WO2017013872A1 - Resource allocation system, resource allocation method, and computer-readable recording medium - Google Patents

Resource allocation system, resource allocation method, and computer-readable recording medium Download PDF

Info

Publication number
WO2017013872A1
WO2017013872A1 PCT/JP2016/003388 JP2016003388W WO2017013872A1 WO 2017013872 A1 WO2017013872 A1 WO 2017013872A1 JP 2016003388 W JP2016003388 W JP 2016003388W WO 2017013872 A1 WO2017013872 A1 WO 2017013872A1
Authority
WO
WIPO (PCT)
Prior art keywords
load
resource allocation
determination unit
urgency
server
Prior art date
Application number
PCT/JP2016/003388
Other languages
French (fr)
Japanese (ja)
Inventor
晴道 横山
智義 菅原
才田　好則
Original Assignee
日本電気株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 日本電気株式会社 filed Critical 日本電気株式会社
Priority to JP2017529458A priority Critical patent/JPWO2017013872A1/en
Publication of WO2017013872A1 publication Critical patent/WO2017013872A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]

Definitions

  • the present invention relates to a resource allocation system, a resource allocation method, and a computer-readable recording medium.
  • Patent Document 1 describes a technique for making computer resources follow a load of an application in a short time.
  • the resource allocation control server described in Patent Literature 1 uses the collected load information and the load information collected in the past to calculate an increase / decrease amount of the load when the application is executed. Then, the resource allocation control server described in Patent Literature 1 determines the resource amount to be allocated to the application based on the calculated increase / decrease amount.
  • Patent Document 2 describes a technology that appropriately determines the type of a resource that is insufficient according to the operating status of the system, and enables the addition of the resource.
  • the load state of the operating system is measured by the operation status monitoring unit, and the additional resource determination unit determines whether or not hardware resources need to be added based on the resource addition policy. Is determined.
  • the server addition unit starts operation using the spare server.
  • Patent Document 3 describes a technique for preventing frequent reassignment of resources by reacting sensitively to load fluctuations.
  • the computer system described in Patent Literature 3 calculates the number of resources required for the service for each of the plurality of resource number calculation formulas according to the plurality of resource number calculation formulas defined by the information held in the resource number calculation formula table. .
  • the effective resource number determination mechanism determines the effective resource number selected from the effective resource number determination expression table by one of the required resource numbers calculated for each of the plurality of resource number calculation expressions from the effective resource number determination expression selection mechanism. The number of effective resources is determined according to the formula.
  • the allocated resource number adjustment mechanism adjusts the number of resources allocated to the corresponding service based on the determined effective resource number.
  • the effective resource number determination formula is selected by the user.
  • Patent Document 4 describes a technique for effectively managing a network and controlling it stably while maintaining measurement of load information of resources at a configuration, and at the same time reducing measurement overhead.
  • the measurement interval is adjusted based on the measured load information obtained by measuring the resource load information at the measurement interval and the predicted load information obtained by predicting the resource load information according to the prediction algorithm. Is configured to do.
  • JP 2014-164715 A Japanese Patent Laid-Open No. 2006-11860 JP 2005-173928 A JP 2005-86803 A
  • resource allocation is controlled based on a single threshold value related to the load on the computer system.
  • a threshold value for example, by setting a threshold value so that resources are allocated to a relatively small load, it is possible to quickly cope with a sudden load fluctuation in the computer system.
  • resource allocation that is not necessarily required may be performed in accordance with load fluctuations in a range close to the threshold value.
  • the number of resources allocated to the corresponding service is adjusted based on, for example, a resource number calculation formula selected by the user. For this reason, for example, when a spike phenomenon occurs when the resource count calculation formula is selected based on the assumption that the spike phenomenon, which is a sudden load increase, does not occur, quick resource allocation is performed. There is no possibility.
  • the present invention has been made in order to solve the above-mentioned problems, and has as its main object to provide a resource allocation system and the like that enable appropriate resource allocation according to load fluctuations.
  • a resource allocation system is a load acquisition unit that acquires information regarding a load of a server that executes a task, and a need for allocation of resources to a task based on an urgency level that is obtained using the information regarding the load
  • the urgency determination means for determining the load, the load determination means for determining the load trend of the server when the urgency determination means determines that it is necessary to check the load trend of the server, and the urgency determination means Resource allocation means for allocating a resource to a task when it is determined that a resource needs to be allocated to the task.
  • the resource allocation method acquires information related to a load of a server that executes a task, and determines the necessity for allocation of a resource to a task based on an urgency level obtained using the information related to the load. However, if it is determined that it is necessary to check the load trend of the server, it is necessary to determine the load trend of the server and, as necessary, the urgency determination means must allocate resources to the task. When it is determined that the resource is assigned to the task.
  • each component of each device represents a functional unit block.
  • Each component of each device can be realized by any combination of an information processing device 1000 and software as shown in FIG. 28, for example.
  • the information processing apparatus 1000 includes the following configuration as an example.
  • the task execution server 11 is realized by the information processing apparatus 1000 described above, for example.
  • an OS Operating System
  • OS etc. a virtual machine managed by a hypervisor operate.
  • an OS or the like that runs on the task execution server 11 an arbitrary task is executed in response to a request from the user.
  • the resources such as the task execution server 11 are assigned to each task within a predetermined range.
  • Examples of resources allocated to each task include a processor core and a memory of the information processing apparatus 1000 that implements the task execution server 11.
  • the resource allocated to each task may be a resource provided in another information processing apparatus connected to the task execution server 11 via the communication network 12. Further, the resource allocated to the task is changed according to a setting made in advance for the task execution server or the like.
  • the resource allocation system 100 may set a task executed in each of the task execution servers 11-1 to 11-n as a resource allocation target. That is, the resource allocation system 100 allocates resources to tasks executed by a plurality of task execution servers from the task execution server 11-1 to the task execution server 11-n.
  • the resource allocation system 100 acquires the CPU usage rate of a server on which a task is executed as a server load and allocates a processor core as a resource to the task. To do.
  • the information regarding the load acquired by the load acquisition unit 110 is notified to the urgency level determination unit 120 described later. Further, the load acquisition unit 110 may store information related to the load in the measurement value storage unit 150. Details of the information stored in the measured value storage unit 150 will be described later.
  • the urgency determination unit 120 determines whether it is necessary to allocate a resource task included in the server based on the server load acquired by the load acquisition unit 110. Specifically, the urgency level determination unit 120 obtains the urgency level using the server load acquired by the load acquisition unit 110. Then, based on the urgency level, it is determined whether it is necessary to allocate the resource task of the server. Note that the urgency determining unit 120 may determine whether it is necessary to allocate a resource task included in the server by other methods.
  • the urgency level indicates the degree of necessity related to allocation of resources provided to the server for tasks.
  • the urgency is obtained based on, for example, the load on the server and the load trend.
  • the urgency level determination unit 120 may read and obtain the information stored in the measurement value storage unit 150 as necessary.
  • the urgency is obtained based on the load on the server at the time of obtaining the urgency and the load trend at the time.
  • a and b are appropriately determined according to which one of the load of the server at the time of obtaining the degree of urgency and the load tendency at the time is important.
  • the load tendency is obtained by using the load measured at the time immediately before the time when the urgency is obtained and the load measured immediately before the time.
  • the load trend may be a load measured at a time two or more times before the most recent time when the urgency level is obtained.
  • the urgency level determination unit 120 may hold in advance the load measured at a time point before the most recent time point described above, and read it as necessary when determining the urgency level. Or the urgency determination part 120 may hold
  • the urgency level determination unit 120 determines whether resources need to be allocated to tasks based on the urgency level.
  • the urgency determination unit 120 determines whether the urgency satisfies the first and second conditions, and determines the necessity of resource allocation for the task based on the two criteria described above.
  • the first condition is a condition indicating that there is a high necessity for resource allocation and that a resource should be allocated to a task.
  • the second condition is a condition that indicates that it is necessary to further check the tendency of the load on the server when the resource allocation is high but the resource should not be allocated to the task. .
  • the first and second conditions are expressed as threshold values for the value obtained as the degree of urgency.
  • E1 is set as the threshold value related to the first condition
  • E2 is set as the threshold value related to the second condition.
  • the urgency level determination unit 120 determines that the urgency level satisfies the first condition when the urgency level satisfies E1 ⁇ E (that is, the level of the urgency level E is equal to or greater than E1). Further, the urgency level determination unit 120 determines that the urgency level is the second level when the urgency level E satisfies E2 ⁇ E ⁇ E1 (that is, the magnitude of the urgency level E is equal to or larger than E2 and smaller than E1). It is determined that the above condition is satisfied.
  • E ⁇ E2 that is, the urgency level is smaller than E2
  • the urgency level determination unit 120 determines that the urgency level does not satisfy both the first and second conditions. To do.
  • the load determination unit 130 may refer to information on the load stored in the measured value storage unit 150 as necessary when determining the load tendency.
  • the load determination unit 130 may determine a load tendency using information regarding loads at a plurality of points in time.
  • the load determination unit 130 refers to the information on the load stored in the measurement value storage unit 150, for example, and has already been at a time before the time when the urgency is determined to satisfy the second condition. Information regarding the measured load may be used.
  • the load determination part 130 may use the information regarding the load in the time after the said time.
  • the load determination unit 130 may determine to allocate resources to the task.
  • the resource allocation unit 140 performs an operation of allocating a resource as will be described later.
  • the load determination may be terminated according to any predetermined condition or the like, or the determination of the load trend may continue. May be continued.
  • the resource allocation unit 140 performs an operation of allocating a resource to a task when the urgency level determination unit 120 determines that the necessity for resource allocation is high. That is, the resource allocation unit 140 performs an operation of allocating resources to a task when the urgency satisfies the first condition described above.
  • the resource allocation unit 140 requests resource allocation to a server that executes a task such as the task execution server 11 shown in FIG.
  • the resource allocation unit 140 also performs an operation of allocating resources to tasks even when the load determination unit determines that the load tends to increase.
  • the resource to be allocated includes, for example, one or more processor cores.
  • the resource allocation unit 140 may allocate an arbitrary resource included in the task execution server 11 or the like.
  • the resource allocation unit 140 may hold in advance information on resources that can be allocated on a server on which a task is executed, such as the task execution server 11.
  • the urgency level determination unit 120 determines whether the urgency level satisfies one of the first or second conditions described above as the determination of the necessity regarding the resource allocation to the task. .
  • the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S101).
  • the urgency level determination unit 120 determines that resource allocation is highly necessary and resources should be allocated to the task. To do. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S104).
  • the urgency level determination unit 120 determines whether the urgency level satisfies the above-described second condition when the urgency level does not satisfy the above-described first condition (step S103: No) (step S105). If the urgency level satisfies the second condition (step S105: Yes), the urgency level determination unit 120 determines that it is necessary to further check the load tendency of the server. In response to this determination, the load determination unit 130 measures a load tendency (step S106). The load determination unit 130 may determine to allocate resources to the resource allocation unit 140 according to the measured load tendency. In this case, the resource allocation unit 140 allocates a resource such as a processor core to the task.
  • step S105 if the urgency level determination unit 120 determines that the urgency level does not satisfy both the first and second conditions (step S105: No), the resource allocation system 100 ends the operation. Note that when the load acquisition unit 110 periodically acquires information regarding the load on the server, the resource allocation system 100 may repeatedly execute the operation of the flowchart illustrated in FIG. 3.
  • the resource allocation system 100 performs resource allocation according to the urgency level obtained by the urgency level determination unit 120.
  • the resource allocation system 100 allocates resources without performing further determination. Do. When it is determined that the level of urgency is medium (that is, when the level of urgency satisfies the second condition described above), the resource allocation system 100 further determines the load tendency. When it is determined that the load tends to increase, the resource allocation system 100 may perform resource allocation as necessary.
  • the resource allocation system 100 can quickly allocate resources when the load is high, such as when the load suddenly increases. Further, when the degree of urgency is medium, the resource allocation system 100 determines a load tendency and allocates resources prior to resource allocation. In this way, the allocation of resources that react sensitively to temporary load fluctuations is avoided. Further, when the above-described resource allocation is performed by the resource allocation system 100, it is not necessary to set a policy in advance regarding resource allocation or the like.
  • the resource allocation system 100 enables appropriate resource allocation according to load fluctuations.
  • FIG. 4 is a diagram showing a resource allocation system in the second exemplary embodiment of the present invention.
  • FIG. 5 is a diagram illustrating an example of information related to a load stored in the measurement value storage unit.
  • FIG. 6 is a flowchart showing an example of the operation of the resource allocation system in the second exemplary embodiment of the present invention.
  • FIG. 7 is a flowchart illustrating an example of the operation of the tendency determination unit included in the resource allocation system according to the second embodiment of the present invention.
  • the resource allocation system 200 includes a load acquisition unit 110, an urgency determination unit 120, an interval adjustment unit 231, a tendency determination unit 232, and a resource allocation unit 140.
  • the load acquisition unit 110, the urgency determination unit 120, and the resource allocation unit 140 are basically the same as the same constituent elements included in the resource allocation system 100 according to the first embodiment of the present invention, except for the following description.
  • the interval adjustment unit 231 changes the acquisition interval of information related to the load by the load acquisition unit 110.
  • the trend determination unit 232 determines whether to allocate resources to the task based on the load trend when the interval adjustment unit 231 shortens the information acquisition interval by the load acquisition unit 110.
  • the tendency determination unit 232 is an example of the load determination unit 130.
  • the resource allocation system 200 may include a measurement value storage unit 150 that stores measurement values related to the server load acquired by the load acquisition unit 110.
  • the task to which the resource allocation system 200 in this embodiment is to be assigned a resource is the same as the task to which the resource allocation system 100 in the first embodiment of this invention is to be assigned a resource.
  • the interval at which the information on the load is acquired based on the instruction from the trend determination unit 232 Is changed to a short interval.
  • the interval for acquiring the information regarding the load is changed to a predetermined initial interval.
  • the measured value storage unit 150 stores information on the load, for example, in the form of a table shown in FIG. In the table shown in FIG. 5, the CPU load is stored together with the measurement time and the measurement time interval.
  • the measurement time interval indicates an interval between one measurement time when the CPU load is measured and the measurement time when the CPU load is measured before that.
  • the interval adjustment unit 231 changes the acquisition interval of the information regarding the load by the load acquisition unit 110 when the urgency determination unit 120 determines that the urgency level satisfies the second condition.
  • the interval adjusting unit 231 sets the acquisition interval of the information related to the load by the load acquiring unit 110 as one of two time intervals, a normal time interval and a time interval that is shorter than the normal time interval. Set it.
  • the trend determination unit 232 determines whether to allocate resources to the task based on the load trend when the load information acquisition interval is shortened. That is, when the load tends to increase, the trend determination unit 232 determines to allocate resources to the task. Specifically, when the acquisition interval of information regarding the load is shortened, the tendency determination unit 232 assumes that the load is increasing when the load continues to increase over a predetermined number of measurements. It is determined that resources are allocated to the task.
  • the trend determination unit 232 determines that resources are not allocated to the task, and changes the load acquisition interval of information related to the load by the load acquisition unit 110 to a normal time interval. May be.
  • the tendency determination unit 232 assumes that the load tends to decrease when the load continues to decrease over a predetermined number of measurements. It is determined that no resource is allocated to the task. And the tendency determination part 232 changes the acquisition interval of the information regarding the load by the load acquisition part 110 to a normal space
  • the tendency determination unit 232 operates as follows.
  • the tendency determination unit 232 reads the latest predetermined number of measurement results from the table storing the load of the measurement value storage unit 150. The trend determination unit 232 then increases the load when all of the loads at a plurality of times recorded as measurement results are increased compared to the load measured at the previous time. Judge that there is. If the trend determination unit 232 determines that the load tends to increase, the trend determination unit 232 determines resource allocation to the task. The resource allocation unit 140 allocates resources to tasks according to the determination.
  • the tendency determination unit 232 determines that the load is reduced when all of the loads at a plurality of time points recorded as measurement results are reduced compared to the load measured at the previous time point. Is determined to be decreasing. Then, the trend determination unit 232 determines that no resource is allocated to the task. In response to this determination, the interval adjustment unit 231 changes the acquisition interval of information regarding the load by the load acquisition unit 110 to a normal interval.
  • the tendency determination unit 232 does not need to determine whether to allocate resources at that time. In this case, the tendency determination unit 232 newly acquires a measurement value of information related to the load, and determines whether to allocate resources again based on the newly acquired load tendency.
  • the trend determination unit 232 may determine whether to allocate resources to tasks based on operations and criteria different from the operations described above.
  • the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S201).
  • the urgency level determination unit 120 determines whether a normal time interval is set as an acquisition interval of information regarding the load in the load acquisition unit 110 (step S202).
  • step S204 determines resource allocation to the task.
  • the resource allocation unit 140 allocates a resource such as a processor core to the task (step S205).
  • steps S204, S205, and S206 are the same as those in steps S103, S104, and S105 in the first embodiment of the present invention.
  • the resource allocation system 200 starts again from step S201 onward when the load acquisition unit 110 next acquires information about the server load. Execute the operation.
  • FIG. 7 is a flowchart showing an operation when the tendency determination unit 232 determines a load tendency.
  • the trend determination unit 232 determines the load trend in step S208, the trend determination unit 232 operates as follows.
  • the tendency determination unit 232 acquires information on the load measured by the load acquisition unit 110 at the short time interval described above (step S251). Moreover, the tendency determination part 232 reads a measurement result, for example from the table with which the measured value memory
  • the tendency determining unit 232 determines whether or not the load is continuously increased over the predetermined number of times described above (step S252). As described above, the trend determination unit 232, when all of the loads measured at each time point of the predetermined number of times described above are larger than the load measured at the previous time point, It is determined that the load tends to increase. In addition, the tendency determination unit 232 compares the measurement value of the latest load stored in the measurement value storage unit 150 and the like with the load newly acquired by the load acquisition unit 110 for a predetermined number of times, You may determine the tendency of load.
  • the tendency determination unit 232 determines that the load is continuously increased over the predetermined number of measurements described above (step S252: Yes), it is determined that the load is increasing. In this case, the trend determination unit 232 determines resource allocation for the task. In response to the determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S253). Then, since the resource allocation has been performed, the tendency determination unit 232 determines to change the acquisition interval of information regarding the load in the load acquisition unit 110 to a normal interval. In response to the determination, the interval adjustment unit 231 changes the acquisition interval of information regarding the load in the load acquisition unit to a normal interval (step S254).
  • the trend determination unit 232 determines whether or not the load is continuously decreased over the predetermined number of times described above (step S255). As described above, the trend determination unit 232, when all of the loads measured at each time point of the predetermined number of times described above are smaller than the load measured at the previous time point, It is determined that the load is decreasing. The tendency determination unit 232 performs this determination using the load acquired from the time when the load satisfies the second condition to the later time.
  • step S255: Yes the trend determination unit 232 determines to change the acquisition interval of information regarding the load in the load acquisition unit 110 to a normal interval.
  • the interval adjustment unit 231 changes the acquisition interval of information regarding the load in the load acquisition unit to a normal interval as the process of step S254 described above.
  • the trend determination unit 232 determines that the load has not decreased continuously over the predetermined number of measurements described above (step S255: No), the load does not tend to increase, but the load trend continues. It is assumed that this is a state that needs to be confirmed. In this case, the trend determination unit 232 ends the process. That is, the load acquisition unit 110 continuously acquires information regarding the load at the short time intervals described above. Moreover, the tendency determination part 232 may determine a load tendency again as mentioned above, for example, after predetermined time passes, for example.
  • the resource allocation system 200 in this embodiment has the same effects as the resource allocation system 100 in the first embodiment. Further, the resource allocation system 200 shortens the load measurement interval and reduces the load trend when the degree of urgency is high enough to allocate resources but not to allocate resources to tasks. Determine. By determining the load trend in this way, the resource allocation system 200 can determine with high accuracy that the load is increasing.
  • the resource allocation system 200 enables more appropriate resource allocation according to load fluctuations than the resource allocation system 100 according to the first embodiment.
  • the pattern matching unit 333 indicates how the load changes with time. It is verified whether it is similar to a predetermined pattern.
  • This predetermined pattern is a pattern representing an increase in load.
  • the pattern matching unit 333 acquires a load in a preset period before the time when it is determined that the degree of urgency satisfies the second condition.
  • the pattern matching unit 333 compares the state of the temporal change of the load during the period with a predetermined pattern that represents a load that increases in time. When these differences satisfy a predetermined condition, the pattern matching unit 333 determines that the temporal change in the load is similar to a predetermined pattern representing an increase in the load. In this case, the pattern matching unit 333 determines allocation of resources to tasks on the assumption that the load tends to increase.
  • the resource allocation unit 140 allocates resources to tasks according to the determination.
  • FIG. 10 is a diagram illustrating an example of patterns stored in the pattern storage unit 335.
  • the pattern shown in FIG. 10 represents a series of load measurement values acquired from time Ts to T0 in FIG. T0 is the time when the urgency obtained based on the load acquired at that time satisfies the second condition for the first time.
  • Ts is a past time point when the load acquisition unit 110 has acquired the load. For example, Ts is determined such that the time from Ts to T0 is 30 times the load acquisition interval by the load acquisition unit 110.
  • T1 is the time when the urgency level obtained based on the load acquired at that time satisfies the first condition for the first time.
  • the similarity determination unit 334 obtains the similarity using, for example, the following calculation formula.
  • the similarity determination unit 334 selects one in advance when a plurality of patterns are stored in the pattern storage unit 335. It is assumed that the load measurement value acquired by the load acquisition unit 110 is expressed as Wi, and the load value included in the pattern stored in the pattern storage unit 335 is expressed as Li. Further, it is assumed that n values are included in each of the load measurement value acquired by the load acquisition unit 110 and the pattern representing the load increase. In this case, the similarity determination unit 334 calculates the similarity R using, for example, the following equation (2).
  • the similarity determination unit 334 determines whether the obtained similarity R satisfies a condition regarding the degree of similarity. As an example, the similarity determination unit 334 compares the obtained similarity R with a threshold value R0 regarding the similarity. When the condition R ⁇ R0 is satisfied (that is, the degree of similarity R is smaller than the threshold value R0), the similarity determination unit 334 changes the temporal change in the load acquired by the load acquisition unit 110. Are determined to be similar. That is, the similarity determination unit 334 determines that the load tends to increase. Then, the similarity determination unit 334 determines resource allocation for the task. The resource allocation unit 140 allocates resources to tasks according to the determination.
  • the similarity determination unit 334 determines the load time acquired by the load acquisition unit 110. It is determined that the state of the general change is different from the pattern to be compared. In this case, the similarity determination unit 334 selects another pattern stored in the pattern storage unit 335 and obtains the similarity R again. If all the patterns stored in the pattern storage unit 335 do not satisfy the condition of R ⁇ R0, it is assumed that the load tendency at that time is not similar to any pattern. In this case, the similarity determination unit 334 determines that the load does not tend to increase.
  • the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S301).
  • the urgency level determination unit 120 calculates the urgency level (step S302). Subsequently, the urgency level determination unit 120 determines whether the urgency level satisfies the first condition (step S303).
  • the urgency level determination unit 120 determines resource allocation to the task. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S304).
  • step S303 When the urgency level does not satisfy the first condition described earlier (step S303: No), the urgency level determination unit 120 determines whether the urgency level satisfies the second condition described above (step S305). ). When the degree of urgency satisfies the second condition (step S305: Yes), the pattern matching unit 333 performs matching between the state of load change with time and the pattern (step S306).
  • the resource allocation system 300 may repeatedly perform the operation of the flowchart illustrated in FIG.
  • FIG. 13 is a flowchart showing an operation when the pattern matching unit 333 determines a load tendency.
  • the pattern matching unit 333 performs pattern matching in step S306 described above, the pattern matching unit 333 operates as follows. In this operation example, it is assumed that the pattern matching unit 333 has the configuration shown in FIG.
  • the similarity determination unit 334 selects and reads one of the patterns stored in the pattern storage unit 335 (step S351). Subsequently, the similarity determination unit 334 obtains the similarity with the pattern read in step S351 of the measured load using the above-described equation (2) (step S352).
  • the degree-of-similarity determination unit 334 determines whether the obtained degree of similarity R satisfies a condition regarding the degree of similarity (step S353). That is, the similarity determination unit 334 compares the similarity R with the threshold value R0. When the similarity R and the threshold value R0 satisfy the condition that R ⁇ R0 (step S353: Yes), the similarity determination unit 334 determines resource allocation to the task. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S354).
  • step S355 Yes
  • the similarity degree determination unit 334 increases the load. It is determined that there is no tendency. In this case, the similarity determination unit 334 ends the process.
  • the resource allocation system 300 is configured according to the urgency level obtained by the urgency level determination unit 120 in the same manner as the resource allocation system 100 according to the first exemplary embodiment. Make an assignment. If it is determined that the degree of urgency is high, the resource allocation system 200 allocates resources without performing further determination. If it is determined that the degree of urgency is not high enough to allocate a resource to a task, the resource allocation system 200 is similar to a pattern in which the temporal change in the load represents a load increase And the load tendency is further determined. When it is determined that the load tends to increase, the resource allocation system 100 allocates resources.
  • the resource allocation system 300 enables the resource allocation according to the load variation more appropriately than the resource allocation system 100 according to the first embodiment.
  • the resource allocation system 400 includes a load acquisition unit 110, an urgency determination unit 420, an interval adjustment unit 231, a tendency determination unit 232, and a resource allocation unit 140.
  • the urgency level determination unit 420 performs basically the same operation as the urgency level determination unit 120 included in the resource allocation system 200 according to the second exemplary embodiment of the present invention, except for portions described below.
  • the load acquisition unit 110 acquires information related to the load of the server that executes a task at a predetermined interval.
  • this interval may be changed to any one of different intervals shorter than a predetermined interval based on an instruction from the interval adjusting unit 231.
  • the urgency determination unit 420 when it is determined that the server load trend needs to be further confirmed, the urgency determination unit 420 further determines the degree of urgency at that time.
  • the urgency level determination unit 420 uses a plurality of conditions with different degrees of urgency level when determining that it is necessary to further confirm the load trend of the server.
  • the urgency level determination unit 420 determines that it is necessary to further check the load tendency of the server when the urgency level satisfies any of the first to third conditions.
  • the second and third conditions are conditions indicating that it is necessary to further determine whether or not resources should be allocated to the task.
  • the second and third conditions are conditions with different degrees of urgency.
  • the third condition indicates that the degree of urgency is lower than that of the second condition, and the necessity for resource allocation to tasks is low.
  • the first condition is the same as the first condition in each of the embodiments described above.
  • each of the first to third conditions is expressed in the form of a threshold for the value obtained as the urgency level.
  • E1 is determined as the threshold for the first condition
  • E2 is set as the threshold for the second condition
  • E3 is set as the threshold for the third condition. It is done.
  • the urgency level determination unit 420 determines that the urgency level satisfies the first condition when the urgency level satisfies E1 ⁇ E (that is, the level of the urgency level E is equal to or greater than E1).
  • the urgency level determination unit 420 sets the urgency level to the second condition when the urgency level E satisfies E2 ⁇ E ⁇ E1 (that is, the level of the urgency level E is equal to or larger than E2 and smaller than E1). It is determined that The urgency level determination unit 420 also sets the urgency level to the third level when the urgency level satisfies E3 ⁇ E ⁇ E2 (that is, when the urgency level E is greater than or equal to E3 and smaller than E2). It is determined that the condition is satisfied. When the urgency level satisfies E ⁇ E3 (that is, the urgency level is smaller than E3), the urgency level determination unit 420 determines that the urgency level does not satisfy any of the conditions.
  • the interval adjustment unit 231 sets the acquisition interval by the load acquisition unit 110 to a shorter time interval than the normal time interval. Then, when the urgency level satisfies the second condition, the interval adjusting unit 231 sets the acquisition interval by the load acquisition unit 110 to be shorter than the time interval when the urgency level satisfies the third condition. Set to interval.
  • the interval adjustment unit 231 uses the load acquisition unit 110 to shorten the time interval according to the level of urgency. Set the acquisition interval of information about the load.
  • the load acquisition unit 110 acquires information related to the load of the server that executes the task (step S401).
  • the urgency determining unit 420 determines whether or not a normal time interval is set as an acquisition interval of information regarding the load in the load acquiring unit 110 (step S402).
  • step S404 determines resource allocation to the task.
  • the resource allocation unit 140 allocates a resource such as a processor core to the task (step S405).
  • step S404 When the urgency level does not satisfy the first condition described above (step S404: No), the urgency level determination unit 420 determines whether the urgency level satisfies the second condition described above (step S406). ). When the degree of urgency satisfies the second condition (step S406: Yes), the interval adjustment unit 231 sets the acquisition interval of the information regarding the load in the load acquisition unit 110 to a shorter time interval than the normal time interval. (Step S407).
  • step S402 when a normal time interval is not set for the load acquisition unit 110 as a load information acquisition interval (step S402: No), a short time interval is set as the load information acquisition interval. It is assumed that In this case, the tendency determination unit 232 determines a load tendency (step S408).
  • step S401 to S408 are the same as the respective steps from step S201 to S208 in the second embodiment of the present invention.
  • step S408 the trend determination unit 232 determines the load trend according to the flowchart shown in FIG.
  • the resource allocation system 400 may repeatedly perform the operation of the flowchart illustrated in FIG. 15.
  • the resource allocation system 400 according to the fourth embodiment of the present invention determines the load tendency by shortening the load measurement interval, similarly to the resource allocation system 200 according to the second embodiment. Therefore, the resource allocation system 400 in this embodiment has the same effect as the resource allocation system 200 in the second embodiment.
  • the urgency level determination unit 420 determines that it is necessary to further check the load tendency of the server, a plurality of conditions having different levels of urgency level are used. . And when setting the acquisition interval of the information regarding the load by the load acquisition unit 110 to be a short interval compared to the normal time interval, the time interval is shortened according to the high degree of urgency determined by the urgency determination unit 420. Set the time interval to By doing in this way, it becomes possible for the tendency determination part 232 to determine at an early stage that the load is increasing.
  • the resource allocation system 400 enables the resource allocation according to the load variation more appropriately than the resource allocation system 200 according to the second embodiment.
  • the urgency level determination unit 420 determines whether the urgency level satisfies any of the first to third conditions as an example. And the space
  • FIG. 16 is a diagram showing a resource allocation system in the fifth exemplary embodiment of the present invention.
  • FIG. 17 is a flowchart showing an example of the operation of the resource allocation system in the fifth exemplary embodiment of the present invention.
  • the resource allocation system 500 includes a load acquisition unit 110, an urgency determination unit 520, a tendency determination unit 532, an interval adjustment unit 531, and a pattern matching unit. 533, a resource allocation unit 140, and a measurement value storage unit 150.
  • the load acquisition unit 110 and the resource allocation unit 140 operate basically the same as the same components provided in the resource allocation system 100 according to the first embodiment of the present invention, except for the parts described below. I do.
  • the urgency level determination unit 520 performs basically the same operation as the urgency level determination unit 420 included in the resource allocation system 100 according to the fourth embodiment of the present invention, except as described below.
  • the trend determination unit 532 and the interval adjustment unit 531 are basically the same as the trend determination unit 232 and the interval adjustment unit 231 included in the resource allocation system 100 according to the second embodiment of the present invention, except for the portions described below. Perform the operation.
  • the pattern matching unit 533 basically performs the same operation as that of the pattern matching unit 333 included in the resource allocation system 300 according to the third embodiment of the present invention, except for the portions described below.
  • the resource allocation system 500 includes the tendency determination unit 232 and the interval adjustment unit 231 according to the second embodiment, and the pattern matching unit 333 according to the third embodiment. And different.
  • the load acquisition unit 110 acquires information related to the load of the server that executes a task at a predetermined interval. In addition, this interval may be changed based on an instruction from the interval adjusting unit 531.
  • the urgency level determination unit 520 determines, as in the fourth embodiment, whether one of a plurality of conditions with different degrees of urgency is satisfied as the determination of the necessity regarding the allocation of resources to tasks. That is, the urgency determination unit 520 determines whether or not the urgency satisfies any of the first to third conditions described above.
  • the interval adjustment unit 531, the tendency determination unit 532, and the pattern matching unit 533 perform an operation related to resource allocation according to the determination of the urgency level determination unit 520.
  • the interval adjustment unit 531 changes the acquisition interval of the information regarding the load by the load acquisition unit 110 when the urgency determination unit 520 determines that the urgency level satisfies, for example, the third condition described above.
  • the interval adjusting unit 531 has a shorter time than the normal time interval for acquiring the load-related information by the load acquiring unit 110 when the urgency level is determined to satisfy the second condition. Set to interval.
  • the pattern matching unit 533 is similar to a predetermined pattern in which the state of temporal change in load indicates an increase in load when the urgency level determination unit 520 determines that the urgency level satisfies the second condition described above, for example. Check if you want to.
  • the pattern matching unit 533 has the configuration described in the third embodiment of the present invention as a specific example.
  • the resource allocation system 500 in this embodiment determines that it is necessary to check the load tendency of the server, it checks the load tendency of the server in different procedures depending on the degree of urgency. By doing in this way, the resource allocation system 500 can perform quick resource allocation when the load tends to increase while suppressing unnecessary resource allocation.
  • the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S501).
  • the urgency determining unit 520 determines whether or not a normal time interval is set as an acquisition interval of information related to the load in the load acquiring unit 110 (step S502).
  • the urgency level determination unit 520 obtains an urgency level (step S503). Then, the urgency level determination unit 520 determines whether the urgency level satisfies the first condition (step S504).
  • step S504 determines resource allocation to the task.
  • the resource allocation unit 140 allocates a resource such as a processor core to the task (step S505).
  • steps S501 to S505 the same operations as in steps S401 to S405 in the fourth embodiment of the present invention are performed.
  • the urgency level determination unit 520 determines whether the urgency level satisfies the second condition described above (step S506). ). If the degree of urgency satisfies the second condition (step S506: Yes), the pattern matching unit 533 collates the state of the load with time and the pattern (step S507).
  • step S506 the same operation as in step S406 in the fourth embodiment of the present invention is performed.
  • step S507 the same operation as in step S306 in the third embodiment of the present invention is performed. That is, the pattern matching unit 533 performs the operation described above with reference to the flowchart shown in FIG. 13 in the third embodiment.
  • the resource allocation system 500 ends the operation.
  • the pattern matching unit 533 can perform pattern matching without newly acquiring information related to the load, for example, by referring to the measurement value stored in the measurement value storage unit 150. Therefore, the pattern matching unit 533 can quickly allocate resources when necessary.
  • step S506 determines whether the urgency level satisfies the third condition described above.
  • step S508 determines whether the urgency level satisfies the third condition described above.
  • step S509 sets the acquisition interval of information regarding the load in the load acquisition unit 110 to a shorter time interval than the normal time interval. Setting is made (step S509).
  • step S509 the same operation as in step S207 in the second embodiment is performed.
  • the resource allocation system 500 ends the operation.
  • step S508 determines that none of the conditions regarding the urgency is satisfied (step S508: No).
  • step S502 when a normal time interval is not set for the load acquisition unit 110 as the information acquisition interval for the load (step S502: No), a short time interval is set as the information acquisition interval for the load. It is assumed that In this case, the tendency determination unit 532 determines a load tendency (step S510). In step S510, the trend determination unit 532 determines the load trend according to the flowchart shown in FIG. When the process in step S510 is performed, the resource allocation system 500 ends the operation.
  • the resource allocation system 500 may repeatedly execute the operation of the flowchart shown in FIG.
  • the resource allocation system 500 includes the interval adjustment unit 531, the tendency determination unit 532, and the pattern matching unit 533. Therefore, the resource allocation system 500 in the present embodiment has the same effects as the resource allocation system 200 in the second embodiment and the resource allocation system 300 in the third embodiment.
  • the pattern matching unit 533 compares the load stored in the measured value storage unit 150 with the pattern to determine the load tendency, whereby the load tendency is quickly determined.
  • the above-described pattern matching is performed when it is determined that it is necessary to check the load trend of the server and the degree of urgency is relatively high. Therefore, when the load tends to increase, it is possible to quickly allocate resources.
  • interval adjustment unit 531 and the trend determination unit 532 newly determine the load trend using the load acquired by the load acquisition unit 110 at a shorter interval than normal, so that the load trend is accurate. Will be judged. Further, in this embodiment, when it is determined that it is necessary to check the load tendency of the server and the degree of urgency is relatively low, the load is obtained using the load acquired at a short interval described above. The tendency is determined. Therefore, when the load tends to increase, appropriate resource allocation is performed.
  • the resource allocation system 500 in the present embodiment enables appropriate resource allocation in accordance with load fluctuations with higher accuracy than in the other embodiments.
  • FIG. 18 is a diagram illustrating the configuration of the resource allocation system and the task execution server in the first embodiment.
  • the resource allocation system 200 and the task execution server 11 on which a task to be allocated a resource is executed are connected via the communication network 12 as shown in FIG.
  • the task execution server 11 is, for example, the above-described information processing apparatus 1000 illustrated in FIG.
  • a program that performs processing in response to a request from another server or the like is executed as a task.
  • the program is stored in the storage device 1005 of the information processing apparatus 1000, for example, and is read from the storage device 1005 and expanded in the RAM 1003 when the program is executed. Further, the task execution server 11 is connected to other servers and the like via the external communication network 13.
  • the task execution server 11 includes a plurality of CPUs.
  • the task execution server 11 may include a multi-core processor having a plurality of processor cores as a CPU. It is assumed that the above-described program is initially executed by one CPU (or one processor core) among a plurality of CPUs.
  • the resource allocation system 200 is realized as a program executed by the information processing apparatus 1000 described above shown in FIG.
  • the resource allocation system 200 and the task execution server 11 are connected via the communication network 12.
  • the resource allocation system 200 (a program that realizes the resource allocation system 200) acquires information regarding the load on the task execution server 11 via the communication network 12.
  • the measurement value storage unit 150 is realized as a table of a relational database (Relational Database, hereinafter referred to as “RDB”) held in the storage device 1005 of the information processing apparatus 1000.
  • the RDB table includes, as attributes, the load measurement time, the CPU load at the measurement time, and the measurement time interval between the measurement time and the previous measurement time.
  • the acquisition interval of the information regarding the load by the load acquisition unit 110 is stored in the RAM 1003 in advance.
  • the load acquisition unit 110 acquires a load with a normal time interval of 5 seconds and a short time interval of 1 second. Information about the acquired load is recorded as a record of a table constituting the RDB that becomes the measurement value storage unit 150 each time measurement is performed. An index is attached to the measurement time, which is an attribute of the table included in the measurement value storage unit 150, and the search or alignment of values is performed at high speed.
  • the urgency level determination unit 120 specifically calculates the urgency level E using the following equation (3).
  • the expression (3) is an expression in which a is 1.0 and b is 0.5 in the above-described expression (1).
  • the urgency level determination unit 120 determines that the urgency level satisfies the second condition when the urgency level exceeds 60.
  • the urgency level determination unit 120 determines that the urgency level satisfies the first condition when the urgency level exceeds 90.
  • the tendency determination unit 232 determines that the load tends to increase, Is determined to be assigned. Also, in this case, when the load continues to decrease three times, the trend determination unit 232 determines that the load is in a decreasing trend and does not allocate resources to the task.
  • FIG. 19 is a table showing information about the load acquired by the load acquisition unit 110 and an example of the degree of urgency obtained for the load.
  • the information regarding the load indicates the load of the CPU.
  • FIG. 20 is a graph showing the load shown in FIG. 19 with the horizontal axis as the measurement time and the vertical axis as the load.
  • the urgency level at time 5:42:40 (“5:42:40” represents 5:42:40.
  • the time is similarly expressed) is time 5:42: Using the load value 27 at 40 and the load value 25 at time 5:42:35, the following is obtained.
  • the urgency level determination unit 120 performs derivation of the urgency level and determination of the urgency level for the next measurement result without performing any special processing related to resource allocation or the like.
  • the degree of urgency obtained based on the load measured at each time until 5:42:55 does not satisfy any of the first and second conditions described above. Therefore, the urgency determination unit 120 does not perform special processing related to resource allocation or the like.
  • the emergency level determination unit 120 calculates the emergency level based on the time and the load measured at time 5:42:55, the emergency level is calculated as 100. That is, the urgency satisfies the first condition described above. In this case, the urgency determining unit 120 determines to allocate resources to the task. In response to this determination, the resource allocation unit 140 allocates a CPU (or processor core) as a resource to the task. That is, a task to which resources are allocated is executed using two CPUs (or processor cores).
  • FIG. 21 is a table showing another example of information on the load acquired by the load acquisition unit 110 and the degree of urgency obtained for the load.
  • the information regarding the load indicates the load on the CPU.
  • FIG. 22 is a graph showing the load shown in FIG. 21 with the horizontal axis as the measurement time and the vertical axis as the load.
  • the table shown in FIG. 21 shows information such as the load acquired at a different time from the load shown in the table of FIG. 19 in the resource allocation system 200 shown in FIG.
  • the urgency level is calculated using the equation (3) described above.
  • each of the urgency levels obtained at time 7:12:20 and time 7:12:25 does not satisfy both the first and second conditions.
  • the load acquisition unit 110 acquires information about the load at intervals of 1 second.
  • the load acquired at each of the three times after the time 7:12:31 when the load is acquired at 1 second intervals is a small value compared to the load acquired immediately before that. That is, the load continues to decrease three times.
  • the trend determining unit 232 determines that the resource is not allocated to the task, assuming that the load tends to decrease. Based on this determination, the interval adjustment unit 231 changes the load measurement interval to a normal interval.
  • FIG. 23 is a table showing still another example of the information about the load acquired by the load acquisition unit 110 and the urgency level obtained for the load.
  • the information regarding the load indicates the load on the CPU.
  • FIG. 24 is a graph showing the load shown in FIG. 23 with the horizontal axis as the measurement time and the vertical axis as the load.
  • the table shown in FIG. 23 shows information such as the load acquired at a different time from the load shown in the table of FIG. 19 or FIG. 23 in the resource allocation system 200 shown in FIG.
  • the urgency level is obtained using the equation (3) described above.
  • each of the urgency levels obtained at time 8:25:20 and time 8:25:25 does not satisfy both the first and second conditions.
  • required based on the load acquired at the time 8:25:30 satisfy
  • the load acquisition unit 110 acquires information about the load at intervals of 1 second. And the load acquired at each time of three time points after the time 8:25:31 when the load is acquired at intervals of 1 second is larger than the load acquired immediately before that. That is, the load continues to increase three times.
  • the trend determination unit 232 determines that resources are allocated to the task, assuming that the load tends to increase. Then, based on this determination, the trend determination unit 232 determines resource allocation for the task. In response to this determination, the resource allocation unit 140 allocates a CPU (or processor core) as a resource to the task. That is, a task to which resources are allocated is executed using two CPUs (or processor cores). In addition, when the resource allocation unit 140 performs resource allocation, the interval adjustment unit 231 changes the load measurement interval to a normal interval.
  • the resource allocation system 200 determines that the urgency level satisfies the first condition, the resource is quickly allocated to the task.
  • the load measurement interval was set to a short time interval. When the load tends to increase, resources are allocated to the tasks.
  • the resource allocation system 200 performs appropriate resource allocation according to load fluctuations.
  • Example 2 Next, another example of resource allocation using the resource allocation system in each embodiment of the present invention will be described. First, an example of resource allocation using the resource allocation system 300 according to the third embodiment of the present invention will be described.
  • the resource allocation system 300 and the task execution server 11 on which a task to be allocated a resource is executed are connected via a communication network as shown in FIG.
  • the task execution server 11 is the above-described information processing apparatus 1000 shown in FIG. 28, for example, as in the case of the first embodiment described above.
  • the task execution server 11 includes a plurality of CPUs (or a plurality of processor cores).
  • the program executed by the task execution server is initially assumed to be executed by one CPU (or one processor core) among a plurality of CPUs.
  • the resource allocation system 300 is realized as a program executed by the above-described information processing apparatus 1000 shown in FIG. 28, similarly to the resource allocation system 200 described in the first embodiment.
  • the pattern storage unit 335 of the pattern matching unit 333 is realized as an RDB table held in the storage device 1005 of the information processing apparatus 1000.
  • FIG. 26 is an example of a pattern stored in the pattern storage unit 335.
  • the pattern storage unit 335 stores five patterns. Each of the patterns includes 10 loads. In this embodiment, the load represents a change in load every 5 seconds.
  • the load acquisition unit 110 acquires information about the load at intervals of 5 seconds.
  • the urgency level determination unit 120 specifically calculates the urgency level E using the above-described equation (3). Further, the urgency level determination unit 120 determines that the urgency level satisfies the second condition when the urgency level exceeds 60. The urgency level determination unit 120 determines that the urgency level satisfies the first condition when the urgency level exceeds 90.
  • the measurement value storage unit 150 is at least 10 times from the most recently acquired load so as to enable comparison between the load acquired by the load acquisition unit 110 and the pattern stored in the pattern storage unit 335 described above. Memorize the load.
  • the similarity determination unit 334 of the pattern matching unit 333 sets the threshold value R0 regarding the similarity to 1.0.
  • FIG. 27 is a table showing information on the load acquired by the load acquisition unit 110 and an example of the degree of urgency obtained for the load.
  • the information regarding the load indicates the load on the CPU.
  • the emergency level determination unit 120 calculates the emergency level. At times before 9:54:20, the urgency level does not satisfy both the first and second conditions. Then, at time 9:54:25, when the emergency level determination unit 120 calculates the emergency level, the emergency level is calculated to be 82.0. That is, the urgency satisfies the second condition.
  • the similarity determination unit 334 collates the state of the load with time and the pattern. That is, the similarity determination unit 334 obtains the similarity between the load measured from time 9:53:40 to time 9:54:25 and each of the patterns shown in FIG.
  • the similarity with the pattern with ID 1 is obtained using the above-described equation (2) according to the following equation.
  • the similarity with the pattern with ID 1 is 4.3.
  • the similarity with the pattern with ID 2 is 2.7
  • the similarity with the pattern with ID 3 is 2.0
  • the similarity with the pattern with ID 4 is 2.9
  • the pattern with ID 5 The degree of similarity is obtained as 0.8.
  • required regarding the pattern with ID 5 is smaller than 1.0 which is the value of threshold value R0 mentioned above. Therefore, the similarity determination unit 334 determines that the load tends to increase.
  • the similarity determination unit 334 determines resource allocation for the task.
  • the resource allocation unit 140 allocates a CPU (or processor core) as a resource to the task. That is, a task to which resources are allocated is executed using two CPUs (or processor cores).
  • the resource allocation system 300 performs appropriate resource allocation according to load fluctuations.

