5, the queue management DB 1014 stores a queue ID for identifying a queue and a waiting job ID indicating a job queued by the queue in association with each other. In FIG. 5, for example, the job queued in the first order is “D000001”. The queue management DB 1014 adds the job to the waiting job ID when the job management unit 105 receives a job execution request from the client terminal 300 or when a scheduled time arrives.

  The calendar DB 1015 is a database that manages calendar information such as general date, calendar, and company-specific information. FIG. 6 is a diagram illustrating a configuration example of the calendar DB 1015. As shown in FIG. 6, the calendar DB 1015 stores date, day of the week, and company-specific information (for example, sale schedules such as special sales days) in association with each other as general calendar information. In FIG. 6, for example, in addition to the calendar information of March 2013, the shaded days such as 1st, 5th, 10th, etc. are special sales days. The calendar DB 1015 is set in advance by a system administrator or the like.

  The resource management DB 1016 is a database that manages the operating status of the virtual server 210. FIG. 7 is a diagram illustrating a configuration example of the resource management DB 1016. As shown in FIG. 7, the resource management DB 1016 stores the server ID for identifying the virtual server 210, the CPU usage rate of the virtual server 210, the memory usage amount, and the disk usage amount in association with each other. Yes. In FIG. 7, for example, the CPU usage rate, the memory usage amount, and the disk usage amount of the virtual server 210 identified by the server ID “S00001” are 30%, 50%, and 50%, respectively. The numerical values of the CPU usage rate, the memory usage amount, and the disk usage amount of the resource management DB 1016 are periodically updated by the resource monitoring unit 104 (for example, n times every 5 minutes in the example shown in FIG. 7). Next, returning to FIG. 1, the scale control unit 102 will be described.

  The scale control unit 102 calculates a standard processing time (standard processing time) from the start time and end time of the job registered in the schedule DB 1011 illustrated in FIG. 2, and is stored in the processing history DB 1013 illustrated in FIG. The actual processing time (actual processing time) is calculated for each cycle from the start actual time and end actual time. When the difference between the standard processing time and the average actual processing time is longer than the time managed by the buffer, the scale control unit 102 determines that the actual processing time is subtracted from the standard processing time. If there is a margin after the time managed by the buffer, the buffer of the schedule DB 1011 is updated with the difference time, and the end time of the schedule DB 1011 is updated with the average end result time. By updating the end time and buffer of the schedule DB 1011 by the scale control unit 102, the operation cost and the update cost are suppressed for the user and the administrator who use the cloud 200 as described above, and the CPU of the virtual server Can be used effectively.

  In the following description, a case where the actual processing time has a margin even when the time managed by the buffer is subtracted from the standard processing time is described. On the contrary, the scale control unit 102 calculates When the difference between the standard processing time and the average actual processing time exceeds the time managed by the buffer (ie, subtracting the actual processing time from the standard processing time will use up the time managed by the buffer and make it negative) As in the case described above, the buffer of the schedule DB 1011 can be updated with the difference time, and the end time of the schedule DB 1011 can be updated with the average end result time. By updating the end time and buffer of the schedule DB 1011 by the scale control unit 102, the operation cost and the update cost are suppressed for the user and the administrator who use the cloud 200 as described above, and the CPU of the virtual server Can be used effectively.

