JP5375427B2 - Temperature control device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce the heat deviation of a plurality of information processors whose power consumption is changed. <P>SOLUTION: A temperature control device controls temperatures of the plurality of information processors. A power consumption prediction function part 61 predicts a processing quantity at a predicted time from the time history information 800 of the processing quantity per unit time and predicts the power consumption of the information processor when at least part of the predicted processing quantity is processed from power consumption information 850 which changes according to the processing quantity per unit time specified for each information processor. A device specifying function part 62 specifies the information processor for processing the predicted processing quantity so that the predicted power consumption is below a defined value, and a software execution control function part 63 allocates software which makes the specified information processor generate the predicted processing quantity. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a temperature control device.

  In a data center or server room, a large number of computers are operated, so a large amount of heat is exhausted. This heat heats the surrounding air and raises the room temperature, exceeding the operable temperature defined by the computer specifications, which may cause malfunction or failure. For this reason, a cooling facility for cooling the air around the computer is arranged in the data center and the server room. For this cooling facility, general air conditioning for cooling the entire room, individual air conditioning for cooling only a part of the room or a specific rack, or the like is used.

  In addition, since a large number of servers each generating a large amount of heat are stored in a data center or server room, even within the same floor, variations in temperature distribution occur, that is, a heat bias occurs. A place where the temperature is high and there is no room for cooling is referred to as a “hot spot”, and a “cool spot” where the temperature is low and there is room for cooling. “When the air conditioning capacity is increased and the servers in all locations are cooled to below the stable operating temperature, the temperature drops not only to the hot spot but also to the cool spot, and wasteful power consumption occurs.

  On the other hand, there is a method of reducing the heat bias by grasping the temperature distribution by temperature measurement and physically rearranging the servers. However, although the thermal bias fluctuates with time, there is a disadvantage that physical relocation is not performed frequently due to high cost.

  In order to save energy for cooling a plurality of computers, a management system for managing software executed on the plurality of computers arranged in the data center has been proposed. The management system detects the temperature of the computer or the operation data of the computer, and assigns the software executed on the computer with high heat dissipation to the computer with low heat dissipation.

  In order to prevent the temperature of the computer from rising, a job scheduling apparatus that assigns software to the computer has been proposed. The job scheduling apparatus predicts the temperature of each computer when a desired job is allocated to each computer, using the average power consumption and the average processing time obtained from the detected power consumption and processing time of each computer. The job scheduling apparatus allocates jobs to a plurality of computers using the predicted computer temperature after job allocation so that the computer temperature is equal to or lower than the upper limit temperature.

Japanese Laid-Open Patent Publication No. 2007-179437 JP 2008-242614 A

For example, a central processing unit (CPU) included in an information processing apparatus has a mechanism for reducing power consumption instead of reducing processing capacity when the load is light. Therefore, the power consumption of the information processing apparatus changes according to the load.
However, since the proposed job scheduling apparatus performs computer temperature prediction using average power consumption and average processing time, the job scheduling apparatus performs computer temperature prediction in which the CPU power consumption changes according to the load. It cannot be done accurately. Further, the proposed management system assigns software to a computer with low heat dissipation after detecting the temperature of the computer with high heat dissipation. Therefore, there is a time difference from the temperature detection to the completion of the software assignment process. Within this time difference, there arises a problem that a computer exceeding the operable temperature appears, or that the heat bias exceeds the allowable range and the energy required for cooling increases.

  An object of the temperature control device according to an embodiment is to reduce a heat bias in an information processing device in which power consumption changes according to a processing amount per unit time.

  A temperature control device according to an embodiment is a temperature control device that controls the temperature of a plurality of information processing devices, and the processing amount from a time history of processing amount per unit time processed by the information processing device by execution of software And predicting the power consumption of the information processing device when processing at least a part of the predicted processing amount from the power consumption that varies depending on the processing amount per unit time specified for each information processing device. Predicted by the specified information processing device and the specified information processing device for specifying the information processing device that processes the predicted processing amount so that the predicted power consumption is less than the specified value. A processing unit that implements a software execution control function unit that allocates software that generates a large amount of processing.

  The temperature control device according to an embodiment has an effect of reducing the heat bias in the information processing device in which the power consumption changes according to the processing amount per unit time.

It is a figure which shows an example of the hardware constitutions of a temperature control apparatus. It is a figure which shows an example of the hardware constitutions of the apparatus connected to a temperature control apparatus. An example of the program and data memorize | stored in the memory | storage part of a temperature control apparatus is shown. An example of a function structure of the power consumption prediction function part implement | achieved by the temperature control apparatus, an apparatus specific function part, and a software execution control function part is shown. 2 shows an example of a program and data stored in a storage unit of an information processing device. It is an elevation view which shows an example of the rack which arrange | positions the information processing apparatus connected to a temperature control apparatus. It is a figure which shows an example of arrangement | positioning structure of the rack and cooling equipment which arrange | position the information processing apparatus connected to a temperature control apparatus. It is a figure which shows an example of the sequence which shows acquisition of the processing amount and power consumption by a temperature control apparatus. It is a figure which shows an example of the time history information of a processing amount. It is a figure which shows an example of the relationship between the processing amount registered into the time history information of processing amount, and time. It is a figure which shows an example of the power consumption information when one information processing apparatus runs one software. It is a figure which shows an example of the relationship between the amount of processing, and the power consumption information when one information processing apparatus executes one software. It is a figure which shows an example of power consumption information when one information processing apparatus runs several software. It is a figure which shows an example of the relationship between the amount of processing and power consumption information when one information processing apparatus runs several software. It is a figure which shows an example of apparatus setting information. It is a figure which shows an example of apparatus management information. It is a figure explaining an example of an apparatus specific function part and a software execution control function part. It is a figure explaining an example of an apparatus specific function part and a software execution control function part. It is a figure explaining an example of an apparatus specific function part and a software execution control function part. It is a figure which shows an example of software execution priority information. It is a figure explaining an example of an apparatus specific function part and a software execution control function part. It is a figure which shows an example of queue information. It is a figure explaining an example of an apparatus specific function part and a software execution control function part. It is a figure which shows an example of the processing flow which a temperature control apparatus performs temperature control of an information processing apparatus with the regulation value for every information processing apparatus. It is a figure which shows an example of the processing flow which a temperature control apparatus performs temperature control of information processing apparatus based on the regulation value for every rack. It is a figure which shows an example of the processing flow which a temperature control apparatus performs temperature control of the information processing apparatus which performs several software. It is a figure which shows an example of the processing flow which a temperature control apparatus performs temperature control of the information processing apparatus which stops or starts software.

Hereinafter, an embodiment of a temperature control device will be described with reference to the drawings.
FIG. 1 is a diagram illustrating an example of a hardware configuration of a temperature control device. As shown in FIG. 1, the temperature control device 10 according to an embodiment includes a storage unit 11, a processing unit 12, a communication unit 13, an external storage device 14, a drive device 15, an input unit 17, an output unit 18, and a bus 19. Have.

  The memory | storage part 11 is an apparatus which memorize | stores data electrically using a semiconductor element. The storage unit 11 is, for example, a cache memory, a main memory, or a flash memory. The cache memory is, for example, an SRAM (Static Random Access Memory). The main memory is, for example, a DRAM (Dynamic Random Access Memory). The flash memory is, for example, an EEPROM (Electrically Erasable Programmable ROM). The processing time history information 800, power consumption information 850, device setting information 900, device management information 950, software execution priority information 980, queue information 985, and program 990 stored in the storage unit 11 are shown in FIG. Will be described later.

The processing unit 12 is a device that receives data from the storage unit 11 or the input unit 17 and executes the received data by executing a program 990 stored in the storage unit 11. Then, the processing unit 12 outputs the calculated data to the storage unit 11 and the output unit 18. The processing unit 12 is, for example, a CPU.
The processing unit 12 executes the program 990 to realize a power consumption prediction function unit, a device identification function unit, and a software execution control function unit. The power consumption prediction function unit, the device specifying function unit, and the software execution control function unit will be described later with reference to FIG. 3B.

  The communication unit 13 is a device for connecting to a network cable 5 described later with reference to FIG. 2 and transmitting or receiving data. The communication unit 13 is a network adapter, for example. The external storage device 14 is, for example, a disk array of magnetic disks, a solid state drive (SSD) using flash memory, or an optical disk drive. The external storage device 14 can store programs and data stored in the storage unit 11. The drive device 15 is a device that reads and writes a storage medium 16 such as a floppy (registered trademark) disk, a CD-ROM, and a DVD. The drive device 15 includes a motor that rotates the storage medium 16, a head that reads and writes data on the storage medium 16, and the like.

  The storage medium 16 can store processing time history information 800, power consumption information 850, device setting information 900, device management information 950, software execution priority information 980, queue information 985, and a program 990. The drive device 15 reads the program 990 from the storage medium 16 set in the drive device 15. The processing unit 12 stores the program 990 read by the drive device 15 in the storage unit 11 and / or the external storage device 14. The input unit 17 is, for example, a keyboard or a mouse. The output unit 18 is, for example, a display.

  The bus 19 is connected to the storage unit 11, the processing unit 12, the communication unit 13, the external storage device 14, the drive device 15, the input unit 17, and the output unit 18, and is used for data transmission between the connected devices. It is a transmission line. The bus 19 includes an electronic circuit that functions in accordance with a standard such as PCI Express in order to transmit data.

FIG. 2 is a diagram illustrating an example of a hardware configuration of a device connected to the temperature control device.
As shown in FIG. 2, the temperature control device 10 is connected to the information processing devices 110 to 240, the power monitoring device 410, and the load distribution device 430 via the switch 400 and the network cable 5. In addition, the information processing apparatuses 110 to 240 are a large number of servers that are connected to a plurality of computers 420 via the external network 6 and placed in a data center, for example. The external network 6 is, for example, the Internet or a local area network (LAN). The computer 420 transmits processing requests for software executed by the information processing apparatuses 110 to 240 to the information processing apparatuses 110 to 240. Although one computer 420 is shown in FIG. 2, the number of computers 420 is not limited to the number shown.

  The power monitoring device 410 is a device that has a current sensor 116 and transmits the power consumption of the information processing device 110 detected by the current sensor 116 to the temperature control device 10 via the network cable 5. The current sensor 116 is a sensor for detecting the power consumption of the information processing apparatus 110. The current sensor 116 is, for example, a clamp meter. The current sensor 116 is attached to an electric cable 115 that connects the information processing apparatus 110 and the power supply 50. Similar to the information processing apparatus 110, the current sensor 246 is attached to the electric cable 245 of the information processing apparatus 240. Although not shown in FIG. 1, other information processing apparatuses 120 to 140 and 210 to 230 are similarly attached to an electric cable connected to the power supply 50, and the current sensor monitors the detected power detected. To device 410.