Abstract

Provided is a resource allocation system, etc., with which it is possible to allocate appropriate resources according to fluctuations in load. A resource allocation system 100 is provided with: a load acquisition unit 110 for acquiring information pertaining to the load on a server executing a task; an urgency determination unit 120 for determining need pertaining to allocation of resources for a task on the basis of the urgency determined using the information pertaining to the load; a load determination unit 130 for determining the trend of the load on the server when the urgency determination unit 120 has determined that it is necessary to confirm the trend of the load on the server; and a resource allocation unit 140 for allocating resources for the task when the urgency determination unit 120 has determined that it is necessary to allocate resources for the task.

Description

リソース割当システム、リソース割当方法及びコンピュータ読み取り可能記録媒体Resource allocation system, resource allocation method, and computer-readable recording medium
　本発明は、リソース割当システム、リソース割当方法及びコンピュータ読み取り可能記録媒体に関する。 The present invention relates to a resource allocation system, a resource allocation method, and a computer-readable recording medium.
　複数のタスクを実行する計算機システムにおいては、当該タスクの負荷に応じて、必要なリソースの割り当てが必要となる。また、タスクの負荷が急激に増加した場合には、処理性能を低下させないよう、当該タスクに対する速やかなリソースの割り当てが求められる。 In a computer system that executes multiple tasks, it is necessary to allocate necessary resources according to the load of the task. Further, when the task load increases rapidly, it is required to quickly allocate resources to the task so as not to degrade the processing performance.
　特許文献１には、アプリケーションの負荷に対して計算機リソースを短時間で追従させる技術が記載されている。特許文献１に記載のリソース割当制御サーバは、収集した負荷情報と、過去に収集した負荷情報とを用いて、アプリケーションを実行する際の負荷の増減量を算出する。そして、特許文献１に記載のリソース割当制御サーバは、算出した増減量に基づいて、アプリケーションに割当るリソース量を決定する。 Patent Document 1 describes a technique for making computer resources follow a load of an application in a short time. The resource allocation control server described in Patent Literature 1 uses the collected load information and the load information collected in the past to calculate an increase / decrease amount of the load when the application is executed. Then, the resource allocation control server described in Patent Literature 1 determines the resource amount to be allocated to the application based on the calculated increase / decrease amount.
　特許文献２には、システムの動作状況に応じて不足している資源の種別を適切に判断し、その資源を増設できるようにする技術が記載されている。特許文献２に記載のシステム管理装置においては、運用中のシステムの負荷状態が動作状況監視手段で測定され、追加資源判定手段により、資源追加ポリシーに基づいて、ハードウェア資源の追加の要否等が判定される。そして、特許文献２に記載のシステム管理装置においては、サーバを追加すべきと判定された場合、サーバ追加手段により、予備用サーバによる運用が開始される。 Patent Document 2 describes a technology that appropriately determines the type of a resource that is insufficient according to the operating status of the system, and enables the addition of the resource. In the system management apparatus described in Patent Document 2, the load state of the operating system is measured by the operation status monitoring unit, and the additional resource determination unit determines whether or not hardware resources need to be added based on the resource addition policy. Is determined. In the system management device described in Patent Document 2, when it is determined that a server should be added, the server addition unit starts operation using the spare server.
　特許文献３には、負荷の変動に対して過敏に反応して資源の再割り当てが頻繁に発生するのを防止する技術が記載されている。特許文献３に記載の計算機システムは、資源数算出式テーブルに保持されている情報の定義する複数の資源数算出式に従って、当該複数の資源数算出式毎にサービスに必要な資源数を算出する。有効資源数判定機構は、複数の資源数算出式毎に算出された必要資源数の１つを、有効資源数判定式選択機構によって有効資源数判定式テーブルの中から選択された有効資源数判定式に従って有効資源数として判定する。割り当て資源数調整機構は、判定された有効資源数をもとに、対応するサービスに割当てる資源数を調整する。特許文献３に記載の計算機システムにおいては、有効資源数判定式は、ユーザによって選択される。 Patent Document 3 describes a technique for preventing frequent reassignment of resources by reacting sensitively to load fluctuations. The computer system described in Patent Literature 3 calculates the number of resources required for the service for each of the plurality of resource number calculation formulas according to the plurality of resource number calculation formulas defined by the information held in the resource number calculation formula table. . The effective resource number determination mechanism determines the effective resource number selected from the effective resource number determination expression table by one of the required resource numbers calculated for each of the plurality of resource number calculation expressions from the effective resource number determination expression selection mechanism. The number of effective resources is determined according to the formula. The allocated resource number adjustment mechanism adjusts the number of resources allocated to the corresponding service based on the determined effective resource number. In the computer system described in Patent Document 3, the effective resource number determination formula is selected by the user.
　特許文献４には、リソースの負荷情報の測定を構成度に保ちつつネットワークを効果的に管理して安定に制御すると同時に、測定オーバーヘッドを減少させる技術が記載されている。特許文献４に記載のリソース負荷測定方法においては、リソースの負荷情報を測定間隔で測定した測定負荷情報とリソースの負荷情報を予測アルゴリズムに応じて予測した予測負荷情報とに基づいて測定間隔を調整するように構成されている。 Patent Document 4 describes a technique for effectively managing a network and controlling it stably while maintaining measurement of load information of resources at a configuration, and at the same time reducing measurement overhead. In the resource load measurement method described in Patent Literature 4, the measurement interval is adjusted based on the measured load information obtained by measuring the resource load information at the measurement interval and the predicted load information obtained by predicting the resource load information according to the prediction algorithm. Is configured to do.
特開２０１４－１６４７１５号公報JP 2014-164715 A 特開２００６－１１８６０号公報Japanese Patent Laid-Open No. 2006-11860 特開２００５－１７３９２８号公報JP 2005-173928 A 特開２００５－８６８０３号公報JP 2005-86803 A
　特許文献１又は２に記載の技術においては、計算機システムの負荷に関する単一の閾値に基づいてリソースの割り当てが制御される。この場合においては、例えば相対的に小さな負荷に対してリソースの割り当てが行われるように閾値が設定されることで、計算機システムにおける急激な負荷の変動への速やかな対応が可能となる。しかしながら、このように閾値が設定されることで、閾値に近い範囲での負荷の変動に応じて、必ずしも必要ではないリソースの割り当てが行われる場合がある。 In the technique described in Patent Document 1 or 2, resource allocation is controlled based on a single threshold value related to the load on the computer system. In this case, for example, by setting a threshold value so that resources are allocated to a relatively small load, it is possible to quickly cope with a sudden load fluctuation in the computer system. However, by setting the threshold value in this way, resource allocation that is not necessarily required may be performed in accordance with load fluctuations in a range close to the threshold value.
　一方、特許文献３に記載の技術においては、例えばユーザが選択した資源数算出式に基づいて、対応するサービスに割当てる資源数が調整される。そのため、例えば急激な負荷の上昇であるスパイク現象が発生しないとの想定に基づいて資源数算出式が選択されている際にスパイク現象が発生した場合等においては、速やかなリソースの割り当てが行われない可能性がある。 On the other hand, in the technique described in Patent Document 3, the number of resources allocated to the corresponding service is adjusted based on, for example, a resource number calculation formula selected by the user. For this reason, for example, when a spike phenomenon occurs when the resource count calculation formula is selected based on the assumption that the spike phenomenon, which is a sudden load increase, does not occur, quick resource allocation is performed. There is no possibility.
　すなわち、特許文献１から４に記載の技術においては、負荷の変動に応じた適切なリソースの割り当てが困難であるという課題がある。 That is, in the techniques described in Patent Documents 1 to 4, there is a problem that it is difficult to allocate an appropriate resource according to a load change.
　本発明は、上記課題を解決するためになされたものであって、負荷の変動に応じた適切なリソースの割当てを可能にするリソース割当システム等を提供することを主たる目的とする。 The present invention has been made in order to solve the above-mentioned problems, and has as its main object to provide a resource allocation system and the like that enable appropriate resource allocation according to load fluctuations.
　本発明の一態様におけるリソース割当システムは、タスクを実行するサーバの負荷に関する情報を取得する負荷取得手段と、負荷に関する情報を用いて求められる緊急度に基づいて、タスクに対するリソースの割当てに関する必要性を判定する緊急度判定手段と、緊急度判定手段がサーバの負荷の傾向を確認する必要があると判定した場合に、サーバの負荷の傾向を判定する負荷判定手段と、緊急度判定手段がタスクにリソースを割当てる必要があると判定した場合に、タスクにリソースを割当てるリソース割当手段とを備える。 A resource allocation system according to an aspect of the present invention is a load acquisition unit that acquires information regarding a load of a server that executes a task, and a need for allocation of resources to a task based on an urgency level that is obtained using the information regarding the load The urgency determination means for determining the load, the load determination means for determining the load trend of the server when the urgency determination means determines that it is necessary to check the load trend of the server, and the urgency determination means Resource allocation means for allocating a resource to a task when it is determined that a resource needs to be allocated to the task.
　また、本発明の一態様におけるリソース割り当て方法は、タスクを実行するサーバの負荷に関する情報を取得し、負荷に関する情報を用いて求められる緊急度に基づいて、タスクに対するリソースの割当てに関する必要性を判定し、必要性として、サーバの負荷の傾向を確認する必要があると判定された場合に、サーバの負荷の傾向を判定し、必要性として、緊急度判定手段がタスクにリソースを割当てる必要があると判定された場合に、タスクにリソースを割当てる。 In addition, the resource allocation method according to one aspect of the present invention acquires information related to a load of a server that executes a task, and determines the necessity for allocation of a resource to a task based on an urgency level obtained using the information related to the load. However, if it is determined that it is necessary to check the load trend of the server, it is necessary to determine the load trend of the server and, as necessary, the urgency determination means must allocate resources to the task. When it is determined that the resource is assigned to the task.
　また、本発明の一態様におけるコンピュータ読み取り可能記録媒体は、コンピュータに、タスクを実行するサーバの負荷に関する情報を取得する処理と、負荷に関する情報を用いて求められる緊急度に基づいて、タスクに対するリソースの割当てに関する必要性を判定する処理と、必要性の判定において、サーバの負荷の傾向を確認する必要があると判定された場合に、サーバの負荷の傾向を判定する処理と、必要性の判定において、タスクにリソースを割当てる必要があると判定された場合に、タスクにリソースを割当てる処理とを実行させるプログラムを非一時的に格納する。 In addition, the computer-readable recording medium according to one embodiment of the present invention provides a resource for a task based on a process of acquiring information on a load of a server that executes a task on the computer and an urgency level obtained using the information on the load. The process of determining the necessity for allocation of the server, the process of determining the load trend of the server when it is determined in the determination of necessity that it is necessary to check the load trend of the server, and the determination of necessity In step (1), when it is determined that a resource needs to be allocated to a task, a program that executes a process of allocating a resource to the task is stored temporarily.
　本発明によると、負荷の変動に応じた適切なリソースの割当てを可能にするリソース割当システム等を提供することができる。 According to the present invention, it is possible to provide a resource allocation system or the like that enables appropriate resource allocation according to load fluctuations.
本発明の第１の実施形態におけるリソース割当システムの構成を示す図である。It is a figure which shows the structure of the resource allocation system in the 1st Embodiment of this invention. 本発明の第１の実施形態等におけるリソース割当システムとタスク実行サーバとの接続関係を示す図である。It is a figure which shows the connection relation of the resource allocation system and task execution server in the 1st Embodiment etc. of this invention. 本発明の第１の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。It is a flowchart which shows an example of operation | movement of the resource allocation system in the 1st Embodiment of this invention. 本発明の第２の実施形態におけるリソース割当システムの構成を示す図である。It is a figure which shows the structure of the resource allocation system in the 2nd Embodiment of this invention. 本発明の第２の実施形態において測定値記憶部にて記憶される負荷に関する情報の一例を示す図である。It is a figure which shows an example of the information regarding the load memorize | stored in the measured value memory | storage part in the 2nd Embodiment of this invention. 本発明の第２の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。It is a flowchart which shows an example of operation | movement of the resource allocation system in the 2nd Embodiment of this invention. は、本発明の第２の実施形態におけるリソース割当システムに含まれる傾向判定部の動作の一例を示すフローチャートである。These are flowcharts which show an example of operation | movement of the tendency determination part contained in the resource allocation system in the 2nd Embodiment of this invention. 本発明の第３の実施形態におけるリソース割当システムの構成を示す図である。It is a figure which shows the structure of the resource allocation system in the 3rd Embodiment of this invention. 本発明の第３の実施形態におけるリソース割当システムのパターン照合部に関する具体的な構成例を示す図である。It is a figure which shows the specific structural example regarding the pattern collation part of the resource allocation system in the 3rd Embodiment of this invention. 本発明の第３の実施形態において負荷上昇パターンの一例を示す図である。It is a figure which shows an example of a load increase pattern in the 3rd Embodiment of this invention. 本発明の第３の実施形態において負荷取得部により取得される負荷と負荷上昇パターンとの関係を説明する図である。It is a figure explaining the relationship between the load acquired by the load acquisition part in the 3rd Embodiment of this invention, and a load increase pattern. 本発明の第３の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。It is a flowchart which shows an example of operation | movement of the resource allocation system in the 3rd Embodiment of this invention. 本発明の第３の実施形態におけるリソース割当システムに含まれる傾向判定部の動作の一例を示すフローチャートである。It is a flowchart which shows an example of operation | movement of the tendency determination part contained in the resource allocation system in the 3rd Embodiment of this invention. 本発明の第４の実施形態におけるリソース割当システムを示す図である。It is a figure which shows the resource allocation system in the 4th Embodiment of this invention. 本発明の第４の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。It is a flowchart which shows an example of operation | movement of the resource allocation system in the 4th Embodiment of this invention. 本発明の第５の実施形態におけるリソース割当システムを示す図である。It is a figure which shows the resource allocation system in the 5th Embodiment of this invention. 本発明の第５の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。It is a flowchart which shows an example of operation | movement of the resource allocation system in the 5th Embodiment of this invention. 実施例１におけるリソース割当システムとタスク実行サーバの構成を示す図である。It is a figure which shows the structure of the resource allocation system in Example 1, and a task execution server. 実施例１において、負荷に関する情報と、当該負荷に対して求められた緊急度の例を示す表である。In Example 1, it is the table | surface which shows the information regarding the load, and the example of the urgency degree calculated | required with respect to the said load. 図１９に示すＣＰＵ負荷とＣＰＵ負荷が測定された時刻との関係を示すグラフである。It is a graph which shows the relationship between CPU load shown in FIG. 19, and the time when CPU load was measured. 実施例１において、負荷に関する情報と、当該負荷に対して求められた緊急度の別の例を示す表である。In Example 1, it is a table | surface which shows another example of the information regarding load, and the urgency degree calculated | required with respect to the said load. 図２１に示すＣＰＵ負荷とＣＰＵ負荷が測定された時刻との関係を示すグラフである。It is a graph which shows the relationship between CPU load shown in FIG. 21, and the time when CPU load was measured. 実施例１において、負荷に関する情報と、当該負荷に対して求められた緊急度の更に別の例を示す表である。In Example 1, it is the table | surface which shows another example of the information regarding load, and the urgency degree calculated | required with respect to the said load. 図２３に示すＣＰＵ負荷とＣＰＵ負荷が測定された時刻との関係を示すグラフである。It is a graph which shows the relationship between CPU load shown in FIG. 23, and the time when CPU load was measured. 実施例２におけるリソース割当システムとタスク実行サーバの構成を示す図である。It is a figure which shows the structure of the resource allocation system in Example 2, and a task execution server. 実施例２においてパターン記憶部に記憶されるパターンの例である。It is an example of the pattern memorize | stored in a pattern memory | storage part in Example 2. FIG. 負荷に関する情報と、当該負荷に対して求められた緊急度の例を示す表である。It is a table | surface which shows the example of the urgency calculated | required with respect to the information regarding load, and the said load. 本発明の各実施形態におけるリソース割当システム等を実現する情報処理装置の一例を示す図である。It is a figure which shows an example of the information processing apparatus which implement | achieves the resource allocation system etc. in each embodiment of this invention.
　本発明の各実施形態について、添付の図面を参照して説明する。なお、本発明の各実施形態において、各装置の各構成要素は、機能単位のブロックを示している。各装置の各構成要素は、例えば図２８に示すような情報処理装置１０００とソフトウェアとの任意の組み合わせにより実現することができる。情報処理装置１０００は、一例として、以下のような構成を含む。 Embodiments of the present invention will be described with reference to the accompanying drawings. In each embodiment of the present invention, each component of each device represents a functional unit block. Each component of each device can be realized by any combination of an information processing device 1000 and software as shown in FIG. 28, for example. The information processing apparatus 1000 includes the following configuration as an example.
　　・ＣＰＵ（Ｃｅｎｔｒａｌ　Ｐｒｏｃｅｓｓｉｎｇ　Ｕｎｉｔ）１００１
　　・ＲＯＭ（Ｒｅａｄ　Ｏｎｌｙ　Ｍｅｍｏｒｙ）１００２
　　・ＲＡＭ（Ｒａｎｄｏｍ　Ａｃｃｅｓｓ　Ｍｅｍｏｒｙ）１００３
　　・ＲＡＭ１００３にロードされるプログラム１００４
　　・プログラム１００４を格納する記憶装置１００５
　　・記憶媒体１００６の読み書きを行うドライブ装置１００７
　　・通信ネットワーク１００９と接続する通信インターフェース１００８
　　・データの入出力を行う入出力インターフェース１０１０
　　・各構成要素を接続するバス１０１１
上述した情報処理装置１０００は、複数のＣＰＵ１００１を備えてもよい。ＣＰＵ１００１は、複数のプロセッサコアを含むマルチコアプロセッサであってもよい。各装置の実現方法には様々な変形例がある。例えば、各装置は、専用の装置として実現される。また、各装置は、複数の装置の組み合わせにより実現される。 CPU (Central Processing Unit) 1001
ROM (Read Only Memory) 1002
RAM (Random Access Memory) 1003
A program 1004 loaded into the RAM 1003
A storage device 1005 that stores the program 1004
A drive device 1007 that reads / writes data from / to the storage medium 1006
A communication interface 1008 connected to the communication network 1009
-I / O interface 1010 for inputting / outputting data
-Bus 1011 connecting each component
The information processing apparatus 1000 described above may include a plurality of CPUs 1001. The CPU 1001 may be a multi-core processor including a plurality of processor cores. There are various modifications in the method of realizing each device. For example, each device is realized as a dedicated device. Each device is realized by a combination of a plurality of devices.
　（第１の実施形態）
　まず、本発明の第１の実施形態について説明する。図１は、本発明の第１の実施形態におけるリソース割当システムを示す図である。図２は、本発明の第１の実施形態等におけるリソース割当システムとタスク実行サーバとの接続関係を示す図である。図３は、本発明の第１の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。 (First embodiment)
First, a first embodiment of the present invention will be described. FIG. 1 is a diagram showing a resource allocation system according to the first embodiment of the present invention. FIG. 2 is a diagram showing a connection relationship between the resource allocation system and the task execution server in the first embodiment of the present invention. FIG. 3 is a flowchart showing an example of the operation of the resource allocation system in the first exemplary embodiment of the present invention.
　図１に示すとおり、本発明の第１の実施形態におけるリソース割当システム１００は、負荷取得部１１０と、緊急度判定部１２０と、負荷判定部１３０と、リソース割当部１４０とを備える。負荷取得部１１０は、タスクを実行するサーバの負荷に関する情報を取得する。緊急度判定部１２０は、負荷に関する情報に基づいて、タスクに対するリソースの割当てに関する必要性を判定する。負荷判定部１３０は、緊急度判定部１２０がサーバの負荷の傾向を確認する必要があると判定した場合に、サーバの負荷の傾向を判定する。リソース割当部１４０は、緊急度判定部１２０がタスクにリソースを割当てる必要があると判定した場合に、タスクに対してリソースを割当てる。また、リソース割当システム１００は、測定値記憶部１５０を備えてもよい。測定値記憶部１５０は、負荷取得部１１０にて取得したサーバの負荷に関する情報を記憶する。 As shown in FIG. 1, the resource allocation system 100 according to the first embodiment of the present invention includes a load acquisition unit 110, an urgency determination unit 120, a load determination unit 130, and a resource allocation unit 140. The load acquisition unit 110 acquires information regarding the load on the server that executes the task. The urgency determining unit 120 determines the necessity for allocation of resources to tasks based on the information regarding the load. The load determination unit 130 determines the load tendency of the server when the urgency determination unit 120 determines that the load trend of the server needs to be confirmed. The resource allocation unit 140 allocates a resource to a task when the urgency determination unit 120 determines that a resource needs to be allocated to the task. In addition, the resource allocation system 100 may include a measurement value storage unit 150. The measured value storage unit 150 stores information related to the server load acquired by the load acquisition unit 110.
　本実施形態及び本発明の各実施形態におけるリソース割当システム１００は、一例として、図２（Ａ）に示すタスク実行サーバ１１において実行されるタスクに対するリソースの割当てを行う。リソース割当システム１００は、タスク実行サーバ１１と通信ネットワーク１２を介して接続される。また、タスク実行サーバ１１は、外部通信ネットワーク１３と接続される。 As an example, the resource allocation system 100 in this embodiment and each embodiment of the present invention allocates resources to tasks executed in the task execution server 11 shown in FIG. The resource allocation system 100 is connected to the task execution server 11 via the communication network 12. The task execution server 11 is connected to the external communication network 13.
　タスク実行サーバ１１は、例えば上述した情報処理装置１０００にて実現される。また、タスク実行サーバ１１では、ＯＳ（Ｏｐｅｒａｔｉｎｇ　Ｓｙｓｔｅｍ）や、ハイパーバイザによって管理される仮想機械（以下「ＯＳ等」とする）が動作する。タスク実行サーバ１１で動作するＯＳ等において、ユーザからの要求等に応じて任意のタスクが実行される。 The task execution server 11 is realized by the information processing apparatus 1000 described above, for example. In the task execution server 11, an OS (Operating System) and a virtual machine (hereinafter referred to as “OS etc.”) managed by a hypervisor operate. In an OS or the like that runs on the task execution server 11, an arbitrary task is executed in response to a request from the user.
　各々のタスクに対しては、予め定められた範囲でタスク実行サーバ１１等のリソースが割当てられる。各々のタスクに対して割り当てられるリソースは、例えば、タスク実行サーバ１１を実現する情報処理装置１０００のプロセッサコアやメモリ等がある。また、各々のタスクに対して割り当てられるリソースは、タスク実行サーバ１１と通信ネットワーク１２を介して接続される他の情報処理装置が備えるリソースであってもよい。また、タスクに対して割り当てられるリソースは、タスク実行サーバ等に対して予め行われた設定等に応じて変更される。 The resources such as the task execution server 11 are assigned to each task within a predetermined range. Examples of resources allocated to each task include a processor core and a memory of the information processing apparatus 1000 that implements the task execution server 11. Further, the resource allocated to each task may be a resource provided in another information processing apparatus connected to the task execution server 11 via the communication network 12. Further, the resource allocated to the task is changed according to a setting made in advance for the task execution server or the like.
　リソース割当システム１００は、図２（Ｂ）に示すように、タスク実行サーバ１１－１からタスク実行サーバ１１－ｎの各々において実行されるタスクをリソースの割当ての対象としてもよい。すなわち、リソース割当システム１００は、タスク実行サーバ１１－１からタスク実行サーバ１１－ｎの複数のタスク実行サーバにて実行されるタスクに対するリソースの割当てを行う。 As shown in FIG. 2 (B), the resource allocation system 100 may set a task executed in each of the task execution servers 11-1 to 11-n as a resource allocation target. That is, the resource allocation system 100 allocates resources to tasks executed by a plurality of task execution servers from the task execution server 11-1 to the task execution server 11-n.
　また、本発明の各実施形態におけるリソース割当システム１００、及びリソースの割当ての対象となるタスクは、一つの情報処理装置にて動作するＯＳ等において、それぞれ別のプロセスとして実行されるプログラムであってもよい。 Further, the resource allocation system 100 and the task to which resources are allocated in each embodiment of the present invention are programs that are executed as separate processes in an OS or the like that operates on one information processing apparatus. Also good.
　なお、本実施形態及び本発明の各実施形態において、リソース割当システム１００は、サーバの負荷としてタスクが実行されるサーバのＣＰＵの使用率を取得し、プロセッサコアをリソースとしてタスクに割当てることを想定する。 In this embodiment and each embodiment of the present invention, it is assumed that the resource allocation system 100 acquires the CPU usage rate of a server on which a task is executed as a server load and allocates a processor core as a resource to the task. To do.
　続いて、本発明の第１の実施形態におけるリソース割当システム１００の各構成要素について説明する。 Subsequently, each component of the resource allocation system 100 according to the first embodiment of the present invention will be described.
　負荷取得部１１０は、タスクを実行するサーバの負荷に関する情報を取得する。サーバの負荷に関する情報は、サーバが備えるリソースや、サーバが通信ネットワークを介して利用する外部のリソースに対する負荷の様子を任意の形式で表す情報である。このリソースは、タスクの実行に際して必要となるリソースである。サーバの負荷に関する情報には、サーバが備えるＣＰＵやメモリ、ネットワーク等のリソースの各々の使用率等が含まれる。また、サーバの負荷に関する情報は、上述した使用率等を取得した時刻や、使用率等を取得した間隔などを含んでもよい。以下、サーバの負荷に関する情報によって表されるサーバの負荷を、単に「サーバの負荷」又は「負荷」と表す場合がある。 The load acquisition unit 110 acquires information regarding the load on the server that executes the task. Information regarding the load on the server is information representing the state of the load on the resources included in the server and the external resources used by the server via the communication network in an arbitrary format. This resource is a resource that is required when executing a task. The information related to the server load includes the usage rate of each resource such as CPU, memory, and network provided in the server. Further, the information regarding the load on the server may include the time when the above-described usage rate is acquired, the interval at which the usage rate is acquired, and the like. Hereinafter, the server load represented by the information on the server load may be simply expressed as “server load” or “load”.
　負荷取得部１１０は、例えば図２（Ａ）に示すタスク実行サーバ１１に負荷の測定を指示して、タスク実行サーバ１１の負荷に関する情報を取得する。一例として、負荷取得部１１０は、予め定められた一定の時間間隔に基づいて、タスク実行サーバ１１の負荷に関する情報を取得する。本実施形態においては、負荷取得部１１０は、負荷に関する情報として、例えば、ＣＰＵの使用率、当該ＣＰＵ使用率の測定時刻、当該ＣＰＵ使用率の測定間隔等の情報を取得する。 The load acquisition unit 110, for example, instructs the task execution server 11 illustrated in FIG. 2A to measure the load, and acquires information regarding the load of the task execution server 11. As an example, the load acquisition unit 110 acquires information on the load of the task execution server 11 based on a predetermined time interval. In the present embodiment, the load acquisition unit 110 acquires information such as a CPU usage rate, a measurement time of the CPU usage rate, and a measurement interval of the CPU usage rate as information related to the load.
　負荷取得部１１０にて取得された負荷に関する情報は、後述する緊急度判定部１２０に通知される。また、負荷取得部１１０は、負荷に関する情報を測定値記憶部１５０に記憶させてもよい。測定値記憶部１５０に記憶される情報の詳細については後述する。 The information regarding the load acquired by the load acquisition unit 110 is notified to the urgency level determination unit 120 described later. Further, the load acquisition unit 110 may store information related to the load in the measurement value storage unit 150. Details of the information stored in the measured value storage unit 150 will be described later.
　緊急度判定部１２０は、負荷取得部１１０にて取得されたサーバの負荷に基づいて、サーバが備えるリソースのタスクに対する割当てが必要であるかを判定する。具体的には、緊急度判定部１２０は、負荷取得部１１０によって取得されたサーバの負荷を用いて、緊急度を求める。そして、緊急度に基づいて、サーバが備えるリソースのタスクに対する割当てが必要であるかを判定する。なお、緊急度判定部１２０は、これ以外の方法でサーバが備えるリソースのタスクに対する割当てが必要であるかを判定してもよい。 The urgency determination unit 120 determines whether it is necessary to allocate a resource task included in the server based on the server load acquired by the load acquisition unit 110. Specifically, the urgency level determination unit 120 obtains the urgency level using the server load acquired by the load acquisition unit 110. Then, based on the urgency level, it is determined whether it is necessary to allocate the resource task of the server. Note that the urgency determining unit 120 may determine whether it is necessary to allocate a resource task included in the server by other methods.
　緊急度は、サーバが備えるリソースのタスクに対する割当てに関する必要性の程度を表す。緊急度は、例えば、サーバの負荷と、負荷の傾向とに基づいて求められる。過去の時点におけるサーバの負荷を用いて緊急度を求める場合には、緊急度判定部１２０は、測定値記憶部１５０に記憶された情報を必要に応じて読み出して求めてもよい。 The urgency level indicates the degree of necessity related to allocation of resources provided to the server for tasks. The urgency is obtained based on, for example, the load on the server and the load trend. When the urgency level is obtained using the load on the server at the past time point, the urgency level determination unit 120 may read and obtain the information stored in the measurement value storage unit 150 as necessary.
　負荷取得部１１０が一定の時間間隔で負荷を取得する場合において、緊急度判定部１２０は、具体的な一例として緊急度を以下のように求める。この場合に、緊急度を求める時点において直近となる時点に測定された負荷をＷとする。また、Ｗと、当該時点より一つ前の時点に測定された負荷との差をΔＷとする。そして、ａ及びｂを予め定められた正の係数とする。この場合に、緊急度Ｅは、以下の（１）式の形で求められる。 When the load acquisition unit 110 acquires a load at regular time intervals, the urgency level determination unit 120 calculates the urgency level as follows as a specific example. In this case, let W be the load measured at the most recent time when the degree of urgency is obtained. In addition, a difference between W and a load measured at a time point immediately before the time point is denoted by ΔW. Then, a and b are predetermined positive coefficients. In this case, the urgency E is obtained in the form of the following equation (1).