  The resource monitoring unit 104 monitors the operating status of each virtual server 210 and periodically updates the CPU usage rate, the memory usage amount, and the disk usage amount of the resource management DB 1016 illustrated in FIG. 7 (for example, every 5 minutes). To do. Although the example shown in FIG. 7 shows the case where each value is updated, the resource monitoring unit 104 uses the information such as the minute, hour, day, month, and year cycles (for example, every five minutes). N times) may be stored in the resource management DB 1016 in a history format. In this case, the resource monitoring unit 104 predicts the future operating status from the operating status of the virtual server 210 in the same period in the past, and outputs the result to the scale control unit 103. The scale control unit 103 determines the number of servers in the schedule DB 1011. Or the execution condition may be updated. For example, when the resource monitoring unit 104 accumulates the CPU usage rate, the memory usage amount, and the disk usage amount monitored every 5 minutes for one year, the time zone when the job registered as monthly processing is executed In the case where the CPU usage rate of each virtual server 210 is 60% on average in a certain month and 30% on average in the following month or the next month, the time period in which the job is executed in the future Then, it is determined that the CPU usage rate of the virtual server 210 is 30% on the average, and the result is output to the scale control unit 103. When the scale control unit 103 receives the result from the resource monitoring unit 104, the scale control unit 103 updates the execution condition of the schedule DB 1011 from 60% to 30%, or aggregates the number of servers used in the job and sets the number of servers to eight. It is good also as updating to 4 sets.

  The job management unit 105 queues a job requested to be executed from the client terminal 300 or a job with a scheduled time and causes the virtual server 210 to execute the job. The job management unit 105 causes the virtual server 210 to execute the queued jobs queued in the queue management DB 1014 illustrated in FIG. 5 in ascending order of queue ID (that is, in the order queued earlier). When the job management unit 105 causes the virtual server 210 to execute a job, the job management unit 105 refers to the CPU usage rate of the resource management DB 1016 illustrated in FIG. 7 and the execution condition of the job management DB 1012 illustrated in FIG. The processing is performed using the virtual server 210 that satisfies the execution condition and the number of servers. At this time, the job management unit 105 writes the start time and end time of the job in the processing history DB 1013 shown in FIG.

  If there are more virtual servers 210 that satisfy the job execution condition than the number of servers, the job management unit 105 identifies the virtual server with the lowest CPU usage rate, and then the CPU next to the virtual server 210 is the CPU. Jobs are executed using the virtual servers 210 corresponding to the number of servers in order of increasing CPU usage rate, such as specifying virtual servers with low usage rates. Each time a job is newly executed, the job management unit 105 accesses the resource management DB 1016 to check the CPU usage rate of the virtual server 210 and identifies a virtual server that satisfies the execution condition. In the following description, the virtual server 210 to be used is specified at the timing when a new job is executed. However, the CPU usage rate of the virtual server 210 is checked at the timing when the job is completed, and the virtual server 210 that executes the next job is executed. The server may be specified. In this way, the job management unit 105 causes the virtual server 210 with a low CPU usage rate to execute a job, so that the load balance of the entire cloud 200 can be leveled.

  Further, the job management unit 105 refers to the processing history DB 1013 using the job ID as a key, and in which virtual server 210 (processing server) the other job was executed at the timing when the job was executed in the past. It is also possible to cause the virtual server 210 that has been checked and executed the least job to process the job. Also in this case, the job management unit 105 allocates the job to be executed to the virtual server 210 after grasping the processing tendency such as how many jobs have been executed in the past in each virtual server 210. The load balance of the entire 200 can be leveled.

  Further, the job management unit 105 refers to the queue management DB 1014 using the queue ID as a key, counts the number of jobs waiting for the queue at that time, and further stores the processing history DB 1013 using the waiting job ID corresponding to the queue ID as a key. Refer to and check the past processing time of the job. Then, the job management unit 105 predicts the load on the virtual server 210 from the past processing time of the checked job, and if the virtual server 210 processes the job with the previous queue ID, These jobs may be processed by another virtual server 210. Regarding the load prediction, for example, the job management unit 105 determines that the load applied to the virtual server 210 is larger as the past processing time of the job is longer or the number of virtual servers 210 is larger. Servers other than the virtual server 210 when the processing time exceeds a certain threshold (the processing time exceeds the specified time, the time exceeding the time managed by the buffer exceeds the specified time, the number of virtual servers 210 exceeds the specified number, etc.) Causes the next queued job to be processed. Also in this case, the job management unit 105 allocates jobs to be executed in consideration of the processing time of past jobs to the virtual server 210, so that the load balance of the entire cloud 200 can be leveled. Next, returning to FIG. 1, the cloud 200 will be described.