  In addition, without using the power monitoring device 410, the worker may register the power data actually measured with the equipment corresponding to the power monitoring device 410 in another environment in the temperature control device 10 via the input unit 17. .

  The load balancer 430 can centrally manage processing requests from the computer 420 and transfer the processing requests to a plurality of information processing apparatuses that execute software corresponding to the processing requests. The load balancer 430 can evenly distribute processing requests to the information processing apparatuses. Further, the load distribution device 430 can transfer a processing request according to the processing speed of the information processing device when there is a variation in the capability of the information processing device. In FIG. 2, the load distribution device 430 is illustrated as a device different from the information processing devices 110 to 240. However, any of the information processing apparatuses 110 to 240 or the temperature control apparatus 10 may function as a load distribution apparatus by executing software for load distribution processing.

  It should be noted that other methods for causing the plurality of information processing apparatuses 110 to 240 to execute the same application without using the load distribution apparatus 430 may be used. As such a method, for example, there is a method in which Uniform Resource Locator (URL) is separately provided in each information processing apparatus and the software to be executed is the same. In such a method, the computer 430 identifies an information processing device by a URL, but a plurality of information processing devices each identified by a separate URL may execute the same software. In such a case, a plurality of information processing apparatuses execute the same application without a load distribution apparatus. Therefore, although an embodiment using the load distribution device 430 is shown below, it is an example of a method for causing the plurality of information processing devices 110 to 240 to execute the same application using the load distribution device 430.

The information processing apparatus 110 includes a storage unit 111, a processing unit 112, a communication unit 113, an external storage device 114, and a bus 119.
The storage unit 111 is a memory, for example. The processing unit 112 is, for example, a CPU. The communication unit 113 is, for example, a network adapter. The external storage device 114 is, for example, a disk array of magnetic disks. The bus 119 is connected to the storage unit 111, the processing unit 112, the communication unit 113, and the external storage device 114, and is a transmission path for transmitting data between the connected devices. The bus 119 includes an electronic circuit that functions in accordance with a standard such as, for example, PCI Express in order to transmit data. The program and data stored in the storage unit 111 will be described later with reference to FIG.

  Since the other information processing apparatuses 120 to 240 also have the same hardware configuration as the information processing apparatus 110, the description of the hardware configuration of the other information processing apparatuses 120 to 240 is omitted. Note that the information processing devices 110 to 240 are examples, and the number of information processing devices that are temperature-controlled by the temperature control device 10 is not limited to the number shown in the information processing devices 110 to 240.

  FIG. 3A is a diagram illustrating an example of a program and data stored in the storage unit 11 of the temperature control apparatus 10. 3A stores processing time history information 800, power consumption information 850, device setting information 900, device management information 950, software execution priority information 980, queue information 985, and a program 990. The

  The processing amount time history information 800 is information indicating a time history of processing amount of software per unit time specified for each software and processed by at least one of a plurality of information processing apparatuses. The processing amount is, for example, a transaction processed by one or a plurality of information processing apparatuses according to a processing request from the computer 420 per second. A transaction is a collection of a plurality of processes related to a processing request as one processing unit. Further, the processing amount is not limited to the number of transactions per second, and may be a consumption amount of the network bandwidth per second when data transmission or data reception accompanying software execution by the information processing apparatus is performed. In a data center or server room, a plurality of information processing apparatuses arranged execute common software, and a load distribution apparatus distributes software processing requests of the computer 420 to the plurality of information processing apparatuses that execute the software. May be assigned. Accordingly, the processing amount time history information 800 is not limited to the processing amount of one information processing apparatus, but may be a time history of processing amounts processed by a plurality of information processing apparatuses.

  A processing request from the computer 420 to the information processing apparatus may have a temporal tendency. Therefore, the processing amount time history information 800 may include, for example, a processing amount of time history recorded for each day of the week or date. As will be described later with reference to FIG. 7, the temperature control device 10 generates processing amount time history information 800 by acquiring processing amount information received from each information processing device. Details of the processing time history information 800 will be described later with reference to FIGS.

  The power consumption information 850 is information indicating power consumption that is specified for each information processing apparatus and varies depending on the processing amount per unit time. The information processing apparatus includes a processing unit. Further, the CPU may execute a power consumption control function such as changing the CPU clock according to the processing amount per unit time. Thus, in order to indicate the power consumption that varies depending on the processing amount per unit time, the power consumption information 850 indicates the power consumption that occurs in each processing amount. Further, the power consumption of the information processing apparatus varies depending on the hardware specifications of the information processing apparatus. Therefore, the power consumption information 850 is specified for each information processing apparatus. Details of the power consumption information 850 will be described later with reference to FIGS.

  The apparatus setting information 900 includes identification information of a rack in which an information processing apparatus is arranged or identification information of a rack row in which racks are arranged, and information indicating a cooling heat amount for cooling the information processing apparatus. Details of the apparatus setting information 900 will be described later with reference to FIG.

  The device management information 950 is information including predicted power consumption before execution control of the information processing device and predicted power consumption after execution control of the information processing device.

  The software execution priority information 980 specifies the execution priority of software executed by the information processing apparatus. Details of the software execution priority information 980 will be described later with reference to FIG.

  The queue information 985 is a queue for storing identification information of software whose execution by the information processing apparatus is stopped when there is no information processing apparatus whose predicted power consumption is less than a specified value. Details of the queue information 985 will be described later with reference to FIG.

  The program 990 is a program for causing the temperature control device 10 to realize at least the power consumption prediction function unit, the device specifying function unit, and the software execution control function unit.

  FIG. 3B is a diagram illustrating an example of a functional configuration of a power consumption prediction function unit, a device identification function unit, and a software execution control function unit realized by the temperature control device. As illustrated in FIG. 3B, the processing unit 12 of the temperature control device 10 executes a program 990 to cause the temperature control device 10 to include a power consumption prediction function unit 61, a device identification function unit 62, and a software execution control function unit 63. Is realized.

  The processing unit 12 of the temperature control device 10 uses the power consumption prediction function unit 61 to predict the processing amount at the prediction time after the current time from the processing amount time history information 800. Then, the processing unit 12 of the temperature control apparatus 10 uses the power consumption prediction function unit 61 to predict the power consumption of the information processing apparatus when processing at least a part of the predicted processing amount from the power consumption information 850. . The processing unit 12 of the temperature control device 10 registers the predicted power consumption of the information processing device in the device management information 950.

  The processing unit 12 of the temperature control device 10 uses the device specifying function unit 62 to process the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than the specified value. Identify a processing device. In other words, the device specifying function unit 62 realized by the temperature control device 10 acquires the predicted power consumption from the device management information 950, and the predicted power consumption is less than the specified value obtained from the device setting information 900. A processing device is specified and registered in the device management information 950.

  The device specifying function unit 62 realized by the temperature control device 10 specifies an information processing device that stops execution of software when there is no information processing device whose predicted power consumption is less than a specified value. The software identified as the execution stop target is registered in the queue information 985 by the software execution control function unit 63 and the identification information of the software whose execution has been stopped is registered. In addition, the device specifying function unit 62 realized by the temperature control device 10 refers to the software priority information 980 and controls the information processing device so that the software with a high priority can be continuously executed.

  Using the software execution control function unit 63, the processing unit 12 of the temperature control device 10 refers to the device management information 950 and allocates and executes software to the identified information processing device prior to the predicted time.

  Note that examples of the power consumption prediction function unit 61, the device identification function unit 62, and the software execution control function unit 63 will be described later with reference to FIGS. 15 to 18, 20, and FIGS. 22 to 26.

  FIG. 4 is a diagram illustrating an example of programs and data stored in the storage unit 111 of the information processing apparatus 110. The storage unit 111 illustrated in FIG. 4 stores system software 670, virtualization software 680, and software 690. In FIG. 7, the storage unit 111 of the information processing apparatus 110 is shown for illustration, but other information processing apparatuses also have the same storage unit.

  The system software 670 is application software that allows the information processing apparatus 110 to realize a function of monitoring the processing amount of the information processing apparatus 110. The virtualization software 680 causes a physical information processing apparatus to function as a logical information processing apparatus using hardware virtualization technology. Hereinafter, a logical information processing apparatus is referred to as a “virtual machine”.

The software 690 is application software that realizes a specific function. The information processing apparatus 110 can execute the software 690 in the following form, for example.
(1) The information processing apparatus 110 executes an operating system (OS) on a virtual machine that functions by executing the virtualization software 680, and executes software 690 that operates in cooperation with the OS. In such a case, the virtual machine may be started, stopped, or moved with the start, stop, or movement of the software 690.
(2) The information processing apparatus 110 executes an operating system (OS) without executing the virtualization software 680 and also executes software 690 that operates in cooperation with the OS. When the information processing apparatus 110 does not execute the virtualization software 680, the storage unit 111 does not have to store the virtualization software 680.

  The information processing apparatus can execute the same or different software as other information processing apparatuses.

  Hereinafter, an example of the arrangement of the information processing apparatus and the arrangement of cooling equipment for cooling the information processing apparatus will be described with reference to FIGS. 5 and 6. FIG. 5 is a diagram illustrating an example of a rack in which information processing devices connected to the temperature control device 10 are arranged. As shown in FIG. 5, among the information processing devices 110 to 240 connected to the temperature control device 10, the information processing devices 110 to 140 are arranged in the rack 1, and the information processing devices 210 to 240 are in the rack 2. Be placed.

  FIG. 6A to FIG. 6C are diagrams respectively showing examples of the arrangement configuration of the rack and the cooling facility in which the information processing device connected to the temperature control device is arranged. In the arrangement configuration 300a shown in FIG. 6A, a plurality of racks 301a, cooling equipment 302a, a room 303a, and cold air 304a are shown. The cooling facility 302a cools the information processing apparatus in units of rooms. A plurality of racks 301a and cooling equipment 302a are arranged in the room 303a. The cool air 304a is supplied from the cooling facility 302a to the room 303a. When the cooling facility 302a cools the information processing apparatus in units of rooms, the cooling facility 302a supplies the rack 301a with cool air 304a that is not limited to ducts, air volume adjustment dampers, vents, and the like.

  In the arrangement configuration 300b illustrated in FIG. 6B, a plurality of rows 305b to 307b, a cooling facility 302b, a room 303b, and cold air 304b are illustrated. The plurality of rows 305b to 307b are a plurality of rack rows arranged in units of rows. The cool air 304b is supplied from the cooling facility 302b in units of columns. When the cooling equipment 302b cools the information processing apparatus arranged in the rack in units of rows, ducts and vents (not shown) are installed from the cooling equipment 302b to the rows 305b to 307b. The cold air 304b is supplied to the rows 305b to 307b through ducts and vents. For this reason, the arrangement configuration 300b is less likely to cause cold air diffusion than the arrangement configuration 300a in units of rooms, and the length of the ventilation path from the cooling facility 302b to the plurality of rows 305b to 307b can be shortened.