Figure JPOXMLDOC01-appb-M000001
Figure JPOXMLDOC01-appb-M000001
　すなわち、緊急度は、緊急度を求める時点におけるサーバの負荷と、当該時点における負荷の傾向に基づいて求められる。ａ及びｂは、緊急度を求める時点におけるサーバの負荷と、当該時点における負荷の傾向とのいずれが重要であるか等に応じて適宜定められる。また、上述の（１）式において、負荷の傾向は、緊急度を求める時点において直近となる時点と、当該時点から一つ前の時点において測定された負荷を用いて求められる。しかしながら、負荷の傾向は、緊急度を求める時点において直近となる時点から二つ以上前の時点において測定された負荷を用いてもよい。 That is, the urgency is obtained based on the load on the server at the time of obtaining the urgency and the load trend at the time. a and b are appropriately determined according to which one of the load of the server at the time of obtaining the degree of urgency and the load tendency at the time is important. Further, in the above equation (1), the load tendency is obtained by using the load measured at the time immediately before the time when the urgency is obtained and the load measured immediately before the time. However, the load trend may be a load measured at a time two or more times before the most recent time when the urgency level is obtained.
　なお、緊急度判定部１２０は、上述した直近となる時点より前の時点において測定された負荷を予め保持し、緊急度を求める際に必要に応じて読み出してもよい。または、緊急度判定部１２０は、上負荷取得部１１０が直近より前の時点に測定された負荷を保持してもよい。この場合には、緊急度判定部１２０は、緊急度を求める際に必要に応じて直近より前の時点に測定された負荷取得部１１０から取得する。また、測定値記憶部１５０が備えられている場合には、緊急度判定部１２０は、測定値記憶部１５０に記憶されている、直近より前の時点に測定された負荷を必要に応じて読み出して用いてもよい。 Note that the urgency level determination unit 120 may hold in advance the load measured at a time point before the most recent time point described above, and read it as necessary when determining the urgency level. Or the urgency determination part 120 may hold | maintain the load measured by the upper load acquisition part 110 before the latest. In this case, the urgency level determination unit 120 acquires the urgency level from the load acquisition unit 110 measured at a time point immediately before the latest as necessary. Further, when the measurement value storage unit 150 is provided, the urgency level determination unit 120 reads, as necessary, the load measured in the measurement value storage unit 150 and measured before the most recent time. May be used.
　緊急度判定部１２０は、緊急度に基づいて、タスクに対するリソースの割当てが必要かを判定する。 The urgency level determination unit 120 determines whether resources need to be allocated to tasks based on the urgency level.
　本実施形態においては、緊急度判定部１２０は、二つの基準によってタスクに対するリソースの割当てが必要かを判定する。すなわち、緊急度判定部１２０は、リソースの割当ての必要性が高い場合に、タスクに対してリソースを割当てるべき状態であると判定する。また、緊急度判定部１２０は、リソースの割当ての必要性が高いが、タスクに対してリソースを割当てるべき程度ではない場合に、サーバの負荷の傾向を更に確認する必要があると判定する。 In the present embodiment, the urgency determination unit 120 determines whether it is necessary to allocate resources to tasks according to two criteria. In other words, the urgency determination unit 120 determines that a resource should be allocated to a task when the necessity for resource allocation is high. Further, the urgency determination unit 120 determines that it is necessary to further check the load trend of the server when the resource allocation is highly necessary but the resource is not allocated to the task.
　本実施形態では、緊急度判定部１２０は、緊急度が第１及び第２の条件を満たすか否かを判定して、上述した二つの基準によってタスクに対するリソースの割当ての必要性を判定する。第１の条件は、リソースの割当ての必要性が高く、タスクに対してリソースを割当てるべき状態であることを示す条件である。また、第２の条件は、リソースの割当ての必要性が高いが、タスクに対してリソースを割当てるべき程度ではない場合に、サーバの負荷の傾向を更に確認する必要があることを示す条件である。 In the present embodiment, the urgency determination unit 120 determines whether the urgency satisfies the first and second conditions, and determines the necessity of resource allocation for the task based on the two criteria described above. The first condition is a condition indicating that there is a high necessity for resource allocation and that a resource should be allocated to a task. The second condition is a condition that indicates that it is necessary to further check the tendency of the load on the server when the resource allocation is high but the resource should not be allocated to the task. .
　本実施形態においては、第１及び第２の条件は、緊急度として求められた値に対する閾値として表される。例えば、緊急度Ｅが上述の（１）式にて求められる場合には、第１の条件に関する閾値としてＥ１が、第２の条件に関する閾値としてＥ２が定められる。そして、緊急度判定部１２０は、緊急度がＥ１≦Ｅを満たす（すなわち、緊急度Ｅの大きさがＥ１以上である）場合に、緊急度が第１の条件を満たすと判定する。また、緊急度判定部１２０は、緊急度ＥがＥ２≦Ｅ＜Ｅ１を満たす（すなわち、緊急度Ｅの大きさがＥ２以上であり、かつ、Ｅ１よりも小さい）場合に、緊急度が第２の条件を満たすと判定する。緊急度がＥ＜Ｅ２を満たす（すなわち、緊急度の大きさがＥ２よりも小さい）場合には、緊急度判定部１２０は、緊急度が第１及び第２のいずれの条件も満たさないと判定する。 In the present embodiment, the first and second conditions are expressed as threshold values for the value obtained as the degree of urgency. For example, when the degree of urgency E is obtained by the above-described equation (1), E1 is set as the threshold value related to the first condition, and E2 is set as the threshold value related to the second condition. Then, the urgency level determination unit 120 determines that the urgency level satisfies the first condition when the urgency level satisfies E1 ≦ E (that is, the level of the urgency level E is equal to or greater than E1). Further, the urgency level determination unit 120 determines that the urgency level is the second level when the urgency level E satisfies E2 ≦ E <E1 (that is, the magnitude of the urgency level E is equal to or larger than E2 and smaller than E1). It is determined that the above condition is satisfied. When the urgency level satisfies E <E2 (that is, the urgency level is smaller than E2), the urgency level determination unit 120 determines that the urgency level does not satisfy both the first and second conditions. To do.
　後述する負荷判定部１３０又はリソース割当部１４０は、緊急度判定部１２０における判定の結果に応じて動作を行う。つまり、緊急度判定部１２０は、緊急度が第１又は第２の条件を満たすと判定した場合には、負荷判定部１３０又はリソース割当部１４０の各々は、緊急度判定部１２０が満たすと判定した条件に応じて必要な動作を行う。 The load determination unit 130 or the resource allocation unit 140, which will be described later, operates according to the determination result in the urgency determination unit 120. That is, when the urgency level determination unit 120 determines that the urgency level satisfies the first or second condition, each of the load determination unit 130 and the resource allocation unit 140 determines that the urgency level determination unit 120 satisfies the condition. Perform the necessary operations according to the conditions.
　負荷判定部１３０は、リソースの割当ての必要性が高いが、タスクに対してリソースを割当てるべき程度ではない旨が緊急度判定部１２０によって判定された場合に、サーバの負荷の傾向を判定する。すなわち、負荷判定部１３０は、緊急度が第２の条件を満たす場合に、サーバの負荷の傾向を判定する。負荷判定部１３０は、例えば、負荷取得部１１０にて取得された負荷を、予め定められたリソース割当実行判定式に当てはめることで、負荷の傾向を判定する。リソース割当実行判定式は、一つ以上の任意の数の負荷の測定値を引数として取得し、リソースの割当てが必要か否かを表す判定結果を示す式である。ただし、負荷判定部１３０は、これ以外の任意の手法に基づいてサーバの負荷の傾向を判定してもよい。 The load determination unit 130 determines the load tendency of the server when the urgency determination unit 120 determines that resource allocation is high but the resource should not be allocated to the task. That is, the load determination unit 130 determines the load tendency of the server when the degree of urgency satisfies the second condition. For example, the load determination unit 130 determines the load tendency by applying the load acquired by the load acquisition unit 110 to a predetermined resource allocation execution determination formula. The resource allocation execution determination expression is an expression indicating a determination result indicating whether or not resource allocation is necessary by acquiring one or more arbitrary measured values of loads as an argument. However, the load determination unit 130 may determine the load tendency of the server based on any other method.
　負荷判定部１３０は、負荷の傾向の判定に際して、必要に応じて、測定値記憶部１５０に記憶された負荷に関する情報を参照してもよい。また、負荷判定部１３０は、複数の時点における負荷に関する情報を用いて負荷の傾向を判定してもよい。この場合には、負荷判定部１３０は、測定値記憶部１５０に記憶された負荷に関する情報を参照する等して、緊急度が第２の条件を満たすと判定された時点より前の時点における既に計測された負荷に関する情報を用いてもよい。また、負荷判定部１３０は、当該時点より後の時点における負荷に関する情報を用いてもよい。 The load determination unit 130 may refer to information on the load stored in the measured value storage unit 150 as necessary when determining the load tendency. In addition, the load determination unit 130 may determine a load tendency using information regarding loads at a plurality of points in time. In this case, the load determination unit 130 refers to the information on the load stored in the measurement value storage unit 150, for example, and has already been at a time before the time when the urgency is determined to satisfy the second condition. Information regarding the measured load may be used. Moreover, the load determination part 130 may use the information regarding the load in the time after the said time.
　そして、負荷判定部１３０は、負荷が上昇する傾向にあると判定した場合に、タスクに対してリソースを割当てることを決定してもよい。負荷判定部１３０がタスクに対してリソースを割当てることを決定した場合には、リソース割当部１４０は、後述のようにリソースを割当てる動作を行う。負荷判定部１３０は、負荷が上昇する傾向にあると判定しない場合には、予め定めた任意の条件等に応じて、負荷の傾向の判定を終了してもよいし、引き続き負荷の傾向の判定を継続してもよい。 Then, when the load determination unit 130 determines that the load tends to increase, the load determination unit 130 may determine to allocate resources to the task. When the load determination unit 130 determines to allocate a resource to a task, the resource allocation unit 140 performs an operation of allocating a resource as will be described later. When the load determination unit 130 does not determine that the load tends to increase, the load determination may be terminated according to any predetermined condition or the like, or the determination of the load trend may continue. May be continued.
　リソース割当部１４０は、緊急度判定部１２０によってリソースの割当ての必要性が高いと判定された場合に、タスクに対してリソースを割当てる動作を行う。すなわち、リソース割当部１４０は、緊急度が上述した第１の条件を満たす場合、タスクに対してリソースを割当てる動作を行う。リソース割当部１４０は、例えば、図２に示すタスク実行サーバ１１等のようなタスクを実行するサーバに対して、リソースの割当てを要求する。また、リソース割当部１４０は、負荷判定手段において負荷が上昇する傾向にあると判定された場合にも、タスクに対してリソースを割当てる動作を行う。 The resource allocation unit 140 performs an operation of allocating a resource to a task when the urgency level determination unit 120 determines that the necessity for resource allocation is high. That is, the resource allocation unit 140 performs an operation of allocating resources to a task when the urgency satisfies the first condition described above. The resource allocation unit 140 requests resource allocation to a server that executes a task such as the task execution server 11 shown in FIG. The resource allocation unit 140 also performs an operation of allocating resources to tasks even when the load determination unit determines that the load tends to increase.
　割当ての対象となるリソースは、例えば１個以上のプロセッサコア等がある。ただし、リソース割当部１４０は、タスク実行サーバ１１等が備える任意のリソースを割当ててもよい。リソース割当部１４０は、タスク実行サーバ１１等のタスクが実行されるサーバに置いて割当てが可能となるリソースの情報を予め保持してもよい。 The resource to be allocated includes, for example, one or more processor cores. However, the resource allocation unit 140 may allocate an arbitrary resource included in the task execution server 11 or the like. The resource allocation unit 140 may hold in advance information on resources that can be allocated on a server on which a task is executed, such as the task execution server 11.
　続いて、図３に示すフローチャートを参照して、本発明の第１の実施形態におけるリソース割当システム１００の動作の一例について説明する。図３に示すフローチャートにおいては、緊急度判定部１２０は、タスクに対するリソースの割当てに関する必要性の判定として、緊急度が上述した第１又は第２の条件のいずれかを満たすか否かを判定する。 Subsequently, an example of the operation of the resource allocation system 100 according to the first embodiment of the present invention will be described with reference to the flowchart shown in FIG. In the flowchart shown in FIG. 3, the urgency level determination unit 120 determines whether the urgency level satisfies one of the first or second conditions described above as the determination of the necessity regarding the resource allocation to the task. .
　最初に、負荷取得部１１０は、タスクを実行するサーバの負荷に関する情報を取得する（ステップＳ１０１）。 First, the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S101).
　次に、緊急度判定部１２０は、緊急度を求める（ステップＳ１０２）。続いて、緊急度判定部１２０は、タスクに対するリソースの割当てに関する必要性を判定する。まず、緊急度判定部１２０は、緊急度が先に説明した上述した第１の条件を満たすかを判定する（ステップＳ１０３）。 Next, the urgency level determination unit 120 obtains the urgency level (step S102). Subsequently, the urgency determining unit 120 determines the necessity regarding the allocation of resources to the tasks. First, the urgency level determination unit 120 determines whether the urgency level satisfies the first condition described above (step S103).
　緊急度が上述した第１の条件を満たす場合（ステップＳ１０３：Ｙｅｓ）には、緊急度判定部１２０は、リソースの割当ての必要性が高く、タスクに対してリソースを割当てるべき状態であると判定する。この判定に応じて、リソース割当部１４０は、プロセッサコア等のリソースをタスクに対して割当てる（ステップＳ１０４）。 When the urgency level satisfies the first condition described above (step S103: Yes), the urgency level determination unit 120 determines that resource allocation is highly necessary and resources should be allocated to the task. To do. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S104).
　緊急度判定部１２０は、緊急度が上述した第１の条件を満たさない場合（ステップＳ１０３：Ｎｏ）に、緊急度が上述した第２の条件を満たすかを判定する（ステップＳ１０５）。緊急度が第２の条件を満たす（ステップＳ１０５：Ｙｅｓ）場合には、緊急度判定部１２０は、サーバの負荷の傾向を更に確認する必要があると判定する。この判定に応じて、負荷判定部１３０は、負荷傾向を測定する（ステップＳ１０６）。負荷判定部１３０は、測定した負荷傾向に応じて、リソース割当部１４０に対してリソースの割当てを行うことを決定してもよい。この場合には、リソース割当部１４０は、プロセッサコア等のリソースをタスクに対して割当てる。 The urgency level determination unit 120 determines whether the urgency level satisfies the above-described second condition when the urgency level does not satisfy the above-described first condition (step S103: No) (step S105). If the urgency level satisfies the second condition (step S105: Yes), the urgency level determination unit 120 determines that it is necessary to further check the load tendency of the server. In response to this determination, the load determination unit 130 measures a load tendency (step S106). The load determination unit 130 may determine to allocate resources to the resource allocation unit 140 according to the measured load tendency. In this case, the resource allocation unit 140 allocates a resource such as a processor core to the task.
　なお、緊急度判定部１２０において緊急度が第１及び第２のいずれの条件も満たさないと判定された場合（ステップＳ１０５：Ｎｏ）には、リソース割当システム１００は、動作を終了する。なお、負荷取得部１１０が定期的にサーバの負荷に関する情報を取得する場合には、リソース割当システム１００は、図３に示すフローチャートの動作を繰り返して実行してもよい。 Note that if the urgency level determination unit 120 determines that the urgency level does not satisfy both the first and second conditions (step S105: No), the resource allocation system 100 ends the operation. Note that when the load acquisition unit 110 periodically acquires information regarding the load on the server, the resource allocation system 100 may repeatedly execute the operation of the flowchart illustrated in FIG. 3.
　以上のとおり、本発明の第１の実施形態におけるリソース割当システム１００は、緊急度判定部１２０にて求められた緊急度に応じてリソースの割当てを行う。 As described above, the resource allocation system 100 according to the first embodiment of the present invention performs resource allocation according to the urgency level obtained by the urgency level determination unit 120.
　具体的には、緊急度が高いと判定される場合（すなわち、緊急度が上述した第１の条件を満たす場合）には、リソース割当システム１００は、更なる判定を行わずにリソースの割当てを行う。緊急度の高さが中程度であると判定される場合（すなわち、緊急度が上述した第２の条件を満たす場合）には、リソース割当システム１００は、更に負荷の傾向を判定する。そして、負荷が上昇する傾向にあると判定された場合には、リソース割当システム１００は、必要に応じてリソースの割当てを行う場合がある。 Specifically, when it is determined that the urgency level is high (that is, when the urgency level satisfies the first condition described above), the resource allocation system 100 allocates resources without performing further determination. Do. When it is determined that the level of urgency is medium (that is, when the level of urgency satisfies the second condition described above), the resource allocation system 100 further determines the load tendency. When it is determined that the load tends to increase, the resource allocation system 100 may perform resource allocation as necessary.
　すなわち、リソース割当システム１００は、急激に負荷が上昇した場合等、負荷が高い場合には迅速にリソースを割当てることが可能である。また、リソース割当システム１００は、緊急度の高さが中程度である場合に、リソースの割当てに先立ち、負荷の傾向を判定してリソースの割当てを行う。このようにすることで、一時的な負荷の変動に対して過敏に反応したリソースの割当てが回避される。更に、リソース割当システム１００による上述したリソースの割当てに際して、リソースの割当て等に関して事前のポリシーの設定は不要である。 That is, the resource allocation system 100 can quickly allocate resources when the load is high, such as when the load suddenly increases. Further, when the degree of urgency is medium, the resource allocation system 100 determines a load tendency and allocates resources prior to resource allocation. In this way, the allocation of resources that react sensitively to temporary load fluctuations is avoided. Further, when the above-described resource allocation is performed by the resource allocation system 100, it is not necessary to set a policy in advance regarding resource allocation or the like.
　したがって、本実施形態におけるリソース割当システム１００は、負荷の変動に応じた適切なリソースの割当てを可能にする。 Therefore, the resource allocation system 100 according to the present embodiment enables appropriate resource allocation according to load fluctuations.
　（第２の実施形態）
　次に、本発明の第２の実施形態について説明する。図４は、本発明の第２の実施形態におけるリソース割当システムを示す図である。図５は、測定値記憶部にて記憶される負荷に関する情報の一例を示す図である。図６は、本発明の第２の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。図７は、本発明の第２の実施形態におけるリソース割当システムに含まれる傾向判定部の動作の一例を示すフローチャートである。 (Second Embodiment)
Next, a second embodiment of the present invention will be described. FIG. 4 is a diagram showing a resource allocation system in the second exemplary embodiment of the present invention. FIG. 5 is a diagram illustrating an example of information related to a load stored in the measurement value storage unit. FIG. 6 is a flowchart showing an example of the operation of the resource allocation system in the second exemplary embodiment of the present invention. FIG. 7 is a flowchart illustrating an example of the operation of the tendency determination unit included in the resource allocation system according to the second embodiment of the present invention.
　図４に示すとおり、本発明の第２の実施形態におけるリソース割当システム２００は、負荷取得部１１０と、緊急度判定部１２０と、間隔調整部２３１と、傾向判定部２３２と、リソース割当部１４０とを備える。負荷取得部１１０、緊急度判定部１２０及びリソース割当部１４０は、以下に説明がある箇所を除き、基本的に、本発明の第１の実施形態におけるリソース割当システム１００が備える同じ構成要素と同様の動作を行う。間隔調整部２３１は、負荷取得部１１０による負荷に関する情報の取得間隔を変更する。傾向判定部２３２は、間隔調整部２３１が負荷取得部１１０による情報の取得間隔を短くした場合における負荷の傾向に基づいて、タスクに対してリソースを割当てるかを判定する。傾向判定部２３２は、負荷判定部１３０の一つの例である。また、リソース割当システム２００は、負荷取得部１１０にて取得したサーバの負荷に関する計測値を記憶する測定値記憶部１５０を備えてもよい。 As illustrated in FIG. 4, the resource allocation system 200 according to the second exemplary embodiment of the present invention includes a load acquisition unit 110, an urgency determination unit 120, an interval adjustment unit 231, a tendency determination unit 232, and a resource allocation unit 140. With. The load acquisition unit 110, the urgency determination unit 120, and the resource allocation unit 140 are basically the same as the same constituent elements included in the resource allocation system 100 according to the first embodiment of the present invention, except for the following description. Perform the operation. The interval adjustment unit 231 changes the acquisition interval of information related to the load by the load acquisition unit 110. The trend determination unit 232 determines whether to allocate resources to the task based on the load trend when the interval adjustment unit 231 shortens the information acquisition interval by the load acquisition unit 110. The tendency determination unit 232 is an example of the load determination unit 130. Further, the resource allocation system 200 may include a measurement value storage unit 150 that stores measurement values related to the server load acquired by the load acquisition unit 110.
　すなわち、本実施形態におけるリソース割当システム２００は、間隔調整部２３１を備え、更に負荷判定部１３０として傾向判定部２３２を備える点が、第１の実施形態におけるリソース割当システム１００と異なる。それ以外の点については、本実施形態におけるリソース割当システム２００は、第１の実施形態におけるリソース割当システム１００と基本的に同様の構成を備える。 That is, the resource allocation system 200 according to the present embodiment is different from the resource allocation system 100 according to the first embodiment in that the resource allocation system 200 includes an interval adjustment unit 231 and further includes a tendency determination unit 232 as the load determination unit 130. In other respects, the resource allocation system 200 in this embodiment has basically the same configuration as the resource allocation system 100 in the first embodiment.
　なお、本実施形態におけるリソース割当システム２００がリソースの割当ての対象とするタスクは、本発明の第１の実施形態におけるリソース割当システム１００がリソースの割当ての対象とするタスクと同様である。 It should be noted that the task to which the resource allocation system 200 in this embodiment is to be assigned a resource is the same as the task to which the resource allocation system 100 in the first embodiment of this invention is to be assigned a resource.
　続いて、本発明の第２の実施形態におけるリソース割当システム２００の各構成要素について説明する。 Subsequently, each component of the resource allocation system 200 according to the second embodiment of the present invention will be described.
　本実施形態においては、負荷取得部１１０は、予め定められた間隔でタスクを実行するサーバの負荷に関する情報を取得する。また、この間隔は、間隔調整部２３１の指示に基づいて変更される場合がある。 In the present embodiment, the load acquisition unit 110 acquires information related to the load of the server that executes a task at a predetermined interval. In addition, this interval may be changed based on an instruction from the interval adjusting unit 231.
　例えば、緊急度判定部１２０が、リソースの割当てに関する必要性の判定として、第２の条件を満たすと判定した場合には、傾向判定部２３２からの指示に基づいて、負荷に関する情報を取得する間隔は、短い間隔に変更される。また、傾向判定部２３２による判定結果に基づいて、負荷に関する情報を取得する間隔は、予め定められた当初の間隔に変更される。 For example, when the urgency determination unit 120 determines that the second condition is satisfied as the determination of the necessity for resource allocation, the interval at which the information on the load is acquired based on the instruction from the trend determination unit 232 Is changed to a short interval. Moreover, based on the determination result by the tendency determination unit 232, the interval for acquiring the information regarding the load is changed to a predetermined initial interval.
　なお、先に説明したように、負荷取得部１１０にて取得された負荷に関する情報は、測定値記憶部１５０に記憶される場合がある。測定値記憶部１５０は、負荷に関する情報を、例えば図５に示すテーブルの形式にて記憶する。図５に示すテーブルでは、ＣＰＵの負荷が、測定時刻及び測定時刻間隔と共に記憶される。なお、測定時刻間隔は、ＣＰＵの負荷が測定された一つの測定時刻と、その前にＣＰＵの負荷が測定された測定時刻との間隔を示す。 In addition, as described above, information regarding the load acquired by the load acquisition unit 110 may be stored in the measurement value storage unit 150. The measured value storage unit 150 stores information on the load, for example, in the form of a table shown in FIG. In the table shown in FIG. 5, the CPU load is stored together with the measurement time and the measurement time interval. The measurement time interval indicates an interval between one measurement time when the CPU load is measured and the measurement time when the CPU load is measured before that.
　間隔調整部２３１は、緊急度判定部１２０において緊急度が第２の条件を満たすと判定された場合に、負荷取得部１１０による負荷に関する情報の取得間隔を変更する。本実施形態においては、間隔調整部２３１は、負荷取得部１１０による負荷に関する情報の取得間隔を、通常の時間間隔と、通常の時間間隔と比較して短い時間間隔との２つの時間間隔のいずれかに設定する。 The interval adjustment unit 231 changes the acquisition interval of the information regarding the load by the load acquisition unit 110 when the urgency determination unit 120 determines that the urgency level satisfies the second condition. In the present embodiment, the interval adjusting unit 231 sets the acquisition interval of the information related to the load by the load acquiring unit 110 as one of two time intervals, a normal time interval and a time interval that is shorter than the normal time interval. Set it.
　具体的には、間隔調整部２３１は、緊急度が第２の条件を満たす場合に、負荷取得部１１０による負荷に関する情報の取得間隔を短くするように設定する。また、間隔調整部２３１は、後述する傾向判定部２３２の判定結果に基づいて、負荷に関する情報の取得間隔を通常の時間間隔とするように負荷取得部１１０の動作を設定してもよい。 Specifically, the interval adjusting unit 231 sets the acquisition interval of the information regarding the load by the load acquiring unit 110 to be shortened when the urgency satisfies the second condition. Further, the interval adjustment unit 231 may set the operation of the load acquisition unit 110 based on the determination result of the tendency determination unit 232 described later so that the acquisition interval of information related to the load is a normal time interval.
　傾向判定部２３２は、負荷に関する情報の取得間隔が短くされた場合における負荷の傾向に基づいて、タスクに対してリソースを割当てるかを判定する。すなわち、傾向判定部２３２は、負荷が増加傾向にある場合には、タスクに対してリソースを割当てると判定する。具体的には、負荷に関する情報の取得間隔が短くされた場合において、傾向判定部２３２は、負荷が所定の測定回数に亘って継続して増加している場合に、負荷が増加傾向にあるとして、タスクに対してリソースを割当てると判定する。 The trend determination unit 232 determines whether to allocate resources to the task based on the load trend when the load information acquisition interval is shortened. That is, when the load tends to increase, the trend determination unit 232 determines to allocate resources to the task. Specifically, when the acquisition interval of information regarding the load is shortened, the tendency determination unit 232 assumes that the load is increasing when the load continues to increase over a predetermined number of measurements. It is determined that resources are allocated to the task.
　また、傾向判定部２３２は、負荷が減少傾向にある場合には、タスクに対してリソースを割当てないと判定し、かつ、負荷取得部１１０による負荷に関する情報の取得間隔を通常の時間間隔に変更してもよい。 In addition, when the load tends to decrease, the trend determination unit 232 determines that resources are not allocated to the task, and changes the load acquisition interval of information related to the load by the load acquisition unit 110 to a normal time interval. May be.
　具体的には、負荷に関する情報の取得間隔が短くされた場合において、傾向判定部２３２は、負荷が所定の測定回数に亘って継続して減少している場合に、負荷が減少傾向にあるとして、タスクに対してリソースを割当てないと判定する。そして、傾向判定部２３２は、負荷取得部１１０による負荷に関する情報の取得間隔を通常の間隔に変更する。 Specifically, when the acquisition interval of the information related to the load is shortened, the tendency determination unit 232 assumes that the load tends to decrease when the load continues to decrease over a predetermined number of measurements. It is determined that no resource is allocated to the task. And the tendency determination part 232 changes the acquisition interval of the information regarding the load by the load acquisition part 110 to a normal space | interval.
　一例として、取得間隔が短くされた場合における負荷に関する情報の測定値が測定値記憶部１５０に記憶される場合に、傾向判定部２３２は次のように動作する。 As an example, when the measurement value of the information regarding the load when the acquisition interval is shortened is stored in the measurement value storage unit 150, the tendency determination unit 232 operates as follows.
　この場合には、傾向判定部２３２は、測定値記憶部１５０の負荷を記憶するテーブルから、直近の所定の回数の測定結果を読み出す。そして、傾向判定部２３２は、測定結果として記録された複数の時点における負荷の全てが、その一つ前の時点に測定された負荷と比較して増加している場合に、負荷が増加傾向にあると判定する。傾向判定部２３２は、負荷が増加傾向にあると判定した場合には、タスクに対するリソースの割当てを決定する。リソース割当部１４０は、当該決定に応じて、タスクに対してリソースを割当てる。 In this case, the tendency determination unit 232 reads the latest predetermined number of measurement results from the table storing the load of the measurement value storage unit 150. The trend determination unit 232 then increases the load when all of the loads at a plurality of times recorded as measurement results are increased compared to the load measured at the previous time. Judge that there is. If the trend determination unit 232 determines that the load tends to increase, the trend determination unit 232 determines resource allocation to the task. The resource allocation unit 140 allocates resources to tasks according to the determination.
　また、この場合において、傾向判定部２３２は、測定結果として記録された複数の時点における負荷の全てが、その一つ前の時点に測定された負荷と比較して減少している場合に、負荷が減少傾向にあると判定する。そして、傾向判定部２３２は、タスクに対してリソースを割当てない旨を決定する。間隔調整部２３１は、この決定に応じて、負荷取得部１１０による負荷に関する情報の取得間隔を通常の間隔に変更する。 Further, in this case, the tendency determination unit 232 determines that the load is reduced when all of the loads at a plurality of time points recorded as measurement results are reduced compared to the load measured at the previous time point. Is determined to be decreasing. Then, the trend determination unit 232 determines that no resource is allocated to the task. In response to this determination, the interval adjustment unit 231 changes the acquisition interval of information regarding the load by the load acquisition unit 110 to a normal interval.
　なお、傾向判定部２３２は、負荷に関する情報の測定値が上述した場合とは異なる場合には、その時点においてリソースを割当てるか否かの判定を行わなくてもよい。この場合には、傾向判定部２３２は、新たに負荷に関する情報の測定値を取得する等して、当該新たに取得した負荷の傾向に基づいて、改めてリソースを割当てるか否かの判定を行う。 In addition, when the measured value of the information regarding the load is different from the case described above, the tendency determination unit 232 does not need to determine whether to allocate resources at that time. In this case, the tendency determination unit 232 newly acquires a measurement value of information related to the load, and determines whether to allocate resources again based on the newly acquired load tendency.
　また、傾向判定部２３２は、上述した動作とは異なる動作や基準によって、タスクに対してリソースを割当てるかを判定してもよい。 Also, the trend determination unit 232 may determine whether to allocate resources to tasks based on operations and criteria different from the operations described above.
　続いて、図６に示すフローチャートを参照して、本発明の第２の実施形態におけるリソース割当システム２００の動作の一例について説明する。 Subsequently, an example of the operation of the resource allocation system 200 according to the second embodiment of the present invention will be described with reference to the flowchart shown in FIG.