  The cloud 200 is a collection of sharable computing resources such as a network, a server, and a storage. As illustrated in FIG. 1, the cloud 200 includes a plurality of virtual servers 210. The virtual server 210 includes a control unit 2101 and a storage unit 2102. The control unit 2101 is configured by a calculation device such as a CPU, and executes a job in response to a request from the scale control server 100, for example. The storage unit 2101 includes a storage medium such as a memory and an HDD (Hard Disk Drive), and temporarily stores various data used for a job or stores the processing results.

  The client terminal 300 is, for example, a general desktop or notebook computer. A user of the system operates the client terminal 300 to execute a desired job. As illustrated in FIG. 1, the client terminal 300 includes a control unit 301 and a display unit 302. The control unit 301 is composed of an arithmetic device such as a CPU, and acquires job execution requests and results. The display unit 302 includes a display device such as an LCD, and displays a job execution instruction screen and a result screen (both not shown).

  Although not specifically described in the present embodiment, for example, the scale control unit 102 refers to the calendar DB 1015 when updating the end time or buffer of the schedule DB 1011, for example, the weekend, the beginning of the month, the end of the month, If the job is scheduled to be executed on a sale day such as a special sale day, it is determined that the processing time will be longer than usual due to the processing of many jobs, and a spare time will be added at the end time. May be set as the end time of the schedule DB 1011 or a buffer obtained by adding a spare buffer to the buffer may be set in the buffer of the schedule DB 1011. In this case, it is possible to set a schedule that matches the situation in which the job is executed.

  In addition, there are cases where queued jobs have a job to be processed with priority over other jobs. In such a case, the job management unit 105 accesses the job management DB 1012 using the job ID (waiting job ID) of the job queued in the queue management DB 1014 as a key, and refers to the priority corresponding to the job. Then, the job management unit 105 compares the referenced priority with the priorities of the queued jobs before and after, and if it is determined that the referenced job has a higher priority, the job is set as a comparison target. It is also possible to queue the job before the job. In this case, the job can be executed in consideration of the priority. Next, scale control processing performed in this system will be described.

  FIG. 8 is a flowchart showing the processing procedure of the scale control processing. As shown in FIG. 8, in the scale control process, first, the job management unit 105 determines whether or not a job execution request has been received from the client terminal 300, or whether or not a job execution schedule has arrived (step S1). S801) When it is determined that there is no execution request or the schedule for executing the job has not arrived (step S801; No), the process waits until these timings arrive.

  On the other hand, when the job management unit 105 determines that the execution request has been made or the schedule for executing the job has arrived (step S801; Yes), the job is queued in the queue management DB 1014 (step S802). It is determined whether or not the job execution order has been reached (step S803). At this time, if the job management unit 105 determines that the job execution order has not been reached (step S803; No), the job management unit 105 stands by until the order arrives.

  On the other hand, if the job management unit 105 determines that the job execution order has been reached (step S803; Yes), the job management unit 105 refers to the CPU usage rate of the resource management DB 1016 and the execution conditions of the job management DB 1012 to determine the job to be executed. A virtual server 210 that satisfies the execution condition and the number of servers is determined (step S804), and the job processing is executed using the determined virtual server 210 (step S805). Then, the job management unit 105 writes the job start time and end time in the processing history DB 1013 as execution results (step S806).

  The scale control unit 102 calculates the standard processing time of the job registered in the schedule DB 1011 and the actual processing time stored in the processing history DB 1013, and checks (confirms) the current buffer, the number of virtual servers, and the execution conditions. Then, it is determined whether or not these contents are appropriate with the current contents (step S808). Specifically, the scale control unit 102 determines whether or not the difference between the standard processing time and the average actual processing time is longer than the time managed by the buffer.