An arrangement 300c shown in FIG. 6C shows a plurality of racks 301c, cooling equipment 302c, room 303c, and cold air 304c. The cool air 304c is supplied in rack units from the cooling facility 302c.
When the cooling facility 302c cools the apparatus in units of racks, ducts and vents (not shown) are installed from the cooling facility 302c to the rack 301c. The cool air 304c is supplied to the rack 301c through a duct or a vent. Therefore, the arrangement configuration 300c can shorten the length of the ventilation path from the cooling facility 302c to the plurality of racks 301c as compared with the cooling configuration in units of rooms.

  The rooms 303a to 303c may be, for example, a data center, a server room, or a computer room. When a plurality of information processing apparatuses are arranged as shown in the arrangement configurations 300a to 300c, the temperature control device 10 connected to the plurality of information processing apparatuses is arranged in a plurality of data centers, server rooms, or computer rooms. The temperature of the information processing apparatus can be controlled.

  FIG. 7 is a diagram illustrating an example of a sequence indicating acquisition of a processing amount and power consumption by the temperature control device 10. The information processing apparatus 110 detects the processing amount at the time of software execution using the monitoring function, and transmits the processing amount information to the temperature control device 10 (S1001). The processing amount information includes the name of the information processing device that is the transmission source of the processing amount information, the name of the software executed by the transmission source information processing device, the date information of the date on which the processing information is acquired by the transmission source information processing device, And throughput. Similarly, the information processing device 240 detects the processing amount at the time of software execution using the monitoring function, and transmits the detected processing amount information to the temperature control device 10 (S1002). In this manner, the temperature control device 10 receives the processing amount information from the plurality of information processing devices, thereby generating the processing amount time history information 800 (S1003).

  The power monitoring apparatus 410 transmits the detected power information of the information processing apparatus 110 and the information processing apparatus 240 detected by the power sensor to the temperature control apparatus 10 (S1004). The detected power information includes at least the name of the information processing apparatus targeted for power consumption detection, the detection time, and the power consumption. Therefore, the temperature control apparatus 10 can confirm the power consumption of the information processing apparatus and the detection time of the power consumption using the detected power information. The temperature control device 10 associates the information processing device name and software name and time included in the processing amount information with the information processing device name and time included in the detected power information. By doing in this way, the temperature control apparatus 10 can acquire the processing amount and power consumption of software. The temperature control device 10 generates power consumption information 850 (S1005).

  Furthermore, the temperature control device 10 updates the data in the device management information 950 using the information processing device name, the software name, and the time included in the processing amount information (S1006). The data structure of the device management information 950 will be described later with reference to FIG. In this manner, the temperature control device 10 generates the processing time history information 800, the power consumption information 850, and the device management information 950 using the data included in the received processing amount information and detected power information. I can do it.

<Time history information of throughput>
FIG. 8 is a diagram illustrating an example of processing time history information. The processing amount time history information 800a illustrated in FIG. 8 is an example of the processing amount time history information 800 illustrated in FIG. 3A. The processing amount time history information 800a includes an information processing device name column 801a, a software name column 802a, a date column 803a, a day of the week column 804a, a time column 805a, and a processing amount column 806a. As will be described below, the temperature control apparatus 10 registers the received processing amount information in each column of the processing amount time history information 800a.

  In this embodiment, the time history information of the processing amount divided by month, day of the week, and time will be described, but this is only an example of the time history information of the processing amount. Unacquired data may be acquired and / or the nature of the processing amount may be estimated using an average processing amount based on the acquired date, day of the week, time, or the like, or a complement method. In this way, by performing statistical processing on unacquired data, the future processing amount can be predicted from the acquired data.

  The temperature control device 10 registers the name of the information processing device included in the processing amount information in the information processing device name column 801a. The temperature control device 10 registers the software name executed by the information processing device in the software name column 802a. The temperature control apparatus 10 registers the date of the date when the information processing apparatus acquires the processing amount in the date column 803a. The temperature control apparatus 10 registers the day of the week on which the information processing apparatus has acquired the processing amount in the day-of-week column 804a. The temperature control device 10 registers the time when the information processing device acquires the processing amount in the time column 805a. The temperature control device 10 registers the processing amount associated with the software execution processing by the information processing device in the processing amount column 806a.

  In FIG. 8, the processing amount time history information 800 a shows “2008/11/24” the time history information for one day of the processing amount of the information processing apparatus 110 that executes the software A, and “2008 / 11/28 ”and“ 2009/3/2 ”are shown with some data omitted. Although not shown in FIG. 8, the processing amount time history information 800 a includes processing amount time history information per unit time when the information processing apparatus 110 executes software other than the software A. Although not shown in FIG. 8, the processing amount time history information 800 a includes processing amount time history information per unit time when the other information processing apparatuses 120 to 240 execute software.

  “2008/11/24” is Monday, and “2008/11/28” is Friday of the same week as “2008/11/24”. “2008/11/24” and “2009/3/2” are the same day of the week, but the dates are different. As described above, the processing amount time history information 800 includes the processing amount of one day of the processing amount recorded for each day of the week or date, so that the temperature control device can be referred to by referring to the processing amount time history information 800. 10 can recognize different time histories for each day of the week or date.

  FIG. 9 is a diagram illustrating an example of the relationship between the processing amount and time registered in the processing amount time history information 800a. A line 821a illustrated in FIG. 9 indicates a time history of a processing amount of the information processing apparatus 110 executing the software A on “2009/3/2 (Monday)”. A line 822a illustrated in FIG. 9 indicates a time history of the processing amount of the information processing apparatus 110 executing the software A in “2008/11/24 (Monday)”. A line 823a illustrated in FIG. 9 indicates a time history of a processing amount of the information processing apparatus 110 executing the software A on “2008/11/28 (Friday)”.

  As shown by the lines 821a to 823a, the processing amount per unit time becomes maximum at 12:00. From this result, it can be seen that there are many processing requests for the software A at 12:00 regardless of day of the week and date. In this way, the time variation of the processing amount in time is registered in the processing amount time history information 800a.

  Further, the processing amount indicated by the line 822a is larger than the processing amount indicated by the line 823a. From this result, it can be seen that in the same week, Monday requires more use of Software A than Friday. Further, the processing amount indicated by the line 821a is larger than the processing amounts indicated by the line 822a and the line 823a. From this result, it can be seen that there are more processing requests for software A in March than in November. In this way, in the processing amount time history information 800a, the time change of the processing amount on the day of the week is registered, and further, the time change of the processing amount on the date is registered. The trend of the processing amount that changes depending on the date is registered.

  As in the above example, the temperature control apparatus 10 registers the time history of the processing amount in units of time, day of the week, date, etc., so that the time history information 800 of the processing amount includes time, day of the week, or date. The amount of processing classified for each is registered. The temperature control apparatus 10 registers the processing amount for each time, day of the week, or date in the processing amount time history information 800, so that the processing amount time history information 800 shows a temporal trend of the processing amount. It is. Therefore, the temperature control apparatus 10 uses the time history information 800 of the processing amount specified for each time or day of the week or date, as compared with the case of simply predicting using the average processing time and the average processing amount, Power consumption can be predicted with high accuracy.

<Power consumption information>
An example of the power consumption information 850 when one information processing apparatus executes one software will be described with reference to FIG. The power consumption information 850a illustrated in FIG. 10 is an example of the power consumption information 850 illustrated in FIG. 3A. In the power consumption information 850a, power consumption information when one information processing apparatus executes one software is registered. The power consumption information 850a includes an information processing device name column 851a, a software name column 852a, a processing amount column 853a, and a power consumption column 854a. As described below, the temperature control device 10 registers data in each column of the power consumption information 850a.

  The temperature control device 10 registers the name of the information processing device in the information processing device name column 851a. The temperature control device 10 registers the name of software executed by the information processing device in the software name column 852a. The temperature control device 10 registers the processing amount associated with the software execution processing by the information processing device in the processing amount column 853a. The temperature control device 10 registers the processing amount detected by the information processing device using the monitoring function in the processing amount column 853a.

  In the processing amount column 853a, the number of transactions per unit time is shown as an example of the processing amount. Registered in the power consumption column 854a is the power consumption of the information processing device detected by the power monitoring device 410 when the information processing device processes the processing amount shown in the processing amount column 853a. As described above, the processing amount detected by the information processing apparatus and the power consumption detected by the power monitoring apparatus 410 are registered in the power consumption information 850a.

  In FIG. 10, the processing amount and power consumption of the information processing device 110 and the information processing device 240 are shown, and data of other information processing devices are not shown, but the power consumption information 850 includes other information processing device data. The processing amount and power consumption are also registered.

FIG. 11 is a diagram illustrating an example of a relationship between a processing amount and power consumption information when one information processing apparatus executes one software. A solid line 871a illustrated in FIG. 11 indicates an actual measurement value of power consumption that varies depending on a processing amount when the information processing apparatus 110 illustrated in the information processing apparatus name column 851a illustrated in FIG. A broken line 872a illustrated in FIG. 11 indicates power consumption that changes at a constant rate with a change in processing amount when the information processing apparatus 110 illustrated in the information processing apparatus name column 851a illustrated in FIG. The constant ratio accompanying the change in the processing amount indicated by the broken line 872a is obtained by the average value of the processing amount increase indicated by the solid line 871a.
In addition, the point 873a of FIG. 11 is a point specified by the average power consumption which shows the average value of the detected power consumption, and the average processing time which shows the average value of the detected processing time. Since the average processing time indicates the processing time required to process a specific processing amount, the average processing time indicates the average processing amount per second at the same time. Therefore, the average processing time and the average processing power indicate the average processing amount and the average power consumption, as indicated by a point 873a in FIG. Therefore, the prediction device that performs temperature prediction using the average processing amount and the average power consumption cannot predict the power consumption of a device whose power consumption varies depending on the processing amount as indicated by the solid line 871a or the broken line 872a.

  The power consumption with respect to the processing amount indicated by the solid line 871a is nonlinear, unlike the power consumption that increases linearly at a constant ratio indicated by the broken line 872a. The reason why the solid line 871a is shown in a non-linear manner is the fluctuation of power consumption by the power consumption mechanism of the processing unit such as the power saving state of the information processing apparatus, the load fluctuation of the disk I / O, or the fluctuation of power consumption of the storage unit This is the operation of the information processing apparatus. For example, since the power saving mechanism of the processing unit works effectively at low loads, the actual power consumption is lower than the average power consumption at low loads, and the actual power consumption at high loads. Indicates a tendency to be larger than the power consumption that increases on average.