　最初に、負荷取得部１１０は、タスクを実行するサーバの負荷に関する情報を取得する（ステップＳ２０１）。次に、緊急度判定部１２０は、負荷取得部１１０における負荷に関する情報の取得間隔として、通常の時間間隔が設定されているかを判定する（ステップＳ２０２）。 First, the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S201). Next, the urgency level determination unit 120 determines whether a normal time interval is set as an acquisition interval of information regarding the load in the load acquisition unit 110 (step S202).
　負荷取得部１１０に対して、負荷に関する情報の取得間隔として通常の時間間隔が設定されている場合（ステップＳ２０２：Ｙｅｓ）には、緊急度判定部１２０は、緊急度を求める（ステップＳ２０３）。そして、緊急度判定部１２０は、タスクに対するリソースの割当てに関する必要性を判定する。緊急度判定部１２０は、緊急度が第１の条件を満たすかを判定する（ステップＳ２０４）。 When a normal time interval is set as an acquisition interval of information related to the load with respect to the load acquisition unit 110 (step S202: Yes), the urgency level determination unit 120 obtains an urgency level (step S203). Then, the urgency determining unit 120 determines the necessity regarding the allocation of resources to the tasks. The urgency determination unit 120 determines whether the urgency satisfies the first condition (step S204).
　ステップＳ２０３にて求めた緊急度が第１の条件を満たす場合（ステップＳ２０４：Ｙｅｓ）には、緊急度判定部１２０は、タスクに対するリソースの割当てを決定する。この決定に応じて、リソース割当部１４０は、プロセッサコア等のリソースをタスクに対して割当てる（ステップＳ２０５）。 When the urgency level obtained in step S203 satisfies the first condition (step S204: Yes), the urgency level determination unit 120 determines resource allocation to the task. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S205).
　緊急度判定部１２０は、緊急度が先に説明した第１の条件を満たさない場合（ステップＳ２０４：Ｎｏ）に、緊急度が先に説明した第２の条件を満たすかを判定する（ステップＳ２０６）。緊急度が第２の条件を満たす（ステップＳ２０５：Ｙｅｓ）場合には、間隔調整部２３１は、負荷取得部１１０における負荷に関する情報の取得間隔を通常の時間間隔と比較して短い時間間隔とするように設定する（ステップＳ２０７）。ステップＳ２０７における処理が行われると、リソース割当システム２００は、動作を終了する。 When the urgency level does not satisfy the first condition described above (step S204: No), the urgency level determination unit 120 determines whether the urgency level satisfies the second condition described above (step S206). ). When the degree of urgency satisfies the second condition (step S205: Yes), the interval adjusting unit 231 sets the acquisition interval of the information regarding the load in the load acquiring unit 110 to a short time interval compared to the normal time interval. (Step S207). When the process in step S207 is performed, the resource allocation system 200 ends the operation.
　なお、ステップＳ２０４、Ｓ２０５及びＳ２０６の動作は、本発明の第１の実施形態におけるステップＳ１０３、Ｓ１０４及びＳ１０５とそれぞれ同様の動作を行う。 The operations in steps S204, S205, and S206 are the same as those in steps S103, S104, and S105 in the first embodiment of the present invention.
　負荷取得部１１０に対して、負荷に関する情報の取得間隔として通常の時間間隔が設定されていない場合（ステップＳ２０２：Ｎｏ）には、負荷に関する情報の取得間隔として、短い時間間隔が設定されていると想定される。この場合には、傾向判定部２３２は、負荷傾向を判定する（ステップＳ２０８）。ステップＳ２０８における処理は後述する。ステップＳ２０８における処理が行われると、リソース割当システム２００は、動作を終了する。 When a normal time interval is not set as the load information acquisition interval for the load acquisition unit 110 (step S202: No), a short time interval is set as the load information acquisition interval. It is assumed. In this case, the tendency determination unit 232 determines a load tendency (step S208). The process in step S208 will be described later. When the process in step S208 is performed, the resource allocation system 200 ends the operation.
　なお、負荷取得部１１０が定期的にサーバの負荷に関する情報を取得する場合には、リソース割当システム２００は、図６に示すフローチャートの動作を繰り返して行ってもよい。 In addition, when the load acquisition unit 110 periodically acquires information related to the server load, the resource allocation system 200 may repeatedly perform the operation of the flowchart illustrated in FIG.
　なお、負荷取得部１１０が定期的にサーバの負荷に関する情報を取得する場合には、リソース割当システム２００は、負荷取得部１１０が次にサーバの負荷に関する情報を取得する時点から、再びステップＳ２０１以降の動作を実行する。 When the load acquisition unit 110 periodically acquires information about the server load, the resource allocation system 200 starts again from step S201 onward when the load acquisition unit 110 next acquires information about the server load. Execute the operation.
　図７は、傾向判定部２３２が負荷傾向を判定する場合における動作を示すフローチャートである。ステップＳ２０８において傾向判定部２３２が負荷傾向を判定する場合には、傾向判定部２３２は、次のように動作する。 FIG. 7 is a flowchart showing an operation when the tendency determination unit 232 determines a load tendency. When the trend determination unit 232 determines the load trend in step S208, the trend determination unit 232 operates as follows.
　最初に、傾向判定部２３２は、負荷取得部１１０によって上述した短い時間間隔で計測された負荷に関する情報を取得する（ステップＳ２５１）。また、傾向判定部２３２は、例えば測定値記憶部１５０が備えるテーブルから測定結果を読み出して、直近の測定から予め定められた回数の測定結果を取得する。 First, the tendency determination unit 232 acquires information on the load measured by the load acquisition unit 110 at the short time interval described above (step S251). Moreover, the tendency determination part 232 reads a measurement result, for example from the table with which the measured value memory | storage part 150 is provided, and acquires the measurement result of the predetermined frequency | count from the latest measurement.
　続いて、傾向判定部２３２は、負荷が、上述した所定の回数に亘って連続して増加しているか否かを判定する（ステップＳ２５２）。傾向判定部２３２は、先に説明したように、上述した所定の回数の各々の時点において計測された負荷の全てが、その一つ前の時点に測定された負荷と比較して大きい場合に、負荷が増加傾向にあると判定する。また、傾向判定部２３２は、測定値記憶部１５０等に記憶された直近の負荷の測定値と、新たに負荷取得部１１０にて取得された負荷との比較を所定の回数について行うことで、負荷の傾向を判定してもよい。 Subsequently, the tendency determining unit 232 determines whether or not the load is continuously increased over the predetermined number of times described above (step S252). As described above, the trend determination unit 232, when all of the loads measured at each time point of the predetermined number of times described above are larger than the load measured at the previous time point, It is determined that the load tends to increase. In addition, the tendency determination unit 232 compares the measurement value of the latest load stored in the measurement value storage unit 150 and the like with the load newly acquired by the load acquisition unit 110 for a predetermined number of times, You may determine the tendency of load.
　傾向判定部２３２が、上述した所定の測定回数に亘って連続して負荷が上昇していると判断した場合（ステップＳ２５２：Ｙｅｓ）には、負荷が増加傾向にあると判定する。この場合には、傾向判定部２３２は、タスクに対するリソースの割当を決定する。決定に応じて、リソース割当部１４０は、タスクに対してプロセッサコアなどのリソースを割当てる（ステップＳ２５３）。そして、リソースの割当てが行われたことから、傾向判定部２３２は、負荷取得部１１０における負荷に関する情報の取得間隔を通常の間隔に変更することを決定する。決定に応じて、間隔調整部２３１は、負荷取得部における負荷に関する情報の取得間隔を通常の間隔に変更する（ステップＳ２５４）。 When the tendency determination unit 232 determines that the load is continuously increased over the predetermined number of measurements described above (step S252: Yes), it is determined that the load is increasing. In this case, the trend determination unit 232 determines resource allocation for the task. In response to the determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S253). Then, since the resource allocation has been performed, the tendency determination unit 232 determines to change the acquisition interval of information regarding the load in the load acquisition unit 110 to a normal interval. In response to the determination, the interval adjustment unit 231 changes the acquisition interval of information regarding the load in the load acquisition unit to a normal interval (step S254).
　また、負荷が上述した所定の測定回数に亘って連続して増加していないと傾向判定部２３２が判断した場合（ステップＳ２５２：Ｎｏ）には、タスクに対するリソースの割当てが必要ではないことが想定される。この場合に、傾向判定部２３２は、負荷が上述した所定の回数に亘って連続して減少しているか否かを判定する（ステップＳ２５５）。傾向判定部２３２は、先に説明したように、上述した所定の回数の各々の時点において計測された負荷の全てが、その一つ前の時点に測定された負荷と比較して小さい場合に、負荷が減少傾向にあると判定する。傾向判定部２３２は、負荷が第２の条件を満たした時点から後の時点に取得された負荷を用いてこの判定を行う。 In addition, when the trend determination unit 232 determines that the load has not continuously increased over the predetermined number of measurements described above (step S252: No), it is assumed that it is not necessary to allocate resources to the task. Is done. In this case, the tendency determination unit 232 determines whether or not the load is continuously decreased over the predetermined number of times described above (step S255). As described above, the trend determination unit 232, when all of the loads measured at each time point of the predetermined number of times described above are smaller than the load measured at the previous time point, It is determined that the load is decreasing. The tendency determination unit 232 performs this determination using the load acquired from the time when the load satisfies the second condition to the later time.
　傾向判定部２３２が、上述した所定の測定回数に亘って連続して負荷が減少していると判断した場合（ステップＳ２５５：Ｙｅｓ）には、負荷が減少しておりタスクに対するリソースの割当ては不要であると想定される。そこで、傾向判定部２３２は、負荷取得部１１０における負荷に関する情報の取得間隔を通常の間隔に変更することを決定する。決定に応じて、間隔調整部２３１は、上述したステップＳ２５４の処理として、負荷取得部における負荷に関する情報の取得間隔を通常の間隔に変更する。 When the trend determination unit 232 determines that the load is continuously reduced over the predetermined number of measurements described above (step S255: Yes), the load is reduced and the resource allocation to the task is unnecessary. It is assumed that Therefore, the trend determination unit 232 determines to change the acquisition interval of information regarding the load in the load acquisition unit 110 to a normal interval. In response to the determination, the interval adjustment unit 231 changes the acquisition interval of information regarding the load in the load acquisition unit to a normal interval as the process of step S254 described above.
　傾向判定部２３２が、上述した所定の測定回数に亘って連続して負荷が減少していないと判断した場合（ステップＳ２５５：Ｎｏ）には、負荷は増加傾向にはないものの、引き続き負荷の傾向を確認する必要がある状態であると想定される。この場合に、傾向判定部２３２は、処理を終了する。すなわち、負荷取得部１１０は、上述した短い時間間隔での負荷に関する情報の取得を継続して行う。また、傾向判定部２３２は、例えば予め定められた時間が経過した後に、再び上述のように負荷傾向を判定してもよい。 When the trend determination unit 232 determines that the load has not decreased continuously over the predetermined number of measurements described above (step S255: No), the load does not tend to increase, but the load trend continues. It is assumed that this is a state that needs to be confirmed. In this case, the trend determination unit 232 ends the process. That is, the load acquisition unit 110 continuously acquires information regarding the load at the short time intervals described above. Moreover, the tendency determination part 232 may determine a load tendency again as mentioned above, for example, after predetermined time passes, for example.
　以上のとおり、本発明の第２の実施形態におけるリソース割当システム２００は、第１の実施形態におけるリソース割当システム１００と同様に、緊急度判定部１２０にて求められた緊急度に応じてリソースの割当てを行う。緊急度が高いと判定される場合には、リソース割当システム２００は、更なる判定を行わずにリソースの割当てを行う。緊急度が、タスクに対してリソースを割当てる程度の高さではないと判定される場合には、リソース割当システム２００は、負荷の測定間隔を短くして、更に負荷の傾向を判定する。そして、負荷が上昇する傾向にあると判定された場合には、リソース割当システム１００は、リソースの割当てを行う。 As described above, the resource allocation system 200 according to the second embodiment of the present invention is configured according to the urgency level obtained by the urgency level determination unit 120 in the same manner as the resource allocation system 100 according to the first embodiment. Make an assignment. If it is determined that the degree of urgency is high, the resource allocation system 200 allocates resources without performing further determination. If it is determined that the degree of urgency is not high enough to allocate a resource to a task, the resource allocation system 200 shortens the load measurement interval and further determines the load trend. When it is determined that the load tends to increase, the resource allocation system 100 allocates resources.
　すなわち、本実施形態におけるリソース割当システム２００は、第１の実施形態におけるリソース割当システム１００と同様の効果を奏する。また、リソース割当システム２００は、緊急度の程度が、リソースの割当ての必要性が高いものの、タスクに対してリソースを割当てる程度ではないである場合に、負荷の測定間隔を短くして負荷の傾向を判定する。このように負荷の傾向が判定されることで、リソース割当システム２００は、負荷が増加傾向にあることを高い精度で判定することができる。 That is, the resource allocation system 200 in this embodiment has the same effects as the resource allocation system 100 in the first embodiment. Further, the resource allocation system 200 shortens the load measurement interval and reduces the load trend when the degree of urgency is high enough to allocate resources but not to allocate resources to tasks. Determine. By determining the load trend in this way, the resource allocation system 200 can determine with high accuracy that the load is increasing.
　したがって、本実施形態におけるリソース割当システム２００は、第１の実施形態におけるリソース割当システム１００と比較して、更に適切に負荷の変動に応じたリソースの割当てを可能にする。 Therefore, the resource allocation system 200 according to the present embodiment enables more appropriate resource allocation according to load fluctuations than the resource allocation system 100 according to the first embodiment.
　（第３の実施形態）
　次に、本発明の第３の実施形態について説明する。図８は、本発明の第３の実施形態におけるリソース割当システムを示す図である。図９は、本発明の第３の実施形態におけるリソース割当システムのパターン照合部に関する具体的な構成例を示す図である。図１０は、本発明の第３の実施形態において負荷上昇パターンの一例を示す図である。図１１は、本発明の第３の実施形態において負荷取得部により取得される負荷と負荷上昇パターンとの関係を説明する図である。図１２は、本発明の第３の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。図１３は、本発明の第３の実施形態におけるリソース割当システムに含まれる傾向判定部の動作の一例を示すフローチャートである。 (Third embodiment)
Next, a third embodiment of the present invention will be described. FIG. 8 is a diagram showing a resource allocation system in the third exemplary embodiment of the present invention. FIG. 9 is a diagram illustrating a specific configuration example regarding the pattern matching unit of the resource allocation system in the third exemplary embodiment of the present invention. FIG. 10 is a diagram illustrating an example of a load increase pattern in the third embodiment of the present invention. FIG. 11 is a diagram illustrating the relationship between the load acquired by the load acquisition unit and the load increase pattern in the third embodiment of the present invention. FIG. 12 is a flowchart showing an example of the operation of the resource allocation system in the third exemplary embodiment of the present invention. FIG. 13 is a flowchart illustrating an example of the operation of the tendency determination unit included in the resource allocation system according to the third embodiment of the present invention.
　図８に示すように、本発明の第３の実施形態におけるリソース割当システム３００は、負荷取得部１１０と、緊急度判定部１２０と、パターン照合部３３３と、リソース割当部１４０と、測定値記憶部１５０とを備える。負荷取得部１１０、緊急度判定部１２０及びリソース割当部１４０は、以下に説明がある箇所を除き、基本的に、本発明の第１の実施形態におけるリソース割当システム１００が備える同じ構成要素と同様の動作を行う。パターン照合部３３３は、緊急度判定部１２０において、サーバの負荷の傾向を更に確認する必要があると判定された場合に、負荷の時間的な変化の様子が所定のパターンと類似するかを判定する。 As shown in FIG. 8, the resource allocation system 300 according to the third exemplary embodiment of the present invention includes a load acquisition unit 110, an urgency determination unit 120, a pattern matching unit 333, a resource allocation unit 140, and a measured value storage. Unit 150. The load acquisition unit 110, the urgency determination unit 120, and the resource allocation unit 140 are basically the same as the same constituent elements included in the resource allocation system 100 according to the first embodiment of the present invention, except for the following description. Perform the operation. The pattern matching unit 333 determines whether the temporal change in load is similar to a predetermined pattern when the urgency level determination unit 120 determines that it is necessary to further check the load trend of the server. To do.
　すなわち、本実施形態におけるリソース割当システム３００は、負荷判定部１３０としてパターン照合部３３３を備える点が、第１の実施形態におけるリソース割当システム１００と異なる。それ以外の点については、本実施形態におけるリソース割当システム３００は、第１の実施形態におけるリソース割当システム１００と基本的に同様の構成を備える。 That is, the resource allocation system 300 in the present embodiment is different from the resource allocation system 100 in the first embodiment in that the load determination unit 130 includes a pattern matching unit 333. In other respects, the resource allocation system 300 in the present embodiment has basically the same configuration as the resource allocation system 100 in the first embodiment.
　続いて、本発明の第３の実施形態におけるリソース割当システム３００の各構成要素について説明する。 Subsequently, each component of the resource allocation system 300 according to the third embodiment of the present invention will be described.
　本実施形態においては、負荷取得部１１０は、予め定められた間隔でタスクを実行するサーバの負荷に関する情報を取得する。また、負荷取得部１１０にて取得された負荷に関する情報は、測定値記憶部１５０に記憶される。測定値記憶部１５０は、負荷に関する情報を、例えば第２の実施形態における図５に示すテーブルの形式にて記憶する。 In the present embodiment, the load acquisition unit 110 acquires information related to the load of the server that executes a task at a predetermined interval. In addition, information regarding the load acquired by the load acquisition unit 110 is stored in the measurement value storage unit 150. The measured value storage unit 150 stores information about the load, for example, in the form of a table shown in FIG. 5 in the second embodiment.
　パターン照合部３３３は、緊急度判定部１２０において緊急度が第２の条件を満たし、サーバの負荷の傾向を更に確認する必要があると判定された場合に、負荷の時間的な変化の様子が所定のパターンと類似するかを照合する。この所定のパターンは、負荷の上昇を表すパターンである。 When the urgency level determination unit 120 determines that the urgency level satisfies the second condition and the server load trend needs to be further confirmed, the pattern matching unit 333 indicates how the load changes with time. It is verified whether it is similar to a predetermined pattern. This predetermined pattern is a pattern representing an increase in load.
　例えば、パターン照合部３３３は、緊急度が第２の条件を満たすと判定された時点より前の予め設定された期間における負荷を取得する。パターン照合部３３３は、当該期間における負荷の時間的な変化の様子と、予め用意された、負荷が時間的に上昇することを表す所定のパターンとを比較する。そして、これらの相違が予め定められた条件を満たす場合に、パターン照合部３３３は、負荷の時間的な変化の様子が負荷の上昇を表す所定のパターンと類似すると判定する。この場合には、パターン照合部３３３は、負荷が増加の傾向にあるとして、タスクへのリソースの割当てを決定する。リソース割当部１４０は、当該決定に応じて、タスクに対してリソースを割当てる。 For example, the pattern matching unit 333 acquires a load in a preset period before the time when it is determined that the degree of urgency satisfies the second condition. The pattern matching unit 333 compares the state of the temporal change of the load during the period with a predetermined pattern that represents a load that increases in time. When these differences satisfy a predetermined condition, the pattern matching unit 333 determines that the temporal change in the load is similar to a predetermined pattern representing an increase in the load. In this case, the pattern matching unit 333 determines allocation of resources to tasks on the assumption that the load tends to increase. The resource allocation unit 140 allocates resources to tasks according to the determination.
　パターン照合部３３３は、具体的な一例として、図９に示す構成を有する。図９に示す例では、パターン照合部３３３は、類似度判定部３３４と、パターン記憶部３３５とを有する。 The pattern matching unit 333 has a configuration shown in FIG. 9 as a specific example. In the example illustrated in FIG. 9, the pattern matching unit 333 includes a similarity determination unit 334 and a pattern storage unit 335.
　パターン記憶部３３５は、負荷が時間的に上昇することを表す任意の数のパターンを記憶する。パターン記憶部３３５において記憶されるパターンは、例えば、過去に測定された負荷が上昇する様子である。このパターンは、例えば、シミュレーションによって求められた負荷の上昇の様子であってもよい。また、パターン記憶部３３５において記憶されるパターンの数は特に制限されない。 The pattern storage unit 335 stores an arbitrary number of patterns indicating that the load increases with time. The pattern stored in the pattern storage unit 335 is, for example, a state in which the load measured in the past increases. This pattern may be, for example, an increase in load obtained by simulation. Further, the number of patterns stored in the pattern storage unit 335 is not particularly limited.
　図１０は、パターン記憶部３３５において記憶されるパターンの例を示す図である。図１０に示すパターンは、図１１における時刻ＴｓからＴ０までの間に取得された負荷の計測値の系列を表す。Ｔ０は、その時点に取得された負荷に基づいて求められた緊急度が初めて第２の条件を満たした時刻である。Ｔｓは、負荷取得部１１０による負荷の取得が行われた過去の時点である。Ｔｓは、一例として、ＴｓからＴ０までの時間が負荷取得部１１０による負荷の取得間隔の３０倍となるように定められる。なお、Ｔ１は、その時点に取得された負荷に基づいて求められた緊急度が初めて第１の条件を満たす時刻である。 FIG. 10 is a diagram illustrating an example of patterns stored in the pattern storage unit 335. The pattern shown in FIG. 10 represents a series of load measurement values acquired from time Ts to T0 in FIG. T0 is the time when the urgency obtained based on the load acquired at that time satisfies the second condition for the first time. Ts is a past time point when the load acquisition unit 110 has acquired the load. For example, Ts is determined such that the time from Ts to T0 is 30 times the load acquisition interval by the load acquisition unit 110. Note that T1 is the time when the urgency level obtained based on the load acquired at that time satisfies the first condition for the first time.
　類似度判定部３３４は、負荷取得部１１０にて取得された負荷の時間的な変化の様子と、パターン記憶部３３５において記憶されたパターンの各々との類似度を求める。 The similarity determination unit 334 obtains the similarity between the state of the temporal change in the load acquired by the load acquisition unit 110 and each of the patterns stored in the pattern storage unit 335.
　類似度判定部３３４は、例えば以下の計算式を用いて類似度を求める。類似度判定部３３４は、パターン記憶部３３５に複数のパターンが記憶されている場合には、予め一つを選択する。負荷取得部１１０にて取得された負荷の測定値はＷｉ、パターン記憶部３３５において記憶されたパターンに含まれる負荷の値はＬｉとして表されるとする。また、負荷取得部１１０にて取得された負荷の測定値及び負荷の上昇を表すパターンのそれぞれにｎ個の値が含まれているとする。この場合に、類似度判定部３３４は、類似度Ｒを例えば以下の（２）式を用いて求める。 The similarity determination unit 334 obtains the similarity using, for example, the following calculation formula. The similarity determination unit 334 selects one in advance when a plurality of patterns are stored in the pattern storage unit 335. It is assumed that the load measurement value acquired by the load acquisition unit 110 is expressed as Wi, and the load value included in the pattern stored in the pattern storage unit 335 is expressed as Li. Further, it is assumed that n values are included in each of the load measurement value acquired by the load acquisition unit 110 and the pattern representing the load increase. In this case, the similarity determination unit 334 calculates the similarity R using, for example, the following equation (2).
Figure JPOXMLDOC01-appb-M000002
 　類似度Ｒが求められると、類似度判定部３３４は、求められた類似度Ｒが類似の程度に関する条件を満たすかを判定する。一例として、類似度判定部３３４は、求められた類似度Ｒと類似度に関する閾値Ｒ０との比較を行う。Ｒ＜Ｒ０との条件を満たす（すなわち、類似度Ｒの大きさが閾値Ｒ０よりも小さい）場合には、類似度判定部３３４は、負荷取得部１１０にて取得された負荷の時間的な変化の様子が類似すると判定する。すなわち、類似度判定部３３４は、負荷が上昇傾向にあると判定する。そして、類似度判定部３３４は、タスクに対するリソースの割当てを決定する。リソース割当部１４０は、当該決定に応じて、タスクに対してリソースを割当てる。
Figure JPOXMLDOC01-appb-M000002
 When the similarity R is obtained, the similarity determination unit 334 determines whether the obtained similarity R satisfies a condition regarding the degree of similarity. As an example, the similarity determination unit 334 compares the obtained similarity R with a threshold value R0 regarding the similarity. When the condition R <R0 is satisfied (that is, the degree of similarity R is smaller than the threshold value R0), the similarity determination unit 334 changes the temporal change in the load acquired by the load acquisition unit 110. Are determined to be similar. That is, the similarity determination unit 334 determines that the load tends to increase. Then, the similarity determination unit 334 determines resource allocation for the task. The resource allocation unit 140 allocates resources to tasks according to the determination.
　また、Ｒ＜Ｒ０との条件を満たさない（すなわち、類似度Ｒの大きさが閾値Ｒ０以上である）場合には、類似度判定部３３４は、負荷取得部１１０にて取得された負荷の時間的な変化の様子が比較対象とされたパターンと異なると判定する。この場合には、類似度判定部３３４は、パターン記憶部３３５に記憶された他のパターンを選択して、再び類似度Ｒを求める。パターン記憶部３３５に記憶された全てのパターンについてＲ＜Ｒ０との条件を満たさない場合には、その時点での負荷の傾向がいずれのパターンにも類似しないことが想定される。この場合には、類似度判定部３３４は、負荷が上昇傾向にないと判定する。 When the condition R <R0 is not satisfied (that is, when the similarity R is greater than or equal to the threshold value R0), the similarity determination unit 334 determines the load time acquired by the load acquisition unit 110. It is determined that the state of the general change is different from the pattern to be compared. In this case, the similarity determination unit 334 selects another pattern stored in the pattern storage unit 335 and obtains the similarity R again. If all the patterns stored in the pattern storage unit 335 do not satisfy the condition of R <R0, it is assumed that the load tendency at that time is not similar to any pattern. In this case, the similarity determination unit 334 determines that the load does not tend to increase.
　続いて、図１２に示すフローチャートを参照して、本発明の第３の実施形態におけるリソース割当システム３００の動作の一例について説明する。 Subsequently, an example of the operation of the resource allocation system 300 according to the third embodiment of the present invention will be described with reference to the flowchart shown in FIG.
　最初に、負荷取得部１１０は、タスクを実行するサーバの負荷に関する情報を取得する（ステップＳ３０１）。次に、緊急度判定部１２０は、緊急度を求める（ステップＳ３０２）。続いて、緊急度判定部１２０は、緊急度が第１の条件を満たすかを判定する（ステップＳ３０３）。 First, the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S301). Next, the urgency level determination unit 120 calculates the urgency level (step S302). Subsequently, the urgency level determination unit 120 determines whether the urgency level satisfies the first condition (step S303).
　緊急度が第１の条件を満たす場合（ステップＳ３０３：Ｙｅｓ）には、緊急度判定部１２０は、タスクに対するリソースの割当てを決定する。この決定に応じて、リソース割当部１４０は、プロセッサコア等のリソースをタスクに対して割当てる（ステップＳ３０４）。 If the urgency level satisfies the first condition (step S303: Yes), the urgency level determination unit 120 determines resource allocation to the task. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S304).
　緊急度判定部１２０は、緊急度が先に説明した第１の条件を満たさない場合（ステップＳ３０３：Ｎｏ）に、緊急度が先に説明した第２の条件を満たすかを判定する（ステップＳ３０５）。緊急度が第２の条件を満たす（ステップＳ３０５：Ｙｅｓ）場合には、パターン照合部３３３は、負荷の時間的な変化の様子とパターンとの照合を行う（ステップＳ３０６）。 When the urgency level does not satisfy the first condition described earlier (step S303: No), the urgency level determination unit 120 determines whether the urgency level satisfies the second condition described above (step S305). ). When the degree of urgency satisfies the second condition (step S305: Yes), the pattern matching unit 333 performs matching between the state of load change with time and the pattern (step S306).
　なお、負荷取得部１１０が定期的にサーバの負荷に関する情報を取得する場合には、リソース割当システム３００は、図１２に示すフローチャートの動作を繰り返して行ってもよい。 In addition, when the load acquisition unit 110 periodically acquires information related to the server load, the resource allocation system 300 may repeatedly perform the operation of the flowchart illustrated in FIG.
　図１３は、パターン照合部３３３が負荷傾向を判定する場合における動作を示すフローチャートである。上述のステップＳ３０６においてパターン照合部３３３がパターンの照合を行う場合には、パターン照合部３３３は、次のように動作する。なお、この動作例において、パターン照合部３３３は、図９に示す構成を有することを想定する。 FIG. 13 is a flowchart showing an operation when the pattern matching unit 333 determines a load tendency. When the pattern matching unit 333 performs pattern matching in step S306 described above, the pattern matching unit 333 operates as follows. In this operation example, it is assumed that the pattern matching unit 333 has the configuration shown in FIG.
　最初に、類似度判定部３３４は、パターン記憶部３３５に記憶されたパターンの一つを選択して読み出す（ステップＳ３５１）。続いて、類似度判定部３３４は、上述した（２）式などを用いて、計測された負荷のステップＳ３５１で読み出されたパターンとの類似度を求める（ステップＳ３５２）。 First, the similarity determination unit 334 selects and reads one of the patterns stored in the pattern storage unit 335 (step S351). Subsequently, the similarity determination unit 334 obtains the similarity with the pattern read in step S351 of the measured load using the above-described equation (2) (step S352).
　類似度が求められると、類似度判定部３３４は、求められた類似度Ｒが類似の程度に関する条件を満たすかを判定する（ステップＳ３５３）。すなわち、類似度判定部３３４は、類似度Ｒと閾値Ｒ０との比較を行う。類似度Ｒ及び閾値Ｒ０がＲ＜Ｒ０との条件を満たす場合（ステップＳ３５３：Ｙｅｓ）には、類似度判定部３３４は、タスクに対するリソースの割当てを決定する。この決定に応じて、リソース割当部１４０は、プロセッサコア等のリソースをタスクに対して割当てる（ステップＳ３５４）。 When the degree of similarity is obtained, the degree-of-similarity determination unit 334 determines whether the obtained degree of similarity R satisfies a condition regarding the degree of similarity (step S353). That is, the similarity determination unit 334 compares the similarity R with the threshold value R0. When the similarity R and the threshold value R0 satisfy the condition that R <R0 (step S353: Yes), the similarity determination unit 334 determines resource allocation to the task. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S354).
　類似度Ｒ及び閾値Ｒ０がＲ＜Ｒ０との条件を満たさない場合（ステップＳ３５３：Ｎｏ）には、類似度判定部３３４は、ステップＳ３５５の処理を行う。ステップＳ３５５においては、類似度判定部３３４は、パターン記憶部３３５に記憶された全てのパターンに対してステップＳ３５２の処理において類似度を求めて閾値との比較を行ったかを判定する。 When the similarity R and the threshold value R0 do not satisfy the condition of R <R0 (step S353: No), the similarity determination unit 334 performs the process of step S355. In step S355, the similarity determination unit 334 determines whether similarity is obtained for all patterns stored in the pattern storage unit 335 and compared with a threshold value in step S352.
　パターン記憶部３３５に記憶された全てのパターンに対して類似度Ｒが類似の程度に関する条件を満たさないと判定された場合（ステップＳ３５５：Ｙｅｓ）には、類似度判定部３３４は、負荷が上昇の傾向にないと判定する。この場合には、類似度判定部３３４は、処理を終了する。 When it is determined that the degree of similarity R does not satisfy the condition regarding the degree of similarity for all patterns stored in the pattern storage unit 335 (step S355: Yes), the similarity degree determination unit 334 increases the load. It is determined that there is no tendency. In this case, the similarity determination unit 334 ends the process.