  If the scale control unit 102 determines that these contents are appropriate for the current contents (step S808; Yes), the process is terminated without doing anything. On the other hand, when the scale control unit 102 determines that these contents are not appropriate with the current contents (step S808; No), that is, the time obtained by subtracting the actual processing time from the standard processing time is equal to or longer than the time managed by the buffer. If there is a margin, the buffer of the job management DB 1011 and the schedule DB 1011 is updated with the difference time, the end time of the schedule DB 1011 is updated with the average end result time, and the result screen (for example, FIG. 2 and the job management DB shown in FIG. 3) are displayed on the display unit 302 of the client terminal 300 (step S809).

  Here, only the processing end time is described, but as described above, the average value of the actual use values of each value of the resource such as the CPU usage rate, the memory usage amount, and the disk usage amount is calculated and the value is calculated. When the current execution condition falls below a certain level (for example, 60%, which is 10% lower), the current execution condition may be updated to that value. When the process of step S809 ends, the scale control process shown in FIG. 8 ends.

  As described above, by performing the scale control processing in the present embodiment, it is possible to control the scale of the system according to the load situation and reduce the waste of operating costs of the virtual server.

  The processing executed by the scale control server of the present embodiment is actually executed by a program. For example, a file in an installable format or an executable format is a CD-ROM, a flexible disk (FD), a CD- It can also be provided by being recorded on a computer-readable recording medium such as R or DVD (Digital Versatile Disk). The program may be stored on a computer connected to a network such as the Internet and downloaded via the network, or provided or distributed via a network such as the Internet. Further, the program may be provided by being incorporated in advance in a ROM or the like.

  The program executed by the scale control server of the present embodiment has a module configuration including the above-described units (scale control unit 102, resource monitoring unit 103, job management unit 104, etc.), and actual hardware As the CPU reads out the program from the storage medium and executes it, the above units are loaded onto a main storage device (not shown), and the scale control unit 102, the resource monitoring unit 103, the job management unit 104, etc. It is supposed to be generated above.

  The present invention is not limited to the above-described embodiments, and includes various modifications. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment. For example, when the number of jobs is large, the function of the job management unit 105 is provided in one server, and the scale is Other functions such as the control unit 102 and the resource monitoring unit 103 may be provided in another server to distribute the load.

  In addition, when the scale control unit 102 checks the number of virtual servers and execution conditions to determine whether the contents are appropriate with the current contents, the time obtained by subtracting the actual processing time from the standard processing time is managed by the buffer. Although it is described as more than the time, whether the time obtained by subtracting the actual processing time from the standard processing time is longer than the time managed by the buffer, or shorter (or longer) than the time managed by the buffer It is also possible to set as a reference time and make a determination by comparing the time with the time obtained by subtracting the actual processing time from the standard processing time. That is, the scale control unit 102 does not necessarily check the number of virtual servers and execution conditions by using the buffers shown in the present embodiment to determine whether or not the contents are appropriate for the current contents. It can be judged by various criteria.

1000 Scale control system 100 Scale control server 101 Storage unit 1011 Schedule DB
1012 Job management DB
1013 Processing history DB
1014 Queue management DB
1015 Calendar DB
1016 Resource management DB
102 scale control unit 103 resource control unit 104 job management unit 200 cloud 210 virtual server 2101 control unit 2102 storage unit 300 client terminal 301 control unit 302 display unit N1, N2 network.

Claims (11)