  The temperature control device 10 predicts power consumption using the power consumption and power consumption information indicated by the solid line 871a or the broken line 872a, so that the temperature control device 10 consumes power according to the change in the processing amount per unit time. Can be predicted. Further, by using the power consumption information indicated by the solid line 871a, the power consumption prediction function unit by the temperature control device 10 increases the power consumption of the information processing device even if the relationship between the power consumption and the processing amount is nonlinear. Can be predicted with accuracy.

  A solid line 874a illustrated in FIG. 11 indicates an actual measurement value of power consumption that varies depending on a processing amount when the information processing apparatus 240 illustrated in the information processing apparatus name column 851a illustrated in FIG. A broken line 875a illustrated in FIG. 11 indicates power consumption that changes at a constant rate with a change due to a processing amount when the information processing apparatus 240 illustrated in the information processing apparatus name column 851a illustrated in FIG. A broken line 875a is obtained by averaging measured values of power consumption indicated by a solid line 874a.

  As indicated by a solid line 871a, is the power consumption per processing amount when the information processing apparatus 110 executes the software A larger than the power consumption per processing amount when the information processing apparatus 240 executes the software A? I understand that. As described above, the information processing amount per transaction varies depending on the information processing apparatus. Therefore, by using the power consumption information detected for each information processing device, the temperature control device 10 can accurately predict the power consumption.

  An example of the power consumption information 850 when one information processing apparatus executes a plurality of software will be described with reference to FIG. The power consumption information 850b illustrated in FIG. 12 is an example of the power consumption information 850 illustrated in FIG. 3A. The power consumption information 850b includes power consumption information when one information processing apparatus executes two pieces of software. The power consumption information 850b includes an information processing device name column 851b, software name areas 881a and 881b, first software processing amount areas 882a and 882b, and second software processing amount areas 883a and 883b. The power consumption information 850b further includes power consumption data areas 884a and 884b.

  The temperature control device 10 registers data in each column and area of the power consumption information 850b as described below. The temperature control device 10 registers the name of the information processing device in the information processing device name column 851b. The temperature control apparatus 10 registers the software name executed by the information processing apparatus in the software name areas 881a and 881b. In the power consumption information 850b shown in FIG. 12, the processing amount and power consumption when two pieces of software are executed are registered. Therefore, the temperature control apparatus 10 stores the two software names in the software name areas 881a and 881b. Register each. In the example shown in FIG. 12, the software name “A” and the software name “B” are registered in the software name areas 881a and 881b.

  The temperature control device 10 receives the processing amount by the pre-control software execution from the information processing device 110 or 240, and registers the received processing amount in the processing amount area 882a or 882b, respectively.

  The temperature control device 10 receives the processing amount due to the execution of the second software from the information processing device 110 or 240, and registers the received processing amount in the processing amount region 883a or 883b, respectively.

  The temperature control device 10 registers the power consumption detected by the power monitoring device 410 during processing of the first software and / or the second software in the information processing device at the time of power consumption detection in the power consumption data areas 884a and 884b.

  As described above, the temperature control device 10 registers the processing amount actually measured by the information processing device and the power consumption actually measured by the power monitoring device 410 in the power consumption information 850b. In FIG. 12, the processing amount and power consumption of the information processing devices 110 and 240 are shown for explanation, and data of other information processing devices is not shown, but the power consumption information 850b includes other information processing devices. The processing amount and power consumption are also registered.

An example of the relationship between the processing amount and power consumption information when one information processing apparatus executes a plurality of software will be described with reference to FIG.
The X axis shown in FIG. 13 indicates the amount of processing when the information processing apparatus 110 executes the software A in the power consumption information 850b shown in FIG. The Y axis shown in FIG. 13 indicates the processing amount when the information processing apparatus 110 executes the software B in the power consumption information 850b shown in FIG. The Z axis shown in FIG. 13 indicates power consumption when the information processing apparatus 110 executes the software A and / or software B in the power consumption information 850b shown in FIG.

  Similarly to the case shown by the solid line 871a, the solid line 874a, the broken line 872a, and the broken line 875a in FIG. 11, the power consumption shown in FIG. However, in the example shown in FIG. 13, since the processing amount and power consumption when two pieces of software are executed are shown, even when the processing amount of one software is constant, the processing amount of other software changes, It shows that the power consumption per unit time changes.

  When the information processing device is executing a virtual machine, an example in which one information processing device executes two softwares is one in which one information processing device executes two virtual machines that execute one software. Corresponding to the case.

  In the job scheduling apparatus that predicts the temperature of the information processing apparatus using the average processing amount and the average power consumption by each software, the change in the processing amount per unit time when executing a plurality of software is not taken into consideration at all. However, in the power consumption information 850b shown in FIGS. 12 and 13, changes in power consumption due to changes in the processing amount per unit time of a plurality of software are registered. Therefore, by using the power consumption information 850b, the temperature control apparatus 10 can predict the power consumption of the information processing apparatus when executing a plurality of softwares with high accuracy.

<Device setting information>
An example of the device setting information will be described with reference to FIG.
Device setting information 900a illustrated in FIG. 14 is an example of the device setting information 900 illustrated in FIG. 3A. The device setting information 900a includes an information processing device name column 901a, a column number column 902a, a rack number column 903a, and a cooling heat quantity column 904a.

  The temperature control device 10 registers the name of the information processing device in the information processing device name column 901a. The temperature control device 10 registers a number for identifying the column in which the rack is arranged in the column number column 902a. The temperature control apparatus 10 registers a number for identifying the rack in which the information processing apparatus is arranged in the rack number column 903a. Note that the user of the temperature control device 10 may input data to the information processing device name column 901a, the column number column 902a, and the rack number column 903a via the input unit 17.

  The temperature control device 10 may receive information indicating the cooling heat amount registered in the cooling heat amount column 904a, and calculate the cooling heat amount according to the arrangement configuration of the rack and the cooling facility. The temperature control device 10 registers the amount of cooling heat supplied per information processing device calculated by the temperature control device 10 in the cooling heat amount column 904a. The user of the temperature control device 10 may register the cooling heat quantity value of the cooling device in the temperature control device 10 via the input unit 17.

  For example, when cooling is performed in units of rooms as in the arrangement configuration 300a in FIG. 6A, the temperature control device 10 uses, for example, a finite volume method to calculate the cool air 304a flowing through the room and the cool air flowing through each information processing device. You may calculate by simulation. Further, the temperature control device 10 may calculate the heat of cooling of the information processing device without performing cold air simulation. For example, the temperature control device 10 calculates the cooling heat amount for each rack by dividing the cooling heat amount of the cooling facility 302a by the number of racks. Next, the amount of cooling heat for each information processing device may be calculated by dividing the calculated amount of cooling heat for each rack by the number of information processing devices arranged in the rack.

  For example, as shown in the arrangement configuration 300b of FIG. 6B, when the cooling is performed in units of rows, the temperature control device 10 passes the amount of cooling heat for cooling the information processing devices by the cooling equipment 302b through the rack row. It may be calculated by simulating the cool air 304b. Further, the temperature control device 10 may obtain the amount of cooling heat of the information processing device without performing cold air simulation. For example, the temperature control apparatus 10 may calculate the cooling heat amount for each column by dividing the cooling heat amount of the cooling facility 302b by the number of columns. Next, the temperature control device 10 calculates the cooling heat amount for each rack by dividing the calculated cooling heat amount for each column by the number of racks. Then, the temperature control device 10 may calculate the cooling heat amount of each information processing device by dividing the calculated cooling heat amount for each rack by the number of information processing devices arranged in the rack.

For example, when cooling is performed in units of racks as in the arrangement configuration 300c of FIG. 6C, the temperature control device 10 passes the amount of cooling heat for cooling the information processing devices by the cooling facility 302c through the racks. You may calculate by simulating the cold 304c. Further, the temperature control device 10 may calculate the cooling heat amount of each information processing device by dividing the cooling heat amount of the rack by the number of information processing devices arranged in the rack without performing the cold air simulation. .
In this way, the temperature control device 10 can calculate the cooling heat amount registered in the cooling heat amount column 904a according to the arrangement configuration of the rack and the cooling equipment.

<Device management information>
An example of the device management information will be described with reference to FIG. Device management information 950a illustrated in FIG. 15 is an example of the device management information 950 illustrated in FIG. 3A. The temperature control device 10 registers data in each data area shown in the device management information 950a. The device management information 950a includes an information processing device name column 951a, a rack number column 952a, a pre-control predicted power consumption column 953a, a post-control predicted power consumption column 954a, and a pre-control predicted power consumption column 955a for each rack. The device management information 950a further includes a predicted post-control power consumption column 956a for each rack, a software name column 957a before execution control, and a software name column 958a after execution control.

  The temperature control device 10 registers the name of the information processing device in the information processing device name column 951a. The temperature control apparatus 10 registers a number for identifying the rack in which the information processing apparatus is arranged in the rack number column 952a. The user of the temperature control device 10 may register the numbers in the information processing device name column 951 a and the rack number column 952 a via the input unit 17.

  The temperature control device 10 registers the predicted power of each information processing device calculated by the power consumption prediction function unit in the pre-control predicted power consumption column 953a. The temperature control device 10 registers the predicted power consumption after the software allocation is performed by the software execution control function unit in the post-control predicted power consumption column 954a. The temperature control device 10 registers the total value of the predicted power consumption registered in the pre-control predicted power consumption column 953a of the information processing device arranged in the rack in the pre-control predicted power consumption column 955a for each rack. The temperature control device 10 registers the total value of the predicted power consumption registered in the post-control predicted power consumption column 954a of the information processing device arranged in the rack in the post-control predicted power consumption column 956a for each rack.

  The temperature control device 10 registers the information processing device before the execution control of the software in the software name column 957a before the execution control. The temperature control device 10 registers the name of software executed by the information processing device after execution control of software in the software name column 958a after execution control.

<Power consumption prediction function part>
An example of the power consumption prediction function unit realized by the processing unit 12 will be described with reference to FIG. A point 826 on the line 822a indicates the processing amount of “9:00” of “2008/11/24 (Monday)”. The time indicated by the point 826 is defined as the current time for explanation. The time one hour after the current time is the time “10:00” indicated by a point 827 on the line 822a. The processing amount indicated by the point 827 is “1418 (transaction / second)”. Therefore, the power consumption prediction function unit sets “10 o'clock” that is one hour after “9 o'clock” of the current time “2008/11/24 (Monday)” as the prediction time for predicting power consumption. Then, the power consumption prediction function unit acquires “1418 (transaction / second)” as the processing amount at the predicted time from the processing amount time history information 800. Thus, the power consumption prediction function unit predicts the processing amount of the information processing apparatus from the processing amount time history information 800.