　パターン記憶部３３５に記憶された全てのパターンに対して既に類似度の導出等を行っていない場合（ステップＳ３５５：Ｎｏ）には、類似度判定部３３４は、ステップＳ３５１に戻って処理を継続して行う。すなわち、類似度判定部３３４は、ステップＳ３５１においてパターン記憶部３３５に記憶された他のパターンを選択して、処理を継続する。 If the derivation of similarity is not already performed for all the patterns stored in the pattern storage unit 335 (step S355: No), the similarity determination unit 334 returns to step S351 and continues the processing. Do it. That is, the similarity determination unit 334 selects another pattern stored in the pattern storage unit 335 in step S351 and continues the process.
　以上のとおり、本発明の第３の実施形態におけるリソース割当システム３００は、第１の実施形態におけるリソース割当システム１００と同様に、緊急度判定部１２０にて求められた緊急度に応じてリソースの割当てを行う。緊急度が高いと判定される場合には、リソース割当システム２００は、更なる判定を行わずにリソースの割当てを行う。緊急度が、タスクに対してリソースを割当てる程度の高さではないと判定される場合には、リソース割当システム２００は、負荷の時間的な変化の様子が負荷の上昇を表すパターンと類似するかを判定して、更に負荷の傾向を判定する。そして、負荷が上昇する傾向にあると判定された場合には、リソース割当システム１００は、リソースの割当てを行う。 As described above, the resource allocation system 300 according to the third exemplary embodiment of the present invention is configured according to the urgency level obtained by the urgency level determination unit 120 in the same manner as the resource allocation system 100 according to the first exemplary embodiment. Make an assignment. If it is determined that the degree of urgency is high, the resource allocation system 200 allocates resources without performing further determination. If it is determined that the degree of urgency is not high enough to allocate a resource to a task, the resource allocation system 200 is similar to a pattern in which the temporal change in the load represents a load increase And the load tendency is further determined. When it is determined that the load tends to increase, the resource allocation system 100 allocates resources.
　すなわち、本実施形態におけるリソース割当システム３００は、第１の実施形態におけるリソース割当システム１００と同様の効果を奏する。また、リソース割当システム３００は、緊急度の高さが中程度である場合に、負荷の時間的な変化の様子が負荷の上昇を表すパターンと類似するかを判定する。そのため、リソース割当システム３００は、種々の負荷上昇のパターンに基づいて、負荷が増加傾向にあることを判定することができる。 That is, the resource allocation system 300 in the present embodiment has the same effects as the resource allocation system 100 in the first embodiment. Further, when the urgency level is medium, the resource allocation system 300 determines whether the temporal change in the load is similar to the pattern representing the load increase. Therefore, the resource allocation system 300 can determine that the load tends to increase based on various load increase patterns.
　したがって、本実施形態におけるリソース割当システム３００は、第１の実施形態におけるリソース割当システム１００と比較して、更に適切に負荷の変動に応じたリソースの割当てを可能にする。 Therefore, the resource allocation system 300 according to the present embodiment enables the resource allocation according to the load variation more appropriately than the resource allocation system 100 according to the first embodiment.
　（第４の実施形態）
　次に、本発明の第４の実施形態について説明する。図１４は、本発明の第４の実施形態におけるリソース割当システムを示す図である。図１５は、本発明の第４の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。 (Fourth embodiment)
Next, a fourth embodiment of the present invention will be described. FIG. 14 is a diagram showing a resource allocation system in the fourth exemplary embodiment of the present invention. FIG. 15 is a flowchart showing an example of the operation of the resource allocation system in the fourth exemplary embodiment of the present invention.
　図１４に示すとおり、本発明の第４の実施形態におけるリソース割当システム４００は、負荷取得部１１０と、緊急度判定部４２０と、間隔調整部２３１と、傾向判定部２３２と、リソース割当部１４０とを備える。緊急度判定部４２０は、以下に説明がある箇所を除き、本発明の第２の実施形態におけるリソース割当システム２００が備える緊急度判定部１２０と基本的に同様の動作を行う。 As illustrated in FIG. 14, the resource allocation system 400 according to the fourth exemplary embodiment of the present invention includes a load acquisition unit 110, an urgency determination unit 420, an interval adjustment unit 231, a tendency determination unit 232, and a resource allocation unit 140. With. The urgency level determination unit 420 performs basically the same operation as the urgency level determination unit 120 included in the resource allocation system 200 according to the second exemplary embodiment of the present invention, except for portions described below.
　本実施形態においては、緊急度判定部４２０は、サーバの負荷の傾向を確認する必要があると判定する場合に、更に、その時点における緊急度の程度を判定する。具体的には、緊急度判定部４２０は、先に説明した第２の条件に相当し、緊急度の程度が異なる複数の条件のいずれを満たすかを判定する。また、間隔調整部２３１は、緊急度の程度が異なる複数の条件のそれぞれに基づいて、負荷取得部１１０による負荷に関する情報の取得間隔を変更する。それ以外の点については、本実施形態におけるリソース割当システム２００は、第１の実施形態におけるリソース割当システム１００と基本的に同様の構成を備える。 In the present embodiment, when determining that it is necessary to check the tendency of the load on the server, the urgency determination unit 420 further determines the degree of urgency at that time. Specifically, the urgency determination unit 420 corresponds to the second condition described above, and determines which of a plurality of conditions with different degrees of urgency is satisfied. In addition, the interval adjustment unit 231 changes the acquisition interval of information related to the load by the load acquisition unit 110 based on each of a plurality of conditions having different degrees of urgency. In other respects, the resource allocation system 200 in this embodiment has basically the same configuration as the resource allocation system 100 in the first embodiment.
　続いて、本発明の第４の実施形態におけるリソース割当システム４００の各構成要素について説明する。 Subsequently, each component of the resource allocation system 400 according to the fourth embodiment of the present invention will be described.
　本実施形態においては、負荷取得部１１０は、予め定められた間隔でタスクを実行するサーバの負荷に関する情報を取得する。また、この間隔は、間隔調整部２３１の指示に基づいて、予め定められた間隔よりも短い異なる間隔のいずれかに変更される場合がある。 In the present embodiment, the load acquisition unit 110 acquires information related to the load of the server that executes a task at a predetermined interval. In addition, this interval may be changed to any one of different intervals shorter than a predetermined interval based on an instruction from the interval adjusting unit 231.
　上述のように、緊急度判定部４２０は、サーバの負荷の傾向を更に確認する必要があると判定する場合に、更に、その時点における緊急度の程度を判定する。本実施形態では、一例として、緊急度判定部４２０は、サーバの負荷の傾向を更に確認する必要があると判定する場合に、緊急度の程度が異なる複数の条件を用いる。 As described above, when it is determined that the server load trend needs to be further confirmed, the urgency determination unit 420 further determines the degree of urgency at that time. In this embodiment, as an example, the urgency level determination unit 420 uses a plurality of conditions with different degrees of urgency level when determining that it is necessary to further confirm the load trend of the server.
　緊急度判定部４２０は、具体的には、緊急度が第１から第３の条件のいずれかを満たす場合に、サーバの負荷の傾向を更に確認する必要があると判定する。第２及び第３の条件は、タスクに対してリソースを割当てるべきか否かを更に判断する必要がある状態であることを示す条件である。また、第２及び第３の条件は、緊急度の程度が異なる条件である。本実施形態においては、第３の条件は、第２の条件と比較して緊急度の程度が低く、タスクに対するリソースの割当ての必要性が低いことを示す。なお、本実施形態においては、第１の条件は、上述した各実施形態における第１の条件と同様の条件となる。 Specifically, the urgency level determination unit 420 determines that it is necessary to further check the load tendency of the server when the urgency level satisfies any of the first to third conditions. The second and third conditions are conditions indicating that it is necessary to further determine whether or not resources should be allocated to the task. The second and third conditions are conditions with different degrees of urgency. In the present embodiment, the third condition indicates that the degree of urgency is lower than that of the second condition, and the necessity for resource allocation to tasks is low. In the present embodiment, the first condition is the same as the first condition in each of the embodiments described above.
　第２の実施形態における緊急度判定部１２０にて用いられる条件と同様に、第１から第３の条件のそれぞれは、緊急度として求められた値に対する閾値の形式で表される。例えば、緊急度Ｅが上述した（１）式にて求められる場合には、第１の条件に関する閾値としてＥ１が、第２の条件に関する閾値としてＥ２が、第３の条件に関する閾値としてＥ３が定められる。そして、緊急度判定部４２０は、緊急度がＥ１≦Ｅを満たす（すなわち、緊急度Ｅの大きさがＥ１以上である）場合に、緊急度が第１の条件を満たすと判定する。緊急度判定部４２０は、緊急度ＥがＥ２≦Ｅ＜Ｅ１を満たす（すなわち、緊急度Ｅの大きさがＥ２以上であり、かつ、Ｅ１よりも小さい）場合に、緊急度が第２の条件を満たすと判定する。また、緊急度判定部４２０は、緊急度がＥ３≦Ｅ＜Ｅ２を満たす（すなわち、緊急度Ｅの大きさがＥ３以上であり、かつ、Ｅ２よりも小さい）場合に、緊急度が第３の条件を満たすと判定する。緊急度がＥ＜Ｅ３を満たす（すなわち、緊急度の大きさがＥ３よりも小さい）場合には、緊急度判定部４２０は、緊急度がいずれの条件も満たさないと判定する。 Similarly to the conditions used in the urgency level determination unit 120 in the second embodiment, each of the first to third conditions is expressed in the form of a threshold for the value obtained as the urgency level. For example, when the degree of urgency E is obtained by the above-described equation (1), E1 is determined as the threshold for the first condition, E2 is set as the threshold for the second condition, and E3 is set as the threshold for the third condition. It is done. The urgency level determination unit 420 determines that the urgency level satisfies the first condition when the urgency level satisfies E1 ≦ E (that is, the level of the urgency level E is equal to or greater than E1). The urgency level determination unit 420 sets the urgency level to the second condition when the urgency level E satisfies E2 ≦ E <E1 (that is, the level of the urgency level E is equal to or larger than E2 and smaller than E1). It is determined that The urgency level determination unit 420 also sets the urgency level to the third level when the urgency level satisfies E3 ≦ E <E2 (that is, when the urgency level E is greater than or equal to E3 and smaller than E2). It is determined that the condition is satisfied. When the urgency level satisfies E <E3 (that is, the urgency level is smaller than E3), the urgency level determination unit 420 determines that the urgency level does not satisfy any of the conditions.
　間隔調整部２３１は、緊急度判定部４２０において緊急度が第２又は第３の条件のいずれかを満たすと判定された場合に、負荷取得部１１０による負荷に関する情報の取得間隔を変更する。本実施形態においては、間隔調整部２３１は、緊急度が第２又は第３の条件のいずれかを満たすと判定された場合に、負荷取得部１１０による負荷に関する情報の取得間隔を通常の時間間隔と比較して短い時間間隔に設定する。 The interval adjustment unit 231 changes the acquisition interval of the information regarding the load by the load acquisition unit 110 when the urgency determination unit 420 determines that the urgency satisfies either the second or the third condition. In the present embodiment, the interval adjusting unit 231 sets the acquisition interval of the information related to the load by the load acquiring unit 110 as a normal time interval when it is determined that the urgency satisfies either the second or the third condition. Set to a shorter time interval.
　例えば、間隔調整部２３１は、緊急度が第３の条件を満たす場合には、負荷取得部１１０による取得間隔を通常の時間間隔と比較して短い時間間隔に設定する。そして、間隔調整部２３１は、緊急度が第２の条件を満たす場合には、負荷取得部１１０による取得間隔を、緊急度が第３の条件を満たす場合における時間間隔と比較して更に短い時間間隔に設定する。 For example, when the degree of urgency satisfies the third condition, the interval adjustment unit 231 sets the acquisition interval by the load acquisition unit 110 to a shorter time interval than the normal time interval. Then, when the urgency level satisfies the second condition, the interval adjusting unit 231 sets the acquisition interval by the load acquisition unit 110 to be shorter than the time interval when the urgency level satisfies the third condition. Set to interval.
　すなわち、間隔調整部２３１は、タスクに対してリソースを割当てるべきか否かを更に判断する必要がある場合に、緊急度の高さに応じて時間間隔が短くなるように、負荷取得部１１０による負荷に関する情報の取得間隔を設定する。 That is, when it is necessary to further determine whether or not resources should be allocated to the task, the interval adjustment unit 231 uses the load acquisition unit 110 to shorten the time interval according to the level of urgency. Set the acquisition interval of information about the load.
　続いて、図１５に示すフローチャート参照して、本発明の第４の実施形態におけるリソース割当システム４００の動作の一例について説明する。 Subsequently, an example of the operation of the resource allocation system 400 according to the fourth embodiment of the present invention will be described with reference to the flowchart shown in FIG.
　最初に、負荷取得部１１０は、タスクを実行するサーバの負荷に関する情報を取得する（ステップＳ４０１）。次に、緊急度判定部４２０は、負荷取得部１１０における負荷に関する情報の取得間隔として、通常の時間間隔が設定されているかを判定する（ステップＳ４０２）。 First, the load acquisition unit 110 acquires information related to the load of the server that executes the task (step S401). Next, the urgency determining unit 420 determines whether or not a normal time interval is set as an acquisition interval of information regarding the load in the load acquiring unit 110 (step S402).
　負荷取得部１１０に対して、負荷に関する情報の取得間隔として通常の時間間隔が設定されている場合（ステップＳ４０２：Ｙｅｓ）には、緊急度判定部４２０は、緊急度を求める（ステップＳ４０３）。そして、緊急度判定部４２０は、緊急度が第１の条件を満たすかを判定する（ステップＳ４０４）。 When a normal time interval is set as an acquisition interval of information related to the load with respect to the load acquisition unit 110 (step S402: Yes), the urgency level determination unit 420 obtains the urgency level (step S403). Then, the urgency determination unit 420 determines whether the urgency satisfies the first condition (step S404).
　ステップＳ４０３にて求めた緊急度が第１の条件を満たす場合（ステップＳ４０４：Ｙｅｓ）には、緊急度判定部４２０は、タスクに対するリソースの割当てを決定する。この決定に応じて、リソース割当部１４０は、プロセッサコア等のリソースをタスクに対して割当てる（ステップＳ４０５）。 When the urgency level obtained in step S403 satisfies the first condition (step S404: Yes), the urgency level determination unit 420 determines resource allocation to the task. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S405).
　緊急度判定部４２０は、緊急度が先に説明した第１の条件を満たさない場合（ステップＳ４０４：Ｎｏ）に、緊急度が先に説明した第２の条件を満たすかを判定する（ステップＳ４０６）。緊急度が第２の条件を満たす（ステップＳ４０６：Ｙｅｓ）場合には、間隔調整部２３１は、負荷取得部１１０における負荷に関する情報の取得間隔を通常の時間間隔と比較して短い時間間隔とするように設定する（ステップＳ４０７）。 When the urgency level does not satisfy the first condition described above (step S404: No), the urgency level determination unit 420 determines whether the urgency level satisfies the second condition described above (step S406). ). When the degree of urgency satisfies the second condition (step S406: Yes), the interval adjustment unit 231 sets the acquisition interval of the information regarding the load in the load acquisition unit 110 to a shorter time interval than the normal time interval. (Step S407).
　また、負荷取得部１１０に対して、負荷に関する情報の取得間隔として通常の時間間隔が設定されていない場合（ステップＳ４０２：Ｎｏ）には、負荷に関する情報の取得間隔として、短い時間間隔が設定されていると想定される。この場合には、傾向判定部２３２は、負荷傾向を判定する（ステップＳ４０８）。 In addition, when a normal time interval is not set for the load acquisition unit 110 as a load information acquisition interval (step S402: No), a short time interval is set as the load information acquisition interval. It is assumed that In this case, the tendency determination unit 232 determines a load tendency (step S408).
　なお、ステップＳ４０１からＳ４０８までの動作は、本発明の第２の実施形態におけるステップＳ２０１からＳ２０８までの各ステップとそれぞれ同様の動作を行う。また、ステップＳ４０８において、傾向判定部２３２は、図７に示すフローチャートに従って負荷傾向の判定を行う。 The operations from step S401 to S408 are the same as the respective steps from step S201 to S208 in the second embodiment of the present invention. In step S408, the trend determination unit 232 determines the load trend according to the flowchart shown in FIG.
　更に、緊急度判定部４２０は、ステップＳ４０５において緊急度が第２の条件を満たさない（ステップＳ４０６：Ｎｏ）場合には、緊急度が先に説明した第３の条件を満たすかを判定する（ステップＳ４０９）。緊急度が第３の条件を満たす（ステップＳ４０９：Ｙｅｓ）場合には、間隔調整部２３１は、負荷取得部１１０における負荷に関する情報の取得間隔を通常の時間間隔と比較して短い時間間隔とするように設定する（ステップＳ４１０）。ステップＳ４１０において設定される時間間隔は、通常の時間間隔と比較して短く、かつ、ステップＳ４０７において設定される時間間隔と比較して長い間隔である。 Furthermore, when the urgency level does not satisfy the second condition in step S405 (step S406: No), the urgency level determination unit 420 determines whether the urgency level satisfies the third condition described earlier ( Step S409). When the degree of urgency satisfies the third condition (step S409: Yes), the interval adjustment unit 231 sets the acquisition interval of information regarding the load in the load acquisition unit 110 to a shorter time interval than the normal time interval. (Step S410). The time interval set in step S410 is shorter than the normal time interval and longer than the time interval set in step S407.
　なお、負荷取得部１１０が定期的にサーバの負荷に関する情報を取得する場合には、リソース割当システム４００は、図１５に示すフローチャートの動作を繰り返し行ってもよい。 Note that when the load acquisition unit 110 periodically acquires information related to the server load, the resource allocation system 400 may repeatedly perform the operation of the flowchart illustrated in FIG. 15.
　以上のとおり、本発明の第４の実施形態におけるリソース割当システム４００は、第２の実施形態におけるリソース割当システム２００と同様に、負荷の測定間隔を短くして負荷の傾向を判定する。したがって、本実施形態におけるリソース割当システム４００は、第２の実施形態におけるリソース割当システム２００と同様の効果を奏する。 As described above, the resource allocation system 400 according to the fourth embodiment of the present invention determines the load tendency by shortening the load measurement interval, similarly to the resource allocation system 200 according to the second embodiment. Therefore, the resource allocation system 400 in this embodiment has the same effect as the resource allocation system 200 in the second embodiment.
　更に、本実施形態におけるリソース割当システム４００では、緊急度判定部４２０が、サーバの負荷の傾向を更に確認する必要があると判断された場合に、緊急度の高さが異なる複数の条件を用いる。そして、負荷取得部１１０による負荷に関する情報の取得間隔を通常と比較して短い間隔に設定する場合に、緊急度判定部４２０にて求められた緊急度の高さに応じて時間間隔が短くなるように時間間隔を設定する。このようにすることで、傾向判定部２３２は、負荷が増加傾向にあることを早い段階で判定することが可能となる。 Furthermore, in the resource allocation system 400 according to the present embodiment, when the urgency level determination unit 420 determines that it is necessary to further check the load tendency of the server, a plurality of conditions having different levels of urgency level are used. . And when setting the acquisition interval of the information regarding the load by the load acquisition unit 110 to be a short interval compared to the normal time interval, the time interval is shortened according to the high degree of urgency determined by the urgency determination unit 420. Set the time interval to By doing in this way, it becomes possible for the tendency determination part 232 to determine at an early stage that the load is increasing.
　したがって、本実施形態におけるリソース割当システム４００は、第２の実施形態におけるリソース割当システム２００等と比較して、更に適切に負荷の変動に応じたリソースの割当てを可能にする。 Therefore, the resource allocation system 400 according to the present embodiment enables the resource allocation according to the load variation more appropriately than the resource allocation system 200 according to the second embodiment.
　なお、本実施形態においては、緊急度判定部４２０は、一例として緊急度が第１から第３の条件のいずれかを満たすか否かを判定した。そして、間隔調整部２３１は、第２又は第３の条件のうち、求められた緊急度が満たす条件に応じて、負荷取得部１１０による負荷に関する情報の取得間隔を変更した。しかしながら、緊急度判定部４２０は、緊急度の程度に関する更に多くの条件を用いてもよい。この場合には、間隔調整部２３１は、例えば、求められた緊急度が満たす条件に応じて、負荷取得部１１０による負荷に関する情報の取得間隔を変更する。 In the present embodiment, the urgency level determination unit 420 determines whether the urgency level satisfies any of the first to third conditions as an example. And the space | interval adjustment part 231 changed the acquisition interval of the information regarding the load by the load acquisition part 110 according to the conditions which the calculated | required urgency degree satisfy | fills among 2nd or 3rd conditions. However, the urgency determination unit 420 may use more conditions regarding the degree of urgency. In this case, for example, the interval adjusting unit 231 changes the acquisition interval of the information regarding the load by the load acquiring unit 110 according to the condition that the obtained urgency level satisfies.
　（第５の実施形態）
　次に、本発明の第５の実施形態について説明する。図１６は、本発明の第５の実施形態におけるリソース割当システムを示す図である。図１７は、本発明の第５の実施形態におけるリソース割当システムの動作の一例を示すフローチャートである。 (Fifth embodiment)
Next, a fifth embodiment of the present invention will be described. FIG. 16 is a diagram showing a resource allocation system in the fifth exemplary embodiment of the present invention. FIG. 17 is a flowchart showing an example of the operation of the resource allocation system in the fifth exemplary embodiment of the present invention.
　図１６に示すように、本発明の第５の実施形態におけるリソース割当システム５００は、負荷取得部１１０と、緊急度判定部５２０と、傾向判定部５３２と、間隔調整部５３１と、パターン照合部５３３と、リソース割当部１４０と、測定値記憶部１５０とを備える。 As illustrated in FIG. 16, the resource allocation system 500 according to the fifth exemplary embodiment of the present invention includes a load acquisition unit 110, an urgency determination unit 520, a tendency determination unit 532, an interval adjustment unit 531, and a pattern matching unit. 533, a resource allocation unit 140, and a measurement value storage unit 150.
　本実施形態において、負荷取得部１１０及びリソース割当部１４０は、以下に説明がある箇所を除き、本発明の第１の実施形態におけるリソース割当システム１００が備える同じ構成要素と基本的に同様の動作を行う。また、緊急度判定部５２０は、以下に説明がある箇所を除き、本発明の第４の実施形態におけるリソース割当システム１００が備える緊急度判定部４２０と基本的に同様の動作を行う。傾向判定部５３２及び間隔調整部５３１は、以下に説明がある箇所を除き、基本的に、本発明の第２の実施形態におけるリソース割当システム１００が備える傾向判定部２３２及び間隔調整部２３１と同様の動作を行う。パターン照合部５３３は、以下に説明がある箇所を除き、基本的に、本発明の第３の実施形態におけるリソース割当システム３００が備えるパターン照合部３３３と同様の動作を行う。 In the present embodiment, the load acquisition unit 110 and the resource allocation unit 140 operate basically the same as the same components provided in the resource allocation system 100 according to the first embodiment of the present invention, except for the parts described below. I do. The urgency level determination unit 520 performs basically the same operation as the urgency level determination unit 420 included in the resource allocation system 100 according to the fourth embodiment of the present invention, except as described below. The trend determination unit 532 and the interval adjustment unit 531 are basically the same as the trend determination unit 232 and the interval adjustment unit 231 included in the resource allocation system 100 according to the second embodiment of the present invention, except for the portions described below. Perform the operation. The pattern matching unit 533 basically performs the same operation as that of the pattern matching unit 333 included in the resource allocation system 300 according to the third embodiment of the present invention, except for the portions described below.
　すなわち、本実施形態におけるリソース割当システム５００は、第２の実施形態における傾向判定部２３２及び間隔調整部２３１と、第３の実施形態におけるパターン照合部３３３とを共に備える点が上述した各実施形態と異なる。 That is, the resource allocation system 500 according to the present embodiment includes the tendency determination unit 232 and the interval adjustment unit 231 according to the second embodiment, and the pattern matching unit 333 according to the third embodiment. And different.
　続いて、本発明の第５の実施形態におけるリソース割当システム５００の各構成要素について説明する。 Subsequently, each component of the resource allocation system 500 according to the fifth embodiment of the present invention will be described.
　本実施形態においては、負荷取得部１１０は、予め定められた間隔でタスクを実行するサーバの負荷に関する情報を取得する。また、この間隔は、間隔調整部５３１の指示に基づいて変更される場合がある。 In the present embodiment, the load acquisition unit 110 acquires information related to the load of the server that executes a task at a predetermined interval. In addition, this interval may be changed based on an instruction from the interval adjusting unit 531.
　緊急度判定部５２０は、第４の実施形態と同様に、タスクに対するリソースの割当てに関する必要性の判定として、緊急度が程度の異なる複数の条件のいずれかを満たすかを判定する。すなわち、緊急度判定部５２０は、緊急度が上述した第１から第３の条件のいずれかを満たすか否かを判定する。間隔調整部５３１、傾向判定部５３２及びパターン照合部５３３は、緊急度判定部５２０の判定に応じてリソースの割当てに関する動作を行う。 The urgency level determination unit 520 determines, as in the fourth embodiment, whether one of a plurality of conditions with different degrees of urgency is satisfied as the determination of the necessity regarding the allocation of resources to tasks. That is, the urgency determination unit 520 determines whether or not the urgency satisfies any of the first to third conditions described above. The interval adjustment unit 531, the tendency determination unit 532, and the pattern matching unit 533 perform an operation related to resource allocation according to the determination of the urgency level determination unit 520.
　間隔調整部５３１は、緊急度判定部５２０において緊急度が例えば上述した第３の条件を満たすと判定された場合に、負荷取得部１１０による負荷に関する情報の取得間隔を変更する。本実施形態においては、間隔調整部５３１は、緊急度が第２の条件を満たすと判定された場合に、負荷取得部１１０による負荷に関する情報の取得間隔を通常の時間間隔と比較して短い時間間隔に設定する。 The interval adjustment unit 531 changes the acquisition interval of the information regarding the load by the load acquisition unit 110 when the urgency determination unit 520 determines that the urgency level satisfies, for example, the third condition described above. In the present embodiment, the interval adjusting unit 531 has a shorter time than the normal time interval for acquiring the load-related information by the load acquiring unit 110 when the urgency level is determined to satisfy the second condition. Set to interval.
　パターン照合部５３３は、緊急度判定部５２０において緊急度が例えば上述した第２の条件を満たすと判定された場合に、負荷の時間的な変化の様子が負荷の上昇を表す所定のパターンと類似するかを照合する。パターン照合部５３３は、具体的な一例として、本発明の第３の実施形態にて説明した構成を有する。 The pattern matching unit 533 is similar to a predetermined pattern in which the state of temporal change in load indicates an increase in load when the urgency level determination unit 520 determines that the urgency level satisfies the second condition described above, for example. Check if you want to. The pattern matching unit 533 has the configuration described in the third embodiment of the present invention as a specific example.
　すなわち、本実施形態におけるリソース割当システム５００は、サーバの負荷の傾向を確認する必要があると判断した場合に、緊急度の程度に応じて異なる手順でサーバの負荷の傾向の確認を行う。このようにすることで、リソース割当システム５００は、無用なリソースの割当てを抑制しつつ、負荷が上昇傾向にある場合に迅速なリソースの割当てを行うことができる。 That is, when the resource allocation system 500 in this embodiment determines that it is necessary to check the load tendency of the server, it checks the load tendency of the server in different procedures depending on the degree of urgency. By doing in this way, the resource allocation system 500 can perform quick resource allocation when the load tends to increase while suppressing unnecessary resource allocation.
　続いて、図１７に示すフローチャート参照して、本発明の第５の実施形態におけるリソース割当システム５００の動作の一例について説明する。 Subsequently, an example of the operation of the resource allocation system 500 according to the fifth embodiment of the present invention will be described with reference to the flowchart shown in FIG.
　最初に、負荷取得部１１０は、タスクを実行するサーバの負荷に関する情報を取得する（ステップＳ５０１）。次に、緊急度判定部５２０は、負荷取得部１１０における負荷に関する情報の取得間隔として、通常の時間間隔が設定されているかを判定する（ステップＳ５０２）。 First, the load acquisition unit 110 acquires information regarding the load of the server that executes the task (step S501). Next, the urgency determining unit 520 determines whether or not a normal time interval is set as an acquisition interval of information related to the load in the load acquiring unit 110 (step S502).
　負荷取得部１１０に対して、負荷に関する情報の取得間隔として通常の時間間隔が設定されている場合（ステップＳ５０２：Ｙｅｓ）には、緊急度判定部５２０は、緊急度を求める（ステップＳ５０３）。そして、緊急度判定部５２０は、緊急度が第１の条件を満たすかを判定する（ステップＳ５０４）。 When a normal time interval is set as an acquisition interval of information related to the load for the load acquisition unit 110 (step S502: Yes), the urgency level determination unit 520 obtains an urgency level (step S503). Then, the urgency level determination unit 520 determines whether the urgency level satisfies the first condition (step S504).
　ステップＳ５０３にて求めた緊急度が第１の条件を満たす場合（ステップＳ５０４：Ｙｅｓ）には、緊急度判定部５２０は、タスクに対するリソースの割当てを決定する。この決定に応じて、リソース割当部１４０は、プロセッサコア等のリソースをタスクに対して割当てる（ステップＳ５０５）。 When the urgency level obtained in step S503 satisfies the first condition (step S504: Yes), the urgency level determination unit 520 determines resource allocation to the task. In response to this determination, the resource allocation unit 140 allocates a resource such as a processor core to the task (step S505).
　なお、ステップＳ５０１からＳ５０５までの各々では、本発明の第４の実施形態におけるステップＳ４０１からＳ４０５までの各々と同様の動作が行われる。
　緊急度判定部５２０は、緊急度が先に説明した第１の条件を満たさない場合（ステップＳ５０４：Ｎｏ）に、緊急度が先に説明した第２の条件を満たすかを判定する（ステップＳ５０６）。緊急度が第２の条件を満たす（ステップＳ５０６：Ｙｅｓ）場合には、パターン照合部５３３は、負荷の時間的な変化の様子とパターンとの照合を行う（ステップＳ５０７）。 In each of steps S501 to S505, the same operations as in steps S401 to S405 in the fourth embodiment of the present invention are performed.
When the urgency level does not satisfy the first condition described above (step S504: No), the urgency level determination unit 520 determines whether the urgency level satisfies the second condition described above (step S506). ). If the degree of urgency satisfies the second condition (step S506: Yes), the pattern matching unit 533 collates the state of the load with time and the pattern (step S507).
　なお、ステップＳ５０６では、本発明の第４の実施形態におけるステップＳ４０６と同様の動作が行われる。また、ステップＳ５０７では、本発明の第３の実施形態におけるステップＳ３０６と同様の動作が行われる。すなわち、パターン照合部５３３は、先に第３の実施形態において図１３に示すフローチャートを用いて説明した動作を行う。ステップＳ５０７における処理が行われると、リソース割当システム５００は、動作を終了する。 In step S506, the same operation as in step S406 in the fourth embodiment of the present invention is performed. In step S507, the same operation as in step S306 in the third embodiment of the present invention is performed. That is, the pattern matching unit 533 performs the operation described above with reference to the flowchart shown in FIG. 13 in the third embodiment. When the processing in step S507 is performed, the resource allocation system 500 ends the operation.
　パターン照合部５３３は、測定値記憶部１５０に記憶されている測定値を参照する等して、新たに負荷に関する情報を取得せずにパターンの照合を行うことが可能である。したがって、パターン照合部５３３は、必要な場合に迅速にリソースの割当てを行うことができる。 The pattern matching unit 533 can perform pattern matching without newly acquiring information related to the load, for example, by referring to the measurement value stored in the measurement value storage unit 150. Therefore, the pattern matching unit 533 can quickly allocate resources when necessary.