A scale control server that scales a cloud having a plurality of servers that execute jobs,
A job management database that stores the job, the number of servers that execute the job, and the execution condition of the job in association with each other; and a buffer that indicates an allowance for the job execution schedule and the job processing time. A storage unit that stores a schedule management database that stores the job in association with each other, a processing history database that stores an execution history of the job, and
A resource monitoring unit for monitoring the load status of each of the plurality of servers;
When a job execution request is received from a client terminal, the job management database stores the job execution conditions stored in the job management database and the load status monitored by the resource monitoring unit. A job management unit that determines a server, executes the job, and stores the execution time in the processing history database as an execution result;
Based on the execution time of the job stored by the job management unit, the execution schedule stored in the schedule management database, and the time managed by the buffer, the execution schedule or the buffer is executed for the job. A scale control that updates in time,
A scale control server comprising: The scale control unit determines whether or not a difference between the execution time of the job and the time managed by the execution schedule and the buffer is greater than or equal to a certain value, and the execution time of the job, the execution schedule, and the buffer Updating the execution schedule or the buffer to the execution time of the job when it is determined that the difference from the managed time is a certain value or more;
The scale control server according to claim 1. The scale control unit reduces the number of servers stored in the job management database, or executes the job when it determines that the execution time of the job is less than or equal to the execution schedule and the buffer. When it is determined that the time is more than a certain amount than the execution schedule and the buffer, the number of servers stored in the job management database is increased.
The scale control server according to claim 1 or 2, characterized by the above. The job management database stores a CPU usage rate as the execution condition,
The scale control unit lowers the CPU usage rate when the execution time of the job is determined to be a certain amount or less than the execution schedule and the buffer, or the execution time of the job is the execution schedule and the buffer. If the CPU usage rate is determined to be more than a certain level, the CPU usage rate is increased.
The scale control server according to any one of claims 1 to 3. When there are more servers than the number of servers satisfying the CPU usage rate, the job management unit causes the job to be executed using the servers for the number of servers in order of increasing CPU usage rate.
The scale control server according to claim 4. The storage unit further stores a queue management database for queuing the job,
The job management database further stores a priority for executing the job,
When the job management unit 105 receives a job execution request from a client terminal, the job management unit 105 queues the job in the queue management database, and the priority of the queued job is higher than that of the job queued first. If it is higher, let the job run first,
The scale control server according to any one of claims 1 to 5, wherein the server is a scale control server. The storage unit further stores a calendar database for storing calendars,
The scale control unit refers to the calendar database when updating the execution schedule and the buffer, and when the job is executed on a specific day including a weekend, the beginning of the month, and the end of the month, many jobs are processed. The processing time is longer than usual, and further reserve time is added to the buffer.
The scale control server according to claim 1, wherein: A scale control method that scales a cloud having a plurality of servers that execute jobs,
A first step of monitoring a load status of each of the plurality of servers;
When a job execution request is received from a client terminal, the job is stored in a job management database that stores the job, the number of servers that execute the job, and the job execution conditions in association with each other. A second step of determining the number of servers to execute the job based on the condition and the monitored load status and executing the job;
A third step of storing the execution time in the processing history database for storing the execution history of the job as an execution result;
The execution based on the execution schedule and the buffer stored in the schedule management database that stores the job execution time, the job execution schedule, and a buffer indicating a margin of the processing time of the job in association with each other A fourth step of updating the schedule and the buffer to the execution time of the job;
A scale control method comprising: In the fourth step, the method further includes a step of determining whether or not a difference between the execution time of the job and the execution schedule and the buffer is greater than or equal to a certain value, and the execution time of the job, the execution schedule, and the buffer; The execution schedule and the buffer are updated to the execution time of the job when it is determined that the difference between the two is greater than or equal to a certain value.
The scale control method according to claim 8. On the computer,
A first step of monitoring a load status of each of a plurality of servers that are included in a cloud to be scaled and execute jobs;
When a job execution request is received from a client terminal, the job is stored in a job management database that stores the job, the number of servers that execute the job, and the job execution conditions in association with each other. A second step of determining the number of servers to execute the job based on the condition and the monitored load status and executing the job;
A third step of storing the execution time in the processing history database for storing the execution history of the job as an execution result;
The execution based on the execution schedule and the buffer stored in the schedule management database that stores the job execution time, the job execution schedule, and a buffer indicating a margin of the processing time of the job in association with each other A fourth step of updating the schedule and the buffer to the execution time of the job;
A scale control program characterized in that In the fourth step, the method further includes a step of determining whether or not a difference between the execution time of the job and the execution schedule and the buffer is greater than or equal to a certain value, and the execution time of the job, the execution schedule, and the buffer; The execution schedule and the buffer are updated to the execution time of the job when it is determined that the difference between the two is greater than or equal to a certain value.
A scale control program characterized by that.

Images