A time of several seconds to several tens of seconds is generated from the start to the end of the software execution control. Therefore, if the time between the current time and the predicted time is extremely short, the software execution control may end after the predicted time has elapsed. Therefore, it is preferable that the predicted time is set to be the time after the end of software execution control.
In the above example, the time interval between the current time and the predicted time is 1 hour. The time of 1 hour is sufficiently longer than the time of several seconds to several tens of seconds, which is the time for stopping or executing the software. Therefore, in the above example, the software execution control can be terminated before the predicted time.

  Further, software execution control may be performed in accordance with the time interval between the current time and the predicted time. For example, as in the above example, if the time interval between the current time and the predicted time is 1 hour, software execution control may be performed once per hour. Since software execution control involves stopping and starting software, there is a possibility that processing requests from an external computer to an information processing apparatus for which software execution control is performed are interrupted in the short term. As in the above example, when the temperature control device 10 performs software execution control once an hour, the interruption time for the processing request is relatively shorter than the time of 1 hour when the software execution control is not performed.

  Next, the power consumption prediction function unit of the temperature control apparatus 10 predicts the power consumption when the information processing apparatus 110 executes the software A with a predetermined processing amount from the power consumption information 850. As described above, the processing amount of the software A is predicted to be “1418 (transaction / second)”. For example, when there are three information processing apparatuses that execute the software A, the amount of processing burdened by the information processing apparatus 110 due to load distribution by the load distribution apparatus is, for example, 1/3 of “1418 (transaction / second)”. It becomes a certain “473 (transaction / second)”. Therefore, the power consumption prediction function unit obtains the power consumption when the information processing apparatus executes with the processing amount “473 (transaction / second)”. Referring to FIG. 11, the power consumption of the information processing apparatus 110 when the processing amount of the software A is 473 (transactions / second) is “450 W”. Therefore, the temperature control apparatus 10 predicts that the power consumption at the predicted time “10 o'clock” is “450 W” by the power consumption prediction function unit.

  As described above, the temperature control device 10 uses the power consumption prediction function unit to predict the processing amount from the processing amount time history information 800 and to process at least a part of the predicted processing amount. The power consumption of the processing device can be predicted from the power consumption information 850.

<Device identification function unit and software execution control function unit>
Examples of the device specifying function unit and the software execution control function unit of the temperature control device 10 will be described using cases 1 to 10 shown below.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 1)]
In Case 1, the temperature control device 10 specifies the information processing device that processes the predicted processing amount and specifies the software so that the predicted power consumption of the information processing device that executes the software is less than the specified value. Assigned to the information processing apparatus. The temperature control device 10 specifies an information processing device that processes the predicted processing amount by using the relationship shown in the following equation 1 by the device specifying function.
Ps> EP1 = CHs × C1 (Formula 1)
Ps: Predicted power consumption for each information processing device EP1: Specified power value CHs: Standard cooling heat amount assigned for each information processing device In the air conditioning management of the data center, it is a cooling heat amount assigned during normal operation, and is a cooling heat amount that is smaller than the design cooling capacity CHsMax of the cooling facility. The standard cooling heat quantity is registered in the cooling heat quantity column 904 shown in FIG.
C1: A predetermined coefficient, for example, 1.1

  The cooling facility can satisfy the design cooling capacity CHsMax for a certain period of time. C1 shown in Formula 1 is set to be less than CHsMax. For example, when CHsMax is defined as when the information processing apparatus maintains power consumption of 1.2 times the amount of cooling heat for 1 minute, C1 is defined as a value smaller than 1.2. In addition, the information processing apparatus may have a function of automatically reducing the processing speed of the CPU when a certain high temperature is reached. C1 may be set as a value such that the CPU processing speed is not reduced by the information processing apparatus.

  The temperature control device 10 causes the identified information processing device to execute the software before the predicted time by the software execution control function unit. Such a software execution control function unit can be realized, for example, by the temperature control device 10 transmitting a start command to the specified information processing device. For example, the temperature control apparatus 10 can cause the information processing apparatus to newly execute the software by transmitting an activation command for starting the software to the information processing apparatus that newly executes the software. Further, when the temperature control device 10 wants to reduce or increase the amount of processing by software processing to the information processing device, the temperature control device 10 instructs the load distribution device 430 to reduce or increase the number of processing requests to the target information processing device. Send. By doing in this way, the temperature control apparatus 10 can newly allocate software to an information processing apparatus, or can allocate the processing amount by software execution to an information processing apparatus.

  An example of the device specifying function unit will be described with reference to FIG. In the example illustrated in FIG. 15, the software execution control function unit defines 1.1 times the amount of cooling heat registered in the cooling heat amount column 904a of the device setting information 900a as the specified power value (EP1) expressed by Equation 1. Since the amount of cooling heat is “400 W” from 904a, the specified power value (EP1) is 440 W from Equation 1. The predicted power consumption before execution control shown in the data area 962a of the device management information 950a is 450W. The 450 W indicated in the data area 962a is larger than the specified power value of 440 W.

  Therefore, the temperature control device 10 uses the device specifying function unit to specify the information processing device that processes the predicted processing amount of the software A so that the predicted power consumption of the information processing device 110 is less than the specified value. . For example, the temperature control apparatus 10 assigns the software A executed by the information processing apparatus 110 to the information processing apparatus 140 as illustrated in the data area 965a, as illustrated in the data area 967a. In addition, the temperature control device 10 reduces a part of the processing amount of the information processing device 110 as shown in the data areas 963a and 964a so that the predicted power consumption of the information processing device 110 is less than the specified value. Assign software A to Therefore, the predicted power consumption of the information processing apparatuses 110 and 140 is “200 W” which is less than “440 W” which is the specified power value (EP1).

  As described above, the temperature control device 10 uses the device specifying function unit to specify the information processing device 140 that executes the software A so that the predicted power consumption of the information processing device that executes the software A is less than the specified value. To do. Then, the temperature control device 10 causes the information processing device 140 to execute the software A executed by the information processing device 110 using the software execution control function unit as shown in the data area 965a.

The power consumption of the information processing apparatus including the CPU changes according to the processing amount. However, in the prior art, jobs are assigned to computers using the average power consumption and average processing amount of the computers.
The temperature control device 10 does not perform temperature control using the average power consumption and average processing amount of the information processing device, but the predicted power consumption acquired using the power consumption information that changes according to the processing amount is less than the specified value. The execution control of the software of the information processing apparatus is performed so that Therefore, the temperature control device 10 can reduce the heat bias in a plurality of information processing devices whose power consumption changes according to the processing amount per unit time.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 2)]
In case 2, the temperature control device 10 determines the specified value shown in case 1 from the total value of the predicted power consumption of the information processing devices arranged in the region. Then, similarly to Case 1, the temperature control device 10 specifies the information processing device that processes the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than the specified value, and Assign the software to the specified information processing device.

  Examples of the area where the information processing apparatus is arranged include a rack and a row. When the object for supplying cold air by the cooling facility is not an information processing apparatus but a facility such as a rack or a row, it may be preferable to reduce the heat bias for each region. For example, when the cool air is supplied to each rack as shown in the arrangement configuration 300c of FIG. 6C, the same amount of cooling heat is supplied to each rack, so it is preferable that the heat generated from each rack is not biased. . In the example shown below, a rack is used as an example of an area where the information processing apparatus is arranged.

In Case 2, the rack in which the total value of the predicted power consumption of the information processing devices arranged in the rack is equal to or greater than the specified value is specified using the relationship shown in Equation 2 below.
Rn (ab)> EP2 = CHtrack × C (Equation 2)
Rn (ab): Total value of predicted power consumption of information processing devices arranged in the rack. n is a rack number, b is an information processing device arranged in the rack, and a is a name of software executed by the information processing device.
EP2: Specified power value for each rack CHack: Standard cooling heat amount allocated for each rack C2: Predetermined coefficient, for example, 1.05
R1 (ab) can be defined as R1 (A110, B120, A130, C140). R1 indicates a rack whose identification number is “1”. A110 indicates the information processing apparatus 110 that executes the software A. B120 indicates the information processing apparatus 120 that executes the software B. A 130 indicates the information processing apparatus 130 that executes the software A. C140 indicates the information processing apparatus 140 that executes the software C.

  The temperature control device 10 calculates Rn (ab) shown in Equation 2 using the predicted power registered in the pre-control predicted power consumption sequence 953a shown in FIG. 15, and stores it in the pre-control predicted power consumption sequence 955a for each rack. sign up. When there is a rack that exceeds the specified power value (EP2) for each rack in the pre-control predicted power consumption column 955a, the temperature control device 10 determines that there is a target device that makes the predicted power consumption less than the specified value in the rack. .

The temperature control device 10 arbitrarily identifies an information processing device being executed among information processing devices arranged in racks that exceed the specified power value (EP2) for each rack. At this time, the specified value for reducing the power consumption of the arbitrarily specified information processing apparatus may be specified as follows.
EP1 ′ = EP2− (Rn (ab) −Ps ′) (Formula 3)
EP1 ′: Specified power value Ps ′: Predicted power consumption of an information processing apparatus that is arbitrarily specified

  The temperature control device 10 uses the device specifying function unit to set the predicted processing amount so that the information processing device that is arbitrarily specified is less than the specified power value (EP1 ′) shown in Equation 3. An information processing apparatus to be processed is specified.

For example, when R1 (A110, B120, A130, C140) is “1700 W” and is larger than the specified power value “1680 W” determined in the area, the temperature control device 10 is the information processing device arranged in the rack “1”. It is determined that there is a software execution control target. The power consumption of the information processing apparatus 110 is “435 W”, the power consumption of the information processing apparatus 120 is “425 W”, the power consumption of the information processing apparatus 130 is “430 W”, and the power consumption of the information processing apparatus 140 is “410 W”. Suppose there is. The temperature control device 10 arbitrarily specifies the information processing device 140. In that case, EP1 ′ shown in Equation 3 is calculated as follows.
EP1 '= 1680W- (1700W-410W) = 390W
The temperature control device 10 specifies the information processing device that processes the predicted processing amount so that the power consumption of the information processing device 140 is less than “390 W” that is EP1 ′.

In the above example, when the temperature control device 10 arbitrarily specifies the information processing device 110 with the highest power consumption, EP1 ′ shown in Equation 3 is calculated as follows.
EP1 '= 1680W- (1700W-435W) = 415W
The temperature control device 10 specifies the information processing device that processes the predicted processing amount so that the power consumption of the information processing device 110 is less than “415 W” that is EP1 ′.

  The temperature control device 10 causes the identified information processing device to execute the software before the predicted time by the above-described software execution control function unit. In this way, the temperature control device 10 can reduce the heat bias between the racks in which the information processing devices whose power consumption changes with the processing amount per unit time are arranged.