　続いて、緊急度判定部５２０は、ステップＳ５０６において緊急度が第２の条件を満たさない（ステップＳ５０６：Ｎｏ）場合には、緊急度が先に説明した第３の条件を満たすかを判定する（ステップＳ５０８）。緊急度が第３の条件を満たす（ステップＳ５０８：Ｙｅｓ）場合には、間隔調整部５３１は、負荷取得部１１０における負荷に関する情報の取得間隔を、通常の時間間隔と比較して短い時間間隔に設定する（ステップＳ５０９）。ステップＳ５０９では、第２の実施形態におけるステップＳ２０７等と同様の動作が行われる。ステップＳ５０９における処理が行われると、リソース割当システム５００は、動作を終了する。 Subsequently, when the urgency level does not satisfy the second condition in step S506 (step S506: No), the urgency level determination unit 520 determines whether the urgency level satisfies the third condition described above. (Step S508). When the degree of urgency satisfies the third condition (step S508: Yes), the interval adjustment unit 531 sets the acquisition interval of information regarding the load in the load acquisition unit 110 to a shorter time interval than the normal time interval. Setting is made (step S509). In step S509, the same operation as in step S207 in the second embodiment is performed. When the processing in step S509 is performed, the resource allocation system 500 ends the operation.
　なお、緊急度判定部５２０において緊急度に関するいずれの条件も満たさないと判定された場合（ステップＳ５０８：Ｎｏ）には、リソース割当システム５００は、動作を終了する。 Note that if the urgency determination unit 520 determines that none of the conditions regarding the urgency is satisfied (step S508: No), the resource allocation system 500 ends the operation.
　また、負荷取得部１１０に対して、負荷に関する情報の取得間隔として通常の時間間隔が設定されていない場合（ステップＳ５０２：Ｎｏ）には、負荷に関する情報の取得間隔として、短い時間間隔が設定されていると想定される。この場合には、傾向判定部５３２は、負荷傾向を判定する（ステップＳ５１０）。また、ステップＳ５１０において、傾向判定部５３２は、図７に示すフローチャートに従って負荷傾向の判定を行う。ステップＳ５１０における処理が行われると、リソース割当システム５００は、動作を終了する。 In addition, when a normal time interval is not set for the load acquisition unit 110 as the information acquisition interval for the load (step S502: No), a short time interval is set as the information acquisition interval for the load. It is assumed that In this case, the tendency determination unit 532 determines a load tendency (step S510). In step S510, the trend determination unit 532 determines the load trend according to the flowchart shown in FIG. When the process in step S510 is performed, the resource allocation system 500 ends the operation.
　なお、負荷取得部１１０が定期的にサーバの負荷に関する情報を取得する場合には、リソース割当システム５００は、図１７に示すフローチャートの動作を繰り返し実行してもよい。 In addition, when the load acquisition unit 110 periodically acquires information regarding the server load, the resource allocation system 500 may repeatedly execute the operation of the flowchart shown in FIG.
　以上のとおり、本発明の第５の実施形態におけるリソース割当システム５００は、間隔調整部５３１と、傾向判定部５３２と、パターン照合部５３３とを備える。したがって、本実施形態におけるリソース割当システム５００は、第２の実施形態におけるリソース割当システム２００及び第３の実施形態におけるリソース割当システム３００の各々と同様の効果を奏する。 As described above, the resource allocation system 500 according to the fifth exemplary embodiment of the present invention includes the interval adjustment unit 531, the tendency determination unit 532, and the pattern matching unit 533. Therefore, the resource allocation system 500 in the present embodiment has the same effects as the resource allocation system 200 in the second embodiment and the resource allocation system 300 in the third embodiment.
　また、本実施形態におけるリソース割当システム５００では、サーバの負荷の傾向を更に確認する必要があると判断された場合に、負荷の傾向を異なる手段を用いて確認することが可能となる。 In addition, in the resource allocation system 500 according to the present embodiment, when it is determined that it is necessary to further check the load trend of the server, it is possible to check the load trend using different means.
　すなわち、パターン照合部５３３が、測定値記憶部１５０に記憶された負荷とパターンとを照合して負荷の傾向を判断することで、負荷の傾向が迅速に判断される。また、本実施形態においては、サーバの負荷の傾向を確認する必要があると判定され、かつ、緊急度の程度が相対的に高い場合に、上述したパターンの照合が行われる。したがって、負荷が増加の傾向にある場合に、迅速なリソースの割当てが可能となる。 That is, the pattern matching unit 533 compares the load stored in the measured value storage unit 150 with the pattern to determine the load tendency, whereby the load tendency is quickly determined. In the present embodiment, the above-described pattern matching is performed when it is determined that it is necessary to check the load trend of the server and the degree of urgency is relatively high. Therefore, when the load tends to increase, it is possible to quickly allocate resources.
　また、間隔調整部５３１及び傾向判定部５３２が、新たに負荷取得部１１０にて通常と比較して短い間隔で取得された負荷を用いて負荷の傾向を判断することで、負荷の傾向が正確に判断される。また、本実施形態においては、サーバの負荷の傾向を確認する必要があると判定され、かつ、緊急度の程度が相対的に低い場合に、上述した短い間隔で取得された負荷を用いて負荷の傾向が判断される。したがって、負荷が増加の傾向にある場合に、適切なリソースの割当てが行われる。 In addition, the interval adjustment unit 531 and the trend determination unit 532 newly determine the load trend using the load acquired by the load acquisition unit 110 at a shorter interval than normal, so that the load trend is accurate. Will be judged. Further, in this embodiment, when it is determined that it is necessary to check the load tendency of the server and the degree of urgency is relatively low, the load is obtained using the load acquired at a short interval described above. The tendency is determined. Therefore, when the load tends to increase, appropriate resource allocation is performed.
　したがって、本実施形態におけるリソース割当システム５００は、他の実施形態と比較して更に高い精度で負荷の変動に応じた適切なリソースの割当てを可能にする。 Therefore, the resource allocation system 500 in the present embodiment enables appropriate resource allocation in accordance with load fluctuations with higher accuracy than in the other embodiments.
　（実施例１）
　次に、本発明の各実施形態におけるリソース割当システムを用いたリソース割当の例を説明する。最初に、本発明の第２の実施形態におけるリソース割当システム２００を用いたリソース割当の例を説明する。図１８は、実施例１におけるリソース割当システムとタスク実行サーバの構成を示す図である。 Example 1
Next, an example of resource allocation using the resource allocation system in each embodiment of the present invention will be described. First, an example of resource allocation using the resource allocation system 200 according to the second embodiment of the present invention will be described. FIG. 18 is a diagram illustrating the configuration of the resource allocation system and the task execution server in the first embodiment.
　本実施例においては、リソース割当システム２００と、リソースの割当ての対象となるタスクが実行されるタスク実行サーバ１１は、図１８のように通信ネットワーク１２を介して接続される。 In this embodiment, the resource allocation system 200 and the task execution server 11 on which a task to be allocated a resource is executed are connected via the communication network 12 as shown in FIG.
　タスク実行サーバ１１は、例えば、図２８に示す上述した情報処理装置１０００である。タスク実行サーバ１１においては、タスクとして、他のサーバ等からの要求に応じた処理を行うプログラムが実行される。プログラムは、例えば情報処理装置１０００の記憶装置１００５に記憶されており、プログラムの実行に際して記憶装置１００５から読み出されてＲＡＭ１００３に展開される。また、タスク実行サーバ１１は、外部通信ネットワーク１３を介して他のサーバ等と接続する。 The task execution server 11 is, for example, the above-described information processing apparatus 1000 illustrated in FIG. In the task execution server 11, a program that performs processing in response to a request from another server or the like is executed as a task. The program is stored in the storage device 1005 of the information processing apparatus 1000, for example, and is read from the storage device 1005 and expanded in the RAM 1003 when the program is executed. Further, the task execution server 11 is connected to other servers and the like via the external communication network 13.
　本実施例では、タスク実行サーバ１１は、複数のＣＰＵを備える。または、タスク実行サーバ１１は、ＣＰＵとして、複数のプロセッサコアを有するマルチコアプロセッサを備えてもよい。上述したプログラムは、当初は、複数のＣＰＵのうちの一つのＣＰＵ（又は一つのプロセッサコア）によって実行されることを想定する。 In this embodiment, the task execution server 11 includes a plurality of CPUs. Alternatively, the task execution server 11 may include a multi-core processor having a plurality of processor cores as a CPU. It is assumed that the above-described program is initially executed by one CPU (or one processor core) among a plurality of CPUs.
　また、本実施例においては、リソース割当システム２００は、図２８に示す上述した情報処理装置１０００にて実行されるプログラムとして実現される。リソース割当システム２００とタスク実行サーバ１１は、通信ネットワーク１２を介して接続される。リソース割当システム２００（を実現するプログラム）は、タスク実行サーバ１１における負荷に関する情報を、当該通信ネットワーク１２を介して取得する。 In this embodiment, the resource allocation system 200 is realized as a program executed by the information processing apparatus 1000 described above shown in FIG. The resource allocation system 200 and the task execution server 11 are connected via the communication network 12. The resource allocation system 200 (a program that realizes the resource allocation system 200) acquires information regarding the load on the task execution server 11 via the communication network 12.
　測定値記憶部１５０は、情報処理装置１０００の記憶装置１００５に保持されるリレーショナルデータベース（Ｒｅｌａｔｉｏｎａｌ　Ｄａｔａｂａｓｅ、以下「ＲＤＢ」とする）のテーブルとして実現される。ＲＤＢのテーブルには、属性として、負荷の測定時刻、当該測定時刻におけるＣＰＵ負荷、及び当該測定時刻とその前の測定時刻との測定時刻の間隔が含まれる。負荷取得部１１０による負荷に関する情報の取得間隔は、予めＲＡＭ１００３に保持される。 The measurement value storage unit 150 is realized as a table of a relational database (Relational Database, hereinafter referred to as “RDB”) held in the storage device 1005 of the information processing apparatus 1000. The RDB table includes, as attributes, the load measurement time, the CPU load at the measurement time, and the measurement time interval between the measurement time and the previous measurement time. The acquisition interval of the information regarding the load by the load acquisition unit 110 is stored in the RAM 1003 in advance.
　本実施例において、負荷取得部１１０は、通常の時間間隔を５秒、短い時間間隔を１秒として負荷を取得する。取得した負荷に関する情報は、測定が行われる度に、測定値記憶部１５０となるＲＤＢを構成するテーブルのレコードとして記録される。測定値記憶部１５０に含まれるテーブルの属性である測定時刻にはインデックスが付されており、値の探索又は整列が高速に行われる。 In this embodiment, the load acquisition unit 110 acquires a load with a normal time interval of 5 seconds and a short time interval of 1 second. Information about the acquired load is recorded as a record of a table constituting the RDB that becomes the measurement value storage unit 150 each time measurement is performed. An index is attached to the measurement time, which is an attribute of the table included in the measurement value storage unit 150, and the search or alignment of values is performed at high speed.
　本実施例においては、緊急度判定部１２０は、具体的には、緊急度Ｅを以下の（３）式を用いて求める。（３）式は、上述した（１）式において、ａを１．０、ｂを０．５とした式である。 In the present embodiment, the urgency level determination unit 120 specifically calculates the urgency level E using the following equation (3). The expression (3) is an expression in which a is 1.0 and b is 0.5 in the above-described expression (1).
Figure JPOXMLDOC01-appb-M000003
 　なお、Ｗ及びΔＷは、（１）式における同様の記号と同様の意味である。“＊”は、乗算を表す記号である。また、緊急度判定部１２０は、緊急度が６０を超える場合に、緊急度が第２の条件を満たすと判定する。緊急度判定部１２０は、緊急度が９０を超える場合に、緊急度が第１の条件を満たすと判定する。
Figure JPOXMLDOC01-appb-M000003
 W and ΔW have the same meanings as the same symbols in the equation (1). “*” Is a symbol representing multiplication. Further, the urgency level determination unit 120 determines that the urgency level satisfies the second condition when the urgency level exceeds 60. The urgency level determination unit 120 determines that the urgency level satisfies the first condition when the urgency level exceeds 90.
　傾向判定部２３２は、緊急度が第１の条件を満たし、短い時間間隔にて負荷が取得される場合において、負荷が続けて３回増加すると、負荷が増加傾向にあるとしてタスクに対してリソースを割当てると判定する。また、この場合に、傾向判定部２３２は、負荷が続けて３回減少すると、負荷が減少傾向にあるとしてタスクに対してリソースを割当てないと判定する。 When the urgency satisfies the first condition and the load is acquired at a short time interval, if the load continues to increase three times, the tendency determination unit 232 determines that the load tends to increase, Is determined to be assigned. Also, in this case, when the load continues to decrease three times, the trend determination unit 232 determines that the load is in a decreasing trend and does not allocate resources to the task.
　図１９は、負荷取得部１１０にて取得された負荷に関する情報と、当該負荷に対して求められた緊急度の例を示す表である。図１９において、負荷に関する情報はＣＰＵの負荷を示す。また、図２０は、図１９に示す負荷について、横軸を測定時刻とし、縦軸を負荷として表したグラフである。図１９に示す例において、時刻５：４２：４０（「５：４２：４０」は、５時４２分４０秒を表す。以下、同様に時刻を表す）における緊急度は、時刻５：４２：４０における負荷の値である２７と、時刻５：４２：３５における負荷の値である２５とを用いて次のように求められる。 FIG. 19 is a table showing information about the load acquired by the load acquisition unit 110 and an example of the degree of urgency obtained for the load. In FIG. 19, the information regarding the load indicates the load of the CPU. FIG. 20 is a graph showing the load shown in FIG. 19 with the horizontal axis as the measurement time and the vertical axis as the load. In the example shown in FIG. 19, the urgency level at time 5:42:40 (“5:42:40” represents 5:42:40. Hereinafter, the time is similarly expressed) is time 5:42: Using the load value 27 at 40 and the load value 25 at time 5:42:35, the following is obtained.
Figure JPOXMLDOC01-appb-M000004
 　この時刻においては、緊急度は、上述した第１及び第２のいずれの条件も満たさない。したがって、緊急度判定部１２０は、リソースの割当て等に関する特段の処理を行わず、次の測定結果に対して緊急度の導出及び緊急度の判定を行う。同様に、時刻５：４２：５５までの各々の時刻にて計測された負荷に基づき求められた緊急度は、いずれも上述した第１及び第２のいずれの条件も満たさない。したがって、緊急度判定部１２０は、リソースの割当て等に関する特段の処理を行わない。
Figure JPOXMLDOC01-appb-M000004
 At this time, the degree of urgency does not satisfy either of the first and second conditions described above. Therefore, the urgency level determination unit 120 performs derivation of the urgency level and determination of the urgency level for the next measurement result without performing any special processing related to resource allocation or the like. Similarly, the degree of urgency obtained based on the load measured at each time until 5:42:55 does not satisfy any of the first and second conditions described above. Therefore, the urgency determination unit 120 does not perform special processing related to resource allocation or the like.
　一方、時刻５：４３：００において、緊急度判定部１２０が、当該時刻と時刻５：４２：５５にて計測された負荷に基づいて緊急度を計算すると、緊急度は１００と求められる。すなわち、緊急度は、上述した第１の条件を満たす。この場合には、緊急度判定部１２０は、タスクに対しリソースを割当てる旨を決定する。この決定に応じて、リソース割当部１４０は、タスクに対してリソースとしてＣＰＵ（又はプロセッサコア）を割当てる。すなわち、リソースの割当ての対象となるタスクは、二つのＣＰＵ（又はプロセッサコア）を用いて実行される。 On the other hand, at time 5:43:00, when the emergency level determination unit 120 calculates the emergency level based on the time and the load measured at time 5:42:55, the emergency level is calculated as 100. That is, the urgency satisfies the first condition described above. In this case, the urgency determining unit 120 determines to allocate resources to the task. In response to this determination, the resource allocation unit 140 allocates a CPU (or processor core) as a resource to the task. That is, a task to which resources are allocated is executed using two CPUs (or processor cores).
　図２１は、負荷取得部１１０にて取得された負荷に関する情報と、当該負荷に対して求められた緊急度の別の例を示す表である。図２１において、負荷に関する情報は、ＣＰＵの負荷を示す。また、図２２は、図２１に示す負荷について、横軸を測定時刻とし、縦軸を負荷として表したグラフである。 FIG. 21 is a table showing another example of information on the load acquired by the load acquisition unit 110 and the degree of urgency obtained for the load. In FIG. 21, the information regarding the load indicates the load on the CPU. FIG. 22 is a graph showing the load shown in FIG. 21 with the horizontal axis as the measurement time and the vertical axis as the load.
　図２１に示す表は、図１８に示すリソース割当システム２００において、図１９の表が示す負荷とは異なる時刻に取得された負荷等の情報を示す。図２１に示す表において、緊急度は、先に説明した（３）式を用いて緊急度を求められている。 The table shown in FIG. 21 shows information such as the load acquired at a different time from the load shown in the table of FIG. 19 in the resource allocation system 200 shown in FIG. In the table shown in FIG. 21, the urgency level is calculated using the equation (3) described above.
　図２１に示す表の例では、時刻７：１２：２０及び時刻７：１２：２５に求められた緊急度の各々は、第１及び第２の条件のいずれも満たさない。そして、時刻７：１２：３０に取得された負荷に基づいて求められた緊急度は、上述した第２の条件を満たす。したがって、間隔調整部２３１は、負荷取得部１１０による負荷の取得の間隔を短い時間間隔に変更する。 In the example of the table shown in FIG. 21, each of the urgency levels obtained at time 7:12:20 and time 7:12:25 does not satisfy both the first and second conditions. And the urgency degree calculated | required based on the load acquired at the time 7:12:30 satisfy | fills the 2nd condition mentioned above. Therefore, the interval adjustment unit 231 changes the load acquisition interval by the load acquisition unit 110 to a short time interval.
　負荷取得部１１０は、この場合に、１秒間隔で負荷に関する情報を取得する。しかしながら、１秒間隔で負荷が取得される時刻７：１２：３１以降の３つの時点の各時刻において取得された負荷は、いずれもその直前に取得された負荷と比較して小さな値である。すなわち、負荷が続けて３回減少している。 In this case, the load acquisition unit 110 acquires information about the load at intervals of 1 second. However, the load acquired at each of the three times after the time 7:12:31 when the load is acquired at 1 second intervals is a small value compared to the load acquired immediately before that. That is, the load continues to decrease three times.
　したがって、傾向判定部２３２は、負荷が減少傾向にあるとして、タスクに対してリソースを割当てないと判定する。そして、この判定に基づいて、間隔調整部２３１は、負荷の測定間隔を通常の間隔に変更する。 Therefore, the trend determining unit 232 determines that the resource is not allocated to the task, assuming that the load tends to decrease. Based on this determination, the interval adjustment unit 231 changes the load measurement interval to a normal interval.
　図２３は、負荷取得部１１０にて取得された負荷に関する情報と、当該負荷に対して求められた緊急度の更に別の例を示す表である。図２３において、負荷に関する情報は、ＣＰＵの負荷を示す。また、図２４は、図２３に示す負荷について、横軸を測定時刻とし、縦軸を負荷として表したグラフである。 FIG. 23 is a table showing still another example of the information about the load acquired by the load acquisition unit 110 and the urgency level obtained for the load. In FIG. 23, the information regarding the load indicates the load on the CPU. FIG. 24 is a graph showing the load shown in FIG. 23 with the horizontal axis as the measurement time and the vertical axis as the load.
　図２３に示す表は、図１８に示すリソース割当システム２００において、図１９又は図２３の表が示す負荷とは異なる時刻に取得された負荷等の情報を示す。図２３に示す表においては、緊急度は、先に説明した（３）式を用いて緊急度を求められている。 The table shown in FIG. 23 shows information such as the load acquired at a different time from the load shown in the table of FIG. 19 or FIG. 23 in the resource allocation system 200 shown in FIG. In the table shown in FIG. 23, the urgency level is obtained using the equation (3) described above.
　図２３に示す表の例では、時刻８：２５：２０及び時刻８：２５：２５に求められた緊急度の各々は、第１及び第２の条件のいずれも満たさない。そして、時刻８：２５：３０に取得された負荷に基づいて求められた緊急度は、上述した第２の条件を満たす。したがって、間隔調整部２３１は、負荷取得部１１０による負荷の取得の間隔を短い時間間隔に変更する。 In the example of the table shown in FIG. 23, each of the urgency levels obtained at time 8:25:20 and time 8:25:25 does not satisfy both the first and second conditions. And the urgency degree calculated | required based on the load acquired at the time 8:25:30 satisfy | fills the 2nd condition mentioned above. Therefore, the interval adjustment unit 231 changes the load acquisition interval by the load acquisition unit 110 to a short time interval.
　負荷取得部１１０は、この場合に、１秒間隔で負荷に関する情報を取得する。そして、１秒間隔で負荷が取得される時刻８：２５：３１以降の３つの時点の各時刻において取得された負荷は、いずれもその直前に取得された負荷と比較して大きな値である。すなわち、負荷が続けて３回増加している。 In this case, the load acquisition unit 110 acquires information about the load at intervals of 1 second. And the load acquired at each time of three time points after the time 8:25:31 when the load is acquired at intervals of 1 second is larger than the load acquired immediately before that. That is, the load continues to increase three times.
　したがって、傾向判定部２３２は、負荷が増加傾向にあるとして、タスクに対してリソースを割当てると判定する。そして、この判定に基づいて、傾向判定部２３２は、タスクに対するリソースの割当てを決定する。この決定に応じて、リソース割当部１４０は、タスクに対してリソースとしてＣＰＵ（又はプロセッサコア）を割当てる。すなわち、リソースの割当ての対象となるタスクは、二つのＣＰＵ（又はプロセッサコア）を用いて実行される。また、間隔調整部２３１は、リソース割当部１４０によるリソースの割当てが行われると、負荷の測定間隔を通常の間隔に変更する。 Therefore, the trend determination unit 232 determines that resources are allocated to the task, assuming that the load tends to increase. Then, based on this determination, the trend determination unit 232 determines resource allocation for the task. In response to this determination, the resource allocation unit 140 allocates a CPU (or processor core) as a resource to the task. That is, a task to which resources are allocated is executed using two CPUs (or processor cores). In addition, when the resource allocation unit 140 performs resource allocation, the interval adjustment unit 231 changes the load measurement interval to a normal interval.
　このように、本実施例においては、リソース割当システム２００によって、緊急度が第１の条件を満たすと判定された場合に速やかにタスクに対してリソースの割当てが行われた。また、緊急度が第２の条件を満たすと判定された場合には、負荷の測定の間隔を短い時間間隔に設定された。そして、負荷が増加傾向にある場合に、タスクに対してリソースの割当てが行われた。 As described above, in this embodiment, when the resource allocation system 200 determines that the urgency level satisfies the first condition, the resource is quickly allocated to the task. When it was determined that the urgency level satisfies the second condition, the load measurement interval was set to a short time interval. When the load tends to increase, resources are allocated to the tasks.
　すなわち、リソース割当システム２００によって、負荷の変動に応じた適切なリソースの割当てが行われる。 That is, the resource allocation system 200 performs appropriate resource allocation according to load fluctuations.
　（実施例２）
　次に、本発明の各実施形態におけるリソース割当システムを用いたリソース割当の別の例を説明する。最初に、本発明の第３の実施形態におけるリソース割当システム３００を用いたリソース割当の例を説明する。 (Example 2)
Next, another example of resource allocation using the resource allocation system in each embodiment of the present invention will be described. First, an example of resource allocation using the resource allocation system 300 according to the third embodiment of the present invention will be described.
　本実施例においては、リソース割当システム３００と、リソースの割当ての対象となるタスクが実行されるタスク実行サーバ１１は、図２５のように通信ネットワークを介して接続される。タスク実行サーバ１１は、先に説明した実施例１の場合と同様に、例えば、図２８に示す上述した情報処理装置１０００である。
また、実施例１の倍と同様に、タスク実行サーバ１１は、複数のＣＰＵ（又は複数のプロセッサコア）を備えていることを想定する。そして、タスク実行サーバにて実行されるプログラムは、当初は、複数のＣＰＵのうち、一つのＣＰＵ（又は一つのプロセッサコア）によって実行されることを想定する。 In the present embodiment, the resource allocation system 300 and the task execution server 11 on which a task to be allocated a resource is executed are connected via a communication network as shown in FIG. The task execution server 11 is the above-described information processing apparatus 1000 shown in FIG. 28, for example, as in the case of the first embodiment described above.
As in the case of the first embodiment, it is assumed that the task execution server 11 includes a plurality of CPUs (or a plurality of processor cores). The program executed by the task execution server is initially assumed to be executed by one CPU (or one processor core) among a plurality of CPUs.
　また、本実施例においては、リソース割当システム３００は、第１の実施例において説明したリソース割当システム２００と同様に、図２８に示す上述した情報処理装置１０００にて実行されるプログラムとして実現される。パターン照合部３３３のパターン記憶部３３５は、情報処理装置１０００の記憶装置１００５に保持されるＲＤＢのテーブルとして実現される。図２６は、パターン記憶部３３５に記憶されるパターンの例である。パターン記憶部３３５は、５つのパターンを記憶する。また、パターンの各々は、１０個の負荷を含む。本実施例では、負荷は、５秒毎の負荷の変動を表す。 Further, in this embodiment, the resource allocation system 300 is realized as a program executed by the above-described information processing apparatus 1000 shown in FIG. 28, similarly to the resource allocation system 200 described in the first embodiment. . The pattern storage unit 335 of the pattern matching unit 333 is realized as an RDB table held in the storage device 1005 of the information processing apparatus 1000. FIG. 26 is an example of a pattern stored in the pattern storage unit 335. The pattern storage unit 335 stores five patterns. Each of the patterns includes 10 loads. In this embodiment, the load represents a change in load every 5 seconds.
　本実施例においては、負荷取得部１１０は、負荷に関する情報を５秒間隔にて取得する。また、緊急度判定部１２０は、具体的には、緊急度Ｅを上述した（３）式を用いて求める。また、緊急度判定部１２０は、緊急度が６０を超える場合に、緊急度が第２の条件を満たすと判定する。緊急度判定部１２０は、緊急度が９０を超える場合に、緊急度が第１の条件を満たすと判定する。測定値記憶部１５０は、負荷取得部１１０にて取得された負荷と、上述したパターン記憶部３３５に記憶されるパターンとの比較を可能にするよう、直近に取得された負荷から少なくとも１０回分の負荷を記憶する。パターン照合部３３３の類似度判定部３３４は、類似度に関する閾値Ｒ０を１．０とする。 In this embodiment, the load acquisition unit 110 acquires information about the load at intervals of 5 seconds. In addition, the urgency level determination unit 120 specifically calculates the urgency level E using the above-described equation (3). Further, the urgency level determination unit 120 determines that the urgency level satisfies the second condition when the urgency level exceeds 60. The urgency level determination unit 120 determines that the urgency level satisfies the first condition when the urgency level exceeds 90. The measurement value storage unit 150 is at least 10 times from the most recently acquired load so as to enable comparison between the load acquired by the load acquisition unit 110 and the pattern stored in the pattern storage unit 335 described above. Memorize the load. The similarity determination unit 334 of the pattern matching unit 333 sets the threshold value R0 regarding the similarity to 1.0.
　図２７は、負荷取得部１１０にて取得された負荷に関する情報と、当該負荷に対して求められた緊急度の例を示す表である。図２７において、負荷に関する情報は、ＣＰＵの負荷を示す。 FIG. 27 is a table showing information on the load acquired by the load acquisition unit 110 and an example of the degree of urgency obtained for the load. In FIG. 27, the information regarding the load indicates the load on the CPU.
　図２７に示す表の例では、時刻９：５３：４０以降の各々の時刻において負荷が取得された場合に、緊急度判定部１２０は緊急度を求める。時刻９：５４：２０以前の時刻においては、緊急度は第１及び第２の条件のいずれも満たさない。そして、時刻９：５４：２５において、緊急度判定部１２０が、緊急度を計算すると、緊急度は８２．０と求められる。すなわち、緊急度は第２の条件を満たす。 In the example of the table shown in FIG. 27, when the load is acquired at each time after 9:53:40, the emergency level determination unit 120 calculates the emergency level. At times before 9:54:20, the urgency level does not satisfy both the first and second conditions. Then, at time 9:54:25, when the emergency level determination unit 120 calculates the emergency level, the emergency level is calculated to be 82.0. That is, the urgency satisfies the second condition.
　そこで、類似度判定部３３４は、負荷の時間的な変化の様子とパターンとの照合を行う。すなわち、類似度判定部３３４は、時刻９：５３：４０から時刻９：５４：２５までに計測された負荷と、図２６に示すパターンの各々との類似度を求める。ＩＤが１のパターンとの類似度は、上述した（２）式を用いて、以下に示す計算式に沿って求められる。
Figure JPOXMLDOC01-appb-I000001
Therefore, the similarity determination unit 334 collates the state of the load with time and the pattern. That is, the similarity determination unit 334 obtains the similarity between the load measured from time 9:53:40 to time 9:54:25 and each of the patterns shown in FIG. The similarity with the pattern with ID 1 is obtained using the above-described equation (2) according to the following equation.
Figure JPOXMLDOC01-appb-I000001
　すなわち、ＩＤが１のパターンとの類似度は４．３である。同様に、ＩＤが２のパターンとの類似度は２．７、ＩＤが３のパターンとの類似度は２．０、ＩＤが４のパターンとの類似度は２．９、ＩＤが５のパターンとの類似度は０．８と求められる。そして、ＩＤが５のパターンに関して求められた類似度は、上述した閾値Ｒ０の値である１．０よりも小さい。したがって、類似度判定部３３４は、負荷が増加傾向にあると判定する。 That is, the similarity with the pattern with ID 1 is 4.3. Similarly, the similarity with the pattern with ID 2 is 2.7, the similarity with the pattern with ID 3 is 2.0, the similarity with the pattern with ID 4 is 2.9, and the pattern with ID 5 The degree of similarity is obtained as 0.8. And the similarity calculated | required regarding the pattern with ID 5 is smaller than 1.0 which is the value of threshold value R0 mentioned above. Therefore, the similarity determination unit 334 determines that the load tends to increase.
　そこで、類似度判定部３３４は、タスクに対するリソースの割当てを決定する。当該決定に応じて、リソース割当部１４０は、タスクに対してリソースとしてＣＰＵ（又はプロセッサコア）を割当てる。すなわち、リソースの割当ての対象となるタスクは、二つのＣＰＵ（又はプロセッサコア）を用いて実行される。 Therefore, the similarity determination unit 334 determines resource allocation for the task. In response to the determination, the resource allocation unit 140 allocates a CPU (or processor core) as a resource to the task. That is, a task to which resources are allocated is executed using two CPUs (or processor cores).
　すなわち、本実施形態においても、リソース割当システム３００によって、負荷の変動に応じた適切なリソースの割当てが行われる。 That is, also in the present embodiment, the resource allocation system 300 performs appropriate resource allocation according to load fluctuations.
　以上、実施形態及び実施例を参照して本発明を説明したが、本発明は上記実施形態及び実施例に限定されるものではない。本発明の構成や詳細には、本発明のスコープ内で当業者が理解し得る様々な変更をすることができる。また、各実施形態における構成は、本発明のスコープを逸脱しない限りにおいて、互いに組み合わせることが可能である。 As mentioned above, although this invention was demonstrated with reference to embodiment and an Example, this invention is not limited to the said embodiment and Example. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention. The configurations in the embodiments can be combined with each other without departing from the scope of the present invention.
　この出願は、２０１５年７月２１日に出願された日本出願特願２０１５－１４４００３を基礎とする優先権を主張し、その開示の全てをここに取り込む。 This application claims priority based on Japanese Patent Application No. 2015-144003 filed on July 21, 2015, the entire disclosure of which is incorporated herein.
　１１　　タスク実行サーバ
　１２　　通信ネットワーク
　１３　　外部通信ネットワーク
　１００、２００、３００、４００、５００　　リソース割当システム
　１１０　　負荷取得部
　１２０、４２０、５２０　　緊急度判定部
　１３０　　負荷判定部
　２３１、５３１　　間隔調整部
　２３２、５３２　　傾向判定部
　３３３、５３３　　パターン照合部
　３３４　　類似度判定部
　３３５　　パターン記憶部
　１４０　　リソース割当部
　１５０　　測定値記憶部
　１０００　　情報処理装置
　１００１　　ＣＰＵ
　１００２　　ＲＯＭ
　１００３　　ＲＡＭ
　１００４　　プログラム
　１００５　　記憶装置
　１００６　　記憶媒体
　１００７　　ドライブ装置
　１００８　　通信インターフェース
　１００９　　通信ネットワーク
　１０１０　　入出力インターフェース
　１０１１　　バス DESCRIPTION OF SYMBOLS 11 Task execution server 12 Communication network 13 External communication network 100, 200, 300, 400, 500 Resource allocation system 110 Load acquisition part 120, 420, 520 Urgent degree judgment part 130 Load judgment part 231, 531 Interval adjustment part 232, 532 Trend Determination unit 333, 533 Pattern matching unit 334 Similarity determination unit 335 Pattern storage unit 140 Resource allocation unit 150 Measurement value storage unit 1000 Information processing device 1001 CPU
1002 ROM
1003 RAM
1004 Program 1005 Storage Device 1006 Storage Medium 1007 Drive Device 1008 Communication Interface 1009 Communication Network 1010 Input / Output Interface 1011 Bus

Claims (10)

  1. 　タスクを実行するサーバの負荷に関する情報を取得する負荷取得手段と、
    　前記負荷に関する情報に基づいて、前記タスクに対するリソースの割当てに関する必要性を判定する緊急度判定手段と、
    　前記緊急度判定手段が前記サーバの負荷の傾向を確認する必要があると判定した場合に、前記サーバの負荷の傾向を判定する負荷判定手段と、
    　前記緊急度判定手段が前記タスクにリソースを割当てる必要があると判定した場合に、前記タスクに前記リソースを割当てるリソース割当手段とを備える、リソース割当システム。     Load acquisition means for acquiring information on the load of the server executing the task;
    Urgency determination means for determining necessity regarding allocation of resources to the task based on the information regarding the load;
    A load determination unit that determines a load trend of the server when the urgency determination unit determines that the load trend of the server needs to be confirmed;
    A resource allocation system comprising: a resource allocation unit that allocates the resource to the task when the urgency determination unit determines that a resource needs to be allocated to the task.
  2. 　前記負荷取得手段による前記情報の取得間隔を変更する間隔調整手段を備える、請求項１に記載のリソース割当システム。 2. The resource allocation system according to claim 1, further comprising an interval adjustment unit that changes an acquisition interval of the information by the load acquisition unit.
  3. 　前記間隔調整手段は、前記緊急度判定手段が前記サーバの負荷の傾向を確認する必要があると判定した場合に、前記情報の取得間隔をそれまでより短い時間間隔に変更する、請求項２に記載のリソース割当システム。 The said interval adjustment means changes the acquisition interval of the said information to the time interval shorter than before, when the said emergency degree determination means determines that it is necessary to confirm the tendency of the load of the said server. The resource allocation system described.
  4. 　前記間隔調整手段は、前記緊急度判定手段が前記サーバの負荷の傾向を確認する必要があると判定した場合に、前記情報の取得間隔を、前記情報を用いて求められる緊急度の程度に応じて異なる短い時間間隔に変更する、請求項２に記載のリソース割当システム。 The interval adjusting unit determines the information acquisition interval according to the degree of urgency obtained using the information when the urgency determining unit determines that it is necessary to check the load trend of the server. The resource allocation system according to claim 2, wherein the resource allocation system is changed to a different short time interval.
  5. 　前記負荷判定手段は、前記情報の取得間隔が所定の取得間隔と比較して短くされている場合に、前記サーバの負荷の傾向を判定する、請求項３又は４に記載のリソース割当システム。 The resource allocation system according to claim 3 or 4, wherein the load determination unit determines a tendency of the load on the server when the information acquisition interval is shorter than a predetermined acquisition interval.
  6. 　前記負荷判定手段は、前記負荷取得手段において取得された前記情報が、所定の回数に亘って前記サーバの負荷の増加を示す場合に、前記サーバの負荷が増加傾向にあると判定する、請求項５に記載のリソース割当システム。 The load determination unit determines that the load on the server tends to increase when the information acquired by the load acquisition unit indicates an increase in the load on the server over a predetermined number of times. 5. The resource allocation system according to 5.
  7. 　前記負荷判定手段は、前記サーバの負荷の時間的な変化の様子が所定のパターンと類似するか否かを判定することで、前記サーバの負荷の傾向を判定する、請求項１から６のいずれか一項に記載のリソース割当システム。 7. The load determination unit according to claim 1, wherein the load determination unit determines a tendency of the load on the server by determining whether a temporal change in the load on the server is similar to a predetermined pattern. The resource allocation system according to claim 1.
  8. 　前記リソース割当手段は、前記負荷が増加傾向にあると前記負荷判定手段が判定した場合に前記タスクに前記リソースを割当てる、請求項１から７のいずれか一項に記載のリソース割当システム。 The resource allocation system according to any one of claims 1 to 7, wherein the resource allocation unit allocates the resource to the task when the load determination unit determines that the load tends to increase.
  9. 　タスクを実行するサーバの負荷に関する情報を取得し、
    　前記負荷に関する情報を用いて求められる緊急度に基づいて、前記タスクに対するリソースの割当てに関する必要性を判定し、
    　前記必要性として、前記サーバの負荷の傾向を確認する必要があると判定された場合に、前記サーバの負荷の傾向を判定し、
    　前記必要性として、前記緊急度判定手段が前記タスクにリソースを割当てる必要があると判定された場合に、前記タスクに前記リソースを割当てる、リソース割当方法。     Get information about the load on the server that executes the task,
    Determining the need for allocation of resources for the task based on the urgency required using information about the load;
    If it is determined that it is necessary to check the load trend of the server as the necessity, determine the load trend of the server,
    A resource allocation method for allocating the resource to the task when the necessity determination unit determines that the resource needs to be allocated to the task as the necessity.
  10. 　コンピュータに、
    　タスクを実行するサーバの負荷に関する情報を取得する処理と、
    　前記負荷に関する情報を用いて求められる緊急度に基づいて、前記タスクに対するリソースの割当てに関する必要性を判定する処理と、
    　前記必要性の判定において、前記サーバの負荷の傾向を確認する必要があると判定された場合に、前記サーバの負荷の傾向を判定する処理と、
    　前記必要性の判定において、前記タスクにリソースを割当てる必要があると判定された場合に、前記タスクに前記リソースを割当てる処理とを実行させるプログラムを格納したコンピュータ読み取り可能記録媒体。     On the computer,
    Processing to obtain information about the load on the server that executes the task;
    A process of determining the necessity for allocation of resources for the task based on the urgency obtained using the information about the load;
    In the determination of the necessity, when it is determined that it is necessary to check the load trend of the server, a process of determining the load trend of the server;
    A computer-readable recording medium storing a program for executing a process of allocating the resource to the task when it is determined in the determination of necessity that the resource needs to be allocated to the task.
PCT/JP2016/003388 2015-07-21 2016-07-19 Resource allocation system, resource allocation method, and computer-readable recording medium WO2017013872A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2017529458A JPWO2017013872A1 (en) 2015-07-21 2016-07-19 Resource allocation system, resource allocation method, and computer-readable recording medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2015144003 2015-07-21
JP2015-144003 2015-07-21