  With reference to FIG. 15, a case will be described in which an information processing apparatus that processes a predicted processing amount is specified based on a total value of predicted power consumption of information processing apparatuses arranged in a region. As shown in the rack number column 952a in FIG. 15, the number of information processing apparatuses arranged in the rack is four. Therefore, the temperature control device 10 determines that “1600 W”, which is four times the amount of cooling heat registered in 904a of the device setting information 900a, is the cooling heat of the rack. Therefore, the specified power value (EP2) for each rack is “1680 W” from Equation 2.

  The predicted power consumption before execution control shown in the data area 971a of the device management information 950a is “1705 W”, which is larger than the specified power value (EP2) for each rack. Therefore, the temperature control device 10 determines that the total value of the predicted power consumption of the information processing devices arranged in the rack with the number “2” is equal to or greater than the specified value using the device specifying function unit. Then, the temperature control device 10 predicts the information processing device 230 that consumes “435 W” having the highest predicted power consumption among the information processing devices arranged in the rack with the number “2” indicated in the data area 972a. It is determined that the target device has power consumption less than the specified value.

  Then, as shown in Equation 3, since EP1 ′ = 1680W− (1705W−435W) = 410W, the temperature control device 10 reduces the power consumption of the information processing device 230 that is the target device to less than 410W. An information processing apparatus that processes the predicted processing amount is specified.

  Then, as illustrated in the data area 976a, when the temperature control apparatus 10 causes the information processing apparatus 610 to execute the software B, as illustrated in the data area 977a, the power consumption “240W” of the information processing apparatus 230 that is the target apparatus. Is smaller than the specified power value (EP1 ′). Further, as shown in the data area 973a, the predicted power “1560 W” for each rack after the execution control of the rack with the number “2” is smaller than the specified power value (EP2) for each rack. Furthermore, as shown in the data area 974a, the total power value “1420W” of the information processing devices arranged in the rack “6” is smaller than the specified power value (EP2) for each rack.

  Therefore, the temperature control device 10 identifies the information processing device 610 after execution control among the racks with the number “6” as an information processing device that newly processes at least a part of the processing amount of the software B. Then, the temperature control apparatus 10 determines that the software B shown in the data area 975a executed by the information processing apparatus 230 that has been determined to be the target apparatus whose predicted power consumption is less than the specified value, as shown in the data area 976a. To run.

In the example of execution control for each information processing apparatus shown in FIG. 15, the temperature control apparatus 10 indicates that the predicted power “1240 W” for each rack before execution control shown in the data area 961 a is the specified power value (EP2 for each rack). ) Since it is smaller than “1680 W”, execution control for each rack is not performed. Further, in the example of execution control for each rack illustrated in FIG. 15, the temperature control apparatus 10 indicates that the predicted power consumption “435 W” before execution control indicated in the data area 972 a is smaller than the specified power value (EP1) “440 W”. Therefore, execution control for each information processing apparatus is not performed.
Thus, by combining the device specifying function unit shown in case 1 and the device specifying function unit shown in case 2, the heat bias for each information processing device and each rack can be reduced.

  In a data center or a server room, information processing apparatuses are not scattered in a room, but a plurality of information processing apparatuses are often arranged in a region such as a rack. Therefore, the cooling equipment of the information processing apparatus is designed to supply a certain amount of cool air to the area as described with reference to FIG. Therefore, in a data center or server room where cold air is supplied for each area, it is preferable that the power consumption for each area be constant. As described above, the temperature control device 10 performs temperature management for each region where the information processing device is arranged, thereby reducing the heat bias for each region where the information processing device whose power consumption changes according to the processing amount is arranged. It can be made smaller.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 3)]
In case 3, as in case 1, the temperature control device 10 specifies an information processing device that processes the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than a specified value. . Further, in case 3, the temperature control device 10 stops the software executed by the target device that makes the predicted power consumption less than the specified value. That is, the difference between Case 3 and Case 1 is that the temperature control device 10 performs a process of stopping the execution of software for the target device that makes the predicted power consumption less than the specified value.

  As a method of stopping the software, there is a method in which the temperature control device 10 transmits a stop command to the information processing device determined as the target device. In addition, when the information processing apparatus has activated a virtual machine, there is a method of moving the virtual machine from the information processing apparatus determined to be the target apparatus to an apparatus that newly starts executing software.

  An example of the device specifying function unit and the software execution control function unit shown in case 3 and case 4 described later will be described with reference to FIG. Device management information 950b shown in FIG. 16 is an example of device management information 950 shown in FIG. 3A. The device management information 950b includes an information processing device name column 951b, a rack number column 952b, a pre-control predicted power consumption column 953b, a post-control predicted power consumption column 954b, a pre-control predicted power consumption column 955b for each rack, and a post-control for each rack. A predicted power consumption column 956b is included. Furthermore, the device management information 950b has a software name column 957b before execution control and a software name column 958b after execution control. The information processing device name column 951b to the post-execution software name column 958b correspond to the information processing device name column 951a to the post-execution software name column 958a shown in FIG.

  The difference between Case 3 and Case 1 is that the temperature control device 10 performs a process of stopping the execution of software for the target device whose predicted power consumption is less than the specified value, and therefore the difference will be described.

  As described in Case 1, the temperature control apparatus 10 determines that the information processing apparatus 110 that executes the software A is a target apparatus that makes the predicted power consumption less than a specified value. Then, the temperature control apparatus 10 identifies the information processing apparatus 140 as an information processing apparatus that newly processes the predicted processing amount.

  Accordingly, the temperature control device 10 is the information specified as the information processing device that newly processes the predicted processing amount of the software A shown in the data area 964b executed by the specified information processing device 110 as shown in the data region 965b. The processing device 140 is executed. And the temperature control apparatus 10 stops execution of the software A by the information processing apparatus 110, as shown to the data area 966b.

  As shown in the data area 963b, the predicted power consumption “395W” of the information processing apparatus 140 after execution control is lower than the specified power value (EP1). Therefore, the temperature control device 10 stops the software executed by the target device that makes the predicted power consumption less than the specified value, and causes another information processing device to newly execute the software, thereby causing a heat bias between the information processing devices. Can be reduced.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 4)]
In Case 4, as in Case 2, the temperature control device 10 determines the specified value shown in Case 1 from the total predicted power consumption of the information processing devices arranged in the region. Similar to Case 2, the temperature control device 10 identifies an information processing device that processes the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than a specified value. Further, in case 4, the temperature control device 10 stops the software executed by the target device that makes the predicted power consumption less than the specified value. That is, the difference between Case 4 and Case 2 is that the temperature control device 10 performs a process of stopping execution of software for the target device that makes the predicted power consumption less than the specified value. Therefore, the differences will be described below.

  As described in Case 2, the target device whose predicted power consumption is less than the specified value is determined to be the information processing device 230. The temperature control apparatus 10 causes the information processing apparatus 610, which is an information processing apparatus that newly processes the predicted processing amount as illustrated in the data area 976b, to execute the software B illustrated in the data area 975b executed by the information processing apparatus 230. Then, the temperature control device 10 stops the execution of the software B by the information processing device 230 as shown in the data area 977b.

  As shown in the data area 978b, the predicted power consumption of the information processing apparatus 230 after execution control is lower than the specified power value (EP1 '). As shown in the data area 973b, the predicted power consumption “1580 W” for each rack after execution control is lower than the specified power value (EP2) for each rack. Therefore, by stopping the software executed by the target device that makes the predicted power consumption less than the specified value and causing the other information processing device to newly execute the software, the heat bias between the information processing devices can be reduced. .

[Device Specific Function Unit and Software Execution Control Function Unit (Case 5)]
In Case 5, the temperature control device 10 regards a plurality of software executed by one information processing device as one set of software. In Case 5, as in Case 3, the temperature control device 10 identifies an information processing device that processes the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than a specified value. . And the temperature control apparatus 10 stops the software performed with the object apparatus which makes prediction power consumption less than a regulation value. In other words, the difference between Case 5 and Case 3 is that one information processing apparatus executes one software and one information processing apparatus regards a plurality of software as one set of software. is there. The case where one information processing apparatus executes a plurality of software includes a case where one information processing apparatus executes a plurality of virtual machines that execute one software.

  As described with reference to FIGS. 12 and 13, the power consumption information 850 when each information processing apparatus executes a plurality of software is stored in the storage unit 11. Therefore, the temperature control apparatus 10 can predict the power consumption of each information processing apparatus when executing a plurality of software using the power consumption information 850. In the example shown below, the temperature control device 10 regards a plurality of software as one set of software, and implements the device specifying function unit and the software execution control function unit. Hereinafter, differences between the case 3 and the case 5 will be described.

  An example of an apparatus specifying function unit and a software execution control function unit when the information processing apparatus executes a plurality of software as one set of software will be described with reference to FIG. The device management information 950c illustrated in FIG. 17 is an example of the device management information 950 illustrated in FIG. The device management information 950c includes an information processing device name column 951c, a rack number column 952c, a pre-control predicted power consumption column 953c, a post-control predicted power consumption column 954c, a pre-control predicted power consumption column 955c for each rack, and a post-control for each rack. A predicted power consumption column 956c is included. Furthermore, the device management information 950c has a software name column 957c before execution control and a software name column 958c after execution control. The information processing device name column 951c to the software name column 958c after execution control correspond to the information processing device name column 951a to the software name column 958a after execution control shown in FIG.

  The data area 964c indicates that the information processing apparatus 110 executes two pieces of software A and B. Since the predicted power “450 W” before execution control of the information processing device 110 shown in the data area 962 c is larger than the specified value “440 W”, the temperature control device 10 is a target for which the information processing device 110 makes the predicted power consumption less than the specified value. Judged to be a device. In the temperature control apparatus 10, when the information processing apparatus 140 executes the software A and B as shown in the data area 965c, the predicted power “395W” of the information processing apparatus 140 shown in the data area 963c becomes less than the specified value. Therefore, the temperature control apparatus 10 determines that the information processing apparatus 140 is an information processing apparatus that newly processes at least a part of the processing amount. Therefore, the temperature control device 10 stops the plurality of software regarded as one set executed by the information processing device 230 and causes the information processing device 610 to newly execute the plurality of software regarded as one set, thereby performing information processing. It is possible to reduce the deviation of heat between devices.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 6)]
The temperature control device 10 regards the plurality of software as one set of software and performs the same processing as in the case 4.