Publications (1)

Publication Number Publication Date
WO2017013872A1 true WO2017013872A1 (en) 2017-01-26

Family

ID=57833863

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2016/003388 WO2017013872A1 (en) 2015-07-21 2016-07-19 Resource allocation system, resource allocation method, and computer-readable recording medium

Country Status (2)

Country Link
JP (1) JPWO2017013872A1 (en)
WO (1) WO2017013872A1 (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2011118525A (en) * 2009-12-01 2011-06-16 Hitachi Ltd Server management apparatus, server management method, and server management program
JP2011210225A (en) * 2010-07-14 2011-10-20 Nomura Research Institute Ltd System and method for information processing
JP2012181580A (en) * 2011-02-28 2012-09-20 Fujitsu Ltd Resource control device, resource control method, and resource control program

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2011118525A (en) * 2009-12-01 2011-06-16 Hitachi Ltd Server management apparatus, server management method, and server management program
JP2011210225A (en) * 2010-07-14 2011-10-20 Nomura Research Institute Ltd System and method for information processing
JP2012181580A (en) * 2011-02-28 2012-09-20 Fujitsu Ltd Resource control device, resource control method, and resource control program

Also Published As

Publication number Publication date
JPWO2017013872A1 (en) 2018-05-24

Similar Documents

Publication Publication Date Title
US10896055B2 (en) Capacity risk management for virtual machines
JP5218390B2 (en) Autonomous control server, virtual server control method and program
JP5544967B2 (en) Virtual machine management program and virtual machine management apparatus
JP5332065B2 (en) Cluster configuration management method, management apparatus, and program
JP6233413B2 (en) Task assignment determination device, control method, and program
JP4982578B2 (en) Resource allocation device, resource allocation method, and resource allocation control program
US20080222218A1 (en) Risk-modulated proactive data migration for maximizing utility in storage systems
JP5827594B2 (en) Virtual machine placement apparatus and virtual machine placement method
KR101471749B1 (en) Virtual machine allcoation of cloud service for fuzzy logic driven virtual machine resource evaluation apparatus and method
JP2016042284A (en) Parallel computer system, management device, method for controlling parallel computer system, and management device control program
WO2016178316A1 (en) Computer procurement predicting device, computer procurement predicting method, and program
CN111813523A (en) Duration pre-estimation model generation method, system resource scheduling method, device, electronic equipment and storage medium
JP5515889B2 (en) Virtual machine system, automatic migration method and automatic migration program
US9563532B1 (en) Allocation of tasks in large scale computing systems
US9501321B1 (en) Weighted service requests throttling
WO2017013872A1 (en) Resource allocation system, resource allocation method, and computer-readable recording medium
WO2013070226A1 (en) Computer system performance management with control variables, performance metrics and/or desirability functions
JP2021105772A (en) Prediction management system of resource usage amount, and prediction management method of resource usage amount
US20130123950A1 (en) Computer system performance management with control variables, performance metrics and/or desirability functions
JP4999932B2 (en) Virtual computer system, virtual computer weighting setting processing method, and virtual computer weighting setting processing program
KR101718206B1 (en) Method of dynamic spectrum allocation for load balancing
JPWO2014115282A1 (en) Computer system and computer resource allocation method
CN111045815B (en) Method for optimizing deployed resources of multiple processors and extension equipment thereof
JP2017062721A (en) Resource allocation changing apparatus, resource allocation changing method, and program
Alasaly et al. Dynamic capacity planning with comprehensive formation of SLAs in clouds

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16827439

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2017529458

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16827439

Country of ref document: EP

Kind code of ref document: A1