  A data area 975c illustrated in FIG. 17 indicates that the information processing apparatus 230 executes two pieces of software B and D. Since the predicted power “1705W” before execution control shown in the data area 971c is larger than the specified value (EP2) “1680W” for each rack, the temperature control apparatus 10 sets the predicted power consumption in the rack of the number “2” below the specified value. It is determined that the target device is to be placed. Then, the temperature control device 10 determines that the information processing device 230 having the highest power consumption among the information processing devices arranged in the rack “2” is a target device whose predicted power consumption is less than the specified value. From Equation 3, EP1 ′ = 1680W− (1705W−435W) = 410W. Therefore, the temperature control apparatus 10 performs information processing that processes at least a part of the predicted processing amounts of the software B and D so that the power consumption “435 W” of the information processing apparatus 230 indicated in the data area 972 c is less than 410 W. Identify the device. Then, as shown in the data area 974c, when the information processing apparatus 610 executes the software B and D, as shown in the data area 977c, the predicted predicted power consumption is less than the specified power value (EP1 ′). . As shown in the data area 973c, the predicted power consumption “1580 W” for each rack is less than the specified power value (EP2). Then, as shown in the data area 976c, the temperature control apparatus 10 causes the information processing apparatus 610 to execute the software B and the software D that have been executed by the information processing apparatus 230 as one set.

  Thus, even when the information processing device is executing a plurality of software, the temperature control device 10 regards the plurality of software as one software and controls the execution of the software. Therefore, the temperature control device 10 can easily realize a reduction in heat bias in a plurality of information processing devices whose power consumption changes according to the processing amount.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 7)]
The temperature control device 10 identifies an information processing device that executes software with low priority so that the predicted power consumption of a device that executes a plurality of software is less than a specified value. Furthermore, the temperature control device 10 stops the execution of the low priority software executed in the target device that makes the predicted power consumption less than the specified value.

An example of a device specifying function unit and a software execution control function unit when another information processing device executes one of a plurality of pieces of software executed by the information processing device will be described with reference to FIG.
The device management information 950d shown in FIG. 18 is an example of the device management information 950 shown in FIG. The device management information 950d includes an information processing device name column 951d, a rack number column 952d, a pre-control predicted power consumption column 955d for each rack, a post-control predicted power consumption column 956d for each rack, and a pre-control predicted power consumption column 953d. . Further, the device management information 950d includes a post-control predicted power consumption column 954d, a software name column 957d before execution control, and a software name column 958d after execution control. The information processing device name column 951d to the software name column 958d after execution control correspond to the information processing device name column 951a to the software name column 958a after execution control shown in FIG.

  As shown in the data area 961d, the predicted power consumption before the execution control of the information processing apparatus 110 that executes the two softwares A and D is “450 W” as shown in the data area 962d. Greater than (EP1). Therefore, the software execution control function unit determines that the information processing apparatus 110 is a target apparatus that makes the predicted power consumption less than a specified value.

  FIG. 19 is a diagram illustrating an example of software execution priority information. The software execution priority information 980a illustrated in FIG. 19 is an example of the software execution priority information 980 illustrated in FIG. 3A. The temperature control apparatus 10 can confirm that the execution priority of the software A is higher than the execution priority of the software D with reference to the software execution priority information 980a. Therefore, the temperature control apparatus 10 controls the information processing apparatus 110 to stop the execution of the software D and continue the execution of the software A. Therefore, the temperature control device 10 uses the device specifying function unit to process the predicted processing amount of the software D so that the predicted power consumption of the information processing device 110 that executes the software A is less than a specified value. Identify a processing device.

  As shown in the data area 966d, when the software D is assigned to the information processing apparatus 140, the information processing apparatus 110 can continuously execute the software A as shown in the data area 965d. Further, as shown in the data area 967d, the power consumption after the control of the information processing apparatus 110 is “200 W”. Further, as shown in the data area 963d, the predicted power consumption after the execution control of the information processing apparatus 140 is the specified value “200W”. Therefore, the power consumption of the information processing apparatuses 110 and 140 is less than the specified power value (EP1).

  As described above, when the target device whose predicted power consumption is less than the specified value is executing a plurality of software, the temperature control device 10 causes another information processing device to execute the software having a low execution priority.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 8)]
The temperature control device 10 determines the specified value indicated by the case 7 from the total value of the predicted power consumption of the information processing devices arranged in the region, and performs the same processing as the case 7.

  As shown in the data area 969d of FIG. 18, the total power consumption “1705W” of the information processing devices arranged in the rack with the identification number “2” is larger than the specified power value (EP2) “1680W”. As shown in the data area 970d, the temperature control device 10 sets the information processing device 230 having the largest predicted power among the information processing devices arranged in the rack with the number “2” to the predicted power consumption below the specified value. It is determined that it is a target device. From Equation 3, EP1 '= 1680W- (1705W-435W) = 410W. The temperature control device 10 specifies an information processing device that processes at least a part of the predicted processing amount so that the power consumption “435 W” of the information processing device 230 is less than 410 W.

  The temperature control device 10 identifies the information processing device 840 that executes the software D indicated in the data region 973d so that the power consumption predicted by the execution of the software B and D is less than the specified value as indicated in the data region 972d. To do. In this way, the temperature control device 10 stops the execution of the software D by the information processing device 230 and causes the information processing device 840 to execute the software D. As shown in the data area 971d, the predicted power consumption of the information processing apparatus 230 is “200 W”, which is smaller than the specified power value (EP1 ′). As shown in the data area 974d, the total power consumption of the information processing devices arranged in the rack “2” is “1470 W”, which is smaller than the specified power value (EP2).

  The information processing apparatus may have a function such that the CPU reaches a designed high temperature and the CPU reduces the processing speed. Such a function does not consider the execution priority of software being executed by the information processing apparatus. On the other hand, the software execution control function unit of the temperature control apparatus 10 involves stopping and starting of software. Such stopping and starting of software interrupts a processing request from an external computer in a short period of time. Therefore, by continuously executing the software with high priority and causing the other information processing apparatus to newly execute the software with low priority, the temperature control device 10 requests processing from the outside for the software with high priority. Can not be interrupted.

[Device Specific Function Unit and Software Execution Control Function Unit (Case 9)]
When there is no device whose predicted power consumption is less than the specified value, the temperature control device 10 stops the execution of software having a low execution priority.

An example of the device specifying function unit and the software execution control function unit when the software is stopped will be described with reference to FIG.
The device management information 950e illustrated in FIG. 20 is an example of the device management information 950 illustrated in FIG. 3A. The device management information 950e includes an information processing device name column 951e, a rack number column 952e, a predicted pre-control power consumption column 955e for each rack, a post-control predicted power consumption column 956e for each rack, and a pre-control predicted power consumption column 953e. . Further, the device management information 950e includes a post-control predicted power consumption column 954e, a software name column 957e before execution control, and a software name column 958e after execution control. The information processing device name column 951e to the software name column 958e after execution control correspond to the information processing device name column 951a to the software name column 958a after execution control shown in FIG.

  As shown in the data area 961e, the total predicted power consumption “1755W” of the information processing devices arranged in the rack with the identification number “1” is larger than the specified power value (EP2) “1680W” for each rack. Further, as shown in the data area 962e, the predicted power before execution control of the information processing apparatus 110 in the rack with the identification number “1” is larger than the specified power value (EP1). Therefore, the software execution control function unit determines that the information processing apparatus 110 arranged in the rack “1” is a target apparatus that makes the predicted power consumption less than the specified value. Further, the software execution control function unit searches for an information processing apparatus for newly executing the software E indicated in the data area 967e. However, as shown in the data area 963e, any information processing apparatus before execution control has high predicted power. Therefore, when software is assigned to the information processing apparatus indicated in the data area 963e, the predicted power after execution control of the information processing apparatus indicated in the data area 963e is equal to or greater than the specified power value (EP1). In this case, the temperature control device 10 performs processing for stopping execution of software having a low execution priority of software.

  The temperature control apparatus 10 refers to the software execution priority information 980a illustrated in FIG. 19 and confirms that the execution priority of the software F is lower than the execution priority of the software E. Therefore, as shown in the data area 969e of FIG. 20, the temperature control apparatus 10 stops the execution of the software F having a low execution priority by the information processing apparatus 240. Then, the temperature control device 10 causes the software E to be executed as shown in the data area 969e, and assigns a part of the processing amount of the execution process of the software E assigned to the information processing device 110 to the information processing device 240. As shown in the data area 968e, the temperature control apparatus 10 can make the predicted power after execution control of the information processing apparatus 110 smaller than the specified power value (EP1). Further, as shown in the data area 964e, the temperature control device 10 sets the total predicted power of the information processing devices arranged in the rack with the identification number “1” to “1555 W” which is less than the specified power value (EP2). .

  The information processing apparatus may have a function such that the CPU reaches a designed high temperature and the CPU reduces the processing speed. Such a function does not consider the execution priority of software being executed by the information processing apparatus. On the other hand, the software execution control function unit of the temperature control apparatus 10 involves stopping and starting of software. Such stopping and starting of software interrupts a processing request from an external computer in a short period of time. Therefore, as described above, even when there is no target to which software is assigned, the temperature control device 10 causes the information processing device to stop the software with the lower execution priority and continues to execute the software with the higher priority on the information processing device. It can be made.

  FIG. 21 is a diagram illustrating an example of queue information storing software names that are stopped. The queue information 985a illustrated in FIG. 21 is an example of the queue information 985 illustrated in FIG. 3A. The queue information 985a shown in FIG. 21 has a software name management column 986a. In the software name management column 986a, the names of software stopped by the software execution control function unit are registered. In the queue information 985a, for example, a software name may be registered according to first-in first-out (FIFO). For example, in the example of case 9, as shown in the data area 987a, the software F stopped by the software execution control function unit is registered on the dequeue side of the software name management column 986a, and the software F is first extracted according to the FIFO. It is.

[Device Identification Function Unit and Software Execution Control Function Unit (Case 10)]
The temperature control device 10 identifies a device whose predicted power consumption is less than a specified value when the stopped software is executed, and assigns the software to the identified device.

  The case 10 will be described below with reference to FIG. FIG. 22 is a diagram illustrating an example of a software execution control function unit when starting software. The device management information 950f illustrated in FIG. 22 is an example of the device management information 950 illustrated in FIG. The device management information 950f includes an information processing device name column 951f, a rack number column 952f, a pre-control predicted power consumption column 955f for each rack, a post-control predicted power consumption column 956f for each rack, and a pre-control predicted power consumption column 953f. . Furthermore, the device management information 950f includes a post-control predicted power consumption column 954f, a software name column 957f before execution control, and a software name column 958f after execution control. The information processing device name column 951f to the post-execution software name column 958f correspond to the information processing device name column 951a to the post-execution control software name column 958a shown in FIG.

  As described with reference to FIG. 18, the software F having a low execution priority is stopped to make the power consumption of the information processing apparatus 110 less than the specified power value. Then, the temperature control device 10 recognizes that the software F is waiting to be executed with reference to the name of the software F registered in the queue information 985a shown in FIG. The predicted power before execution control shown in FIG. 22 indicates a case where software is no longer executed in the information processing apparatus 210 as time passes, as shown in the data area 961f. As shown in the data area 962f, the temperature control device 10 is an information processing device in which the information processing device 210 processes the predicted processing amount because the information processing device 210 does not exceed the specified power value even when the software F is executed. Judge. Therefore, the temperature control apparatus 10 causes the information processing apparatus 210 to execute the software F registered in the queue information 985a illustrated in FIG. 21 as illustrated in the data area 963f.

  The data area 988a in FIG. 21 is software G for newly installing registered in the queue information 985 after the software F is registered in the data area 987a. The temperature control device 10 refers to the name of the software G registered in the queue information 985a shown in FIG. 21, and recognizes that the software G is waiting to be executed after the software F is dequeued. The predicted power before execution control illustrated in FIG. 22 indicates a case where software is no longer executed in the information processing device 840 as time passes, as illustrated in the data area 964f. As shown in the data area 965f, the temperature control device 10 determines that the information processing device 840 is an information processing device that processes a predicted processing amount that does not exceed the specified power value even when the software G is executed. Therefore, the temperature control apparatus 10 causes the information processing apparatus 840 to execute the software G registered in the queue information 985a illustrated in FIG. 21 as illustrated in the data area 966f.

  In this way, the temperature control device 10 has a predicted power even if it is software that has been stopped due to the high throughput of all the information processing devices, or software that has been newly entered into the information processing device and has not been executed. It can be executed by an information processing device within a specified value.

An example of a processing flow in which the temperature control device performs temperature control of the information processing device using a specified value for each information processing device will be described with reference to FIG.
As shown in FIG. 23, the temperature control device 10 determines whether or not the temperature control time has come (S501). When the temperature control time comes (S501 Yes), the temperature control device 10 executes S502 described later. When the temperature control time is not reached (No in S501), the temperature control device 10 waits until the temperature control time is reached (S501). The temperature control time is a time specified such that the temperature control is performed once per hour, for example.

  The temperature control apparatus 10 predicts the processing amount from the processing amount time history information 800 (S502). The predicted time is, for example, a time one hour after the current time. The temperature control apparatus 10 predicts the predicted power consumption of the information processing apparatus when processing at least a part of the predicted processing amount from the power consumption information 850 (S503). The temperature control apparatus 10 determines whether there is an information processing apparatus whose predicted power consumption is equal to or greater than a specified value (S504). The specified value shown in S504 is, for example, the specified power value (EP1) shown in Equation 1. When there is no information processing apparatus in which the predicted power consumption is equal to or greater than the specified value (No in S504), the temperature control apparatus 10 performs S511 described below. When there is an information processing device in which the predicted power consumption is equal to or higher than the specified value (S504 Yes), the temperature control device 10 determines whether one information processing device whose predicted power consumption is equal to or higher than the specified value is executing one software. Is determined (S505).

  When one information processing apparatus whose predicted power consumption is greater than or equal to the specified value is executing more than one software (No in S505), the temperature control apparatus 10 executes S521 described later. When one information processing apparatus whose predicted power consumption is greater than or equal to a specified value is executing one software (Yes in S505), the temperature control apparatus 10 executes S506, which will be described later. As described in case 5, when the temperature control device 10 regards a plurality of software as one set of software, the temperature control device 10 executes S506 described later.

  The temperature control apparatus 10 determines whether there is an information processing apparatus that can process the predicted processing amount so that the predicted power consumption of the information processing apparatus that executes the software is less than the specified value (S506).

  When there is no information processing device that can process the predicted processing amount so that the predicted power consumption of the information processing device that executes software is less than the specified value (No in S506), the temperature control device 10 executes S531 described later. . When there is an information processing device that can process the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than the specified value (Yes in S506), the temperature control device 10 executes S507 described later. . The temperature control apparatus 10 assigns software to the information processing apparatus that can process the predicted processing amount so that the predicted power consumption of the information processing apparatus that executes the software is less than the specified value (S507). In other words, in S507, the temperature control apparatus 10 is an information processing apparatus that can stop the execution of software in the target apparatus that reduces the predicted power consumption to less than the specified value or reduce the processing amount and process the predicted processing amount. Assign and run software.

An example of a processing flow in which the temperature control device performs temperature control of the information processing device with a specified value for each rack will be described with reference to FIG.
As illustrated in FIG. 24, when the condition determination illustrated in S504 is not satisfied, the temperature control device 10 determines whether there is a rack having a total value of predicted power consumption of the arranged information processing devices equal to or greater than a specified value. (S511). The specified value shown in S511 is, for example, the specified power value (EP2) for each rack shown in Equation 2. When there is no rack in which the total value of the predicted power consumption of the arranged information processing apparatuses is equal to or greater than the specified value (No in S511), the temperature control apparatus 10 returns to S501. When there is a rack in which the total value of the predicted power consumption of the information processing apparatuses arranged is greater than or equal to a specified value (S511 Yes), the temperature control device 10 determines whether each information processing apparatus arranged in the rack is executing one software It is determined whether or not (S512). Note that the information processing apparatus to be determined in S512 is, for example, the information processing apparatus with the largest predicted power consumption among the information processing apparatuses arranged in the rack.

  When each information processing device arranged in the rack does not execute one software (No in S512), the temperature control device 10 executes S521 described later. When the information processing device arranged in the rack is executing one software (S512 Yes), the temperature control device 10 is capable of processing the predicted processing amount so that the predicted power consumption is less than the specified value. It is determined whether or not there is (S513). The specified value shown in S513 is, for example, a specified power value (EP1 ') shown in Equation 3. As described in case 6, when the temperature control apparatus 10 regards a plurality of software as one software, the temperature control apparatus 10 executes S513.

When there is no information processing device that can process the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than the specified value (No in S513), the temperature control device 10 executes S531 described later. . When there is an information processing apparatus that can process the predicted processing amount so that the predicted power consumption of the information processing apparatus that executes the software is less than the specified value (Yes in S513), the temperature control apparatus 10 executes S514 described later. .
The temperature control apparatus 10 assigns software to the information processing apparatus that can process the predicted processing amount so that the predicted power consumption of the information processing apparatus that executes the software is less than the specified value (S514). In other words, in S514, the temperature control apparatus 10 is an information processing apparatus that can stop the execution of software in the target apparatus that reduces the predicted power consumption below the specified value, or reduce the processing amount, and process the predicted processing amount. Assign and run software.

An example of a processing flow in which the temperature control device performs temperature control of the information processing device that executes a plurality of software will be described with reference to FIG.
As shown in FIG. 25, when the condition determination shown in S505 or S512 is not satisfied, the temperature control apparatus 10 executes the next S521. The temperature control apparatus 10 determines whether there is an information processing apparatus capable of processing at least a part of the predicted processing amount by the second software execution so that the predicted power consumption by the first software execution is less than the specified value ( S521). The prescribed value shown in S521 is the prescribed power value (EP1) shown in Equation 1 or the prescribed power value (EP1 ′) shown in Equation 3. The second software is software having a lower execution priority than the first software. When there is no information processing apparatus capable of processing at least a part of the predicted processing amount by the second software execution (No in S521), the temperature control apparatus 10 executes S531 described later shown in FIG. When there is an information processing apparatus capable of processing at least a part of the predicted processing amount due to the execution of the second software (S521 Yes), the temperature control apparatus 10 executes the following S522.

  The temperature control device 10 stops the execution of the second software by the information processing device that executes the first and second software, or reduces the processing amount, and assigns the second software to the specified information processing device. (S522). And the temperature control apparatus 10 waits until it becomes temperature control time again (S501).

An example of a processing flow in which the temperature control apparatus performs temperature control of the information processing apparatus that stops or starts software will be described with reference to FIG.
As shown in FIG. 26, when the condition determination shown in S506, S513, or S521 is not satisfied, the temperature control device 10 refers to the software execution priority information 980 and stops the software with the lower execution priority. (S531). The temperature control apparatus 10 registers the software name of the stopped software in the queue information 985 (S532). The temperature control device 10 determines whether or not the software execution start time has come (S533).

  If the software execution start time is not reached, the temperature control device 10 waits until the software execution start time is reached (No at S533). When the software execution start time is reached (S533 Yes), the temperature control apparatus 10 predicts the processing amount from the processing amount time history information 800 (S534). The temperature control apparatus 10 predicts the power consumption of the information processing apparatus when processing at least a part of the predicted processing amount from the power consumption information 850 (S535). The temperature control apparatus 10 determines whether there is an information processing apparatus that processes the predicted processing amount so that the predicted power consumption becomes less than the specified value when the stopped software is executed (S536). The specified value shown in S536 is the specified power value (EP1) shown in Expression 1 or the specified power value (EP1 ') shown in Expression 3. When there is no information processing device capable of processing the predicted processing amount so that the predicted power consumption is less than the specified value when the stopped software is executed (No in S536), the temperature control device 10 again sets the software execution start time. It waits until it becomes (S533). When there is an information processing apparatus capable of processing the predicted processing amount so that the predicted power consumption is less than the specified value when the stopped software is executed (S536 Yes), the temperature control apparatus 10 performs the following S537. Run.

  The temperature control apparatus 10 acquires the name of the software stored in the queue information 985, and assigns the software corresponding to the software name to the information processing apparatus that can process the predicted processing amount (S537). At this time, the software name stored in the queue information 985 is, for example, software registered in the queue information 985 in S532 or software that is waiting to be newly installed and executed in the information processing apparatus. The temperature control device 10 stands by until the temperature control time is reached again (S501).

Regarding the above embodiment, the following additional notes are disclosed.
[Appendix 1]
A temperature control device for controlling the temperature of a plurality of information processing devices,
Power consumption that varies depending on the processing amount per unit time that is predicted for the processing amount from the time history of the processing amount per unit time that is processed by the information processing device by executing software and that is specified for each information processing device From the power consumption prediction function unit that predicts the power consumption of the information processing apparatus when processing at least a part of the predicted processing amount,
An apparatus specifying function unit that specifies an information processing apparatus that processes the predicted processing amount, and the specified processing amount in the specified information processing apparatus so that the predicted power consumption is less than a specified value. A processing unit that implements a software execution control function unit that assigns software that generates
A temperature control device comprising:
[Appendix 2]
The information processing apparatus is stored in a plurality of areas for each predetermined number,
The temperature control device according to appendix 1, wherein the specified value is a threshold value that makes a total value of the predicted power consumption in an information processing device arranged in the region less than a specified value determined in the region.
[Appendix 3]
The information processing apparatus simultaneously executes first software and second software having a lower execution priority than the first software,
The power consumption that varies depending on the processing amount is the power consumption that varies depending on the processing amount per unit time by the execution of the first and second software,
The power consumption prediction function unit predicts the processing amount of the first and second software from the time history of the processing amount, and the processing amount of the predicted execution of the first and second software Predicting the power consumption of the information processing apparatus when processing at least a part of it from the power consumption that varies depending on the processing amount
The device specifying function unit processes at least a part of a processing amount predicted by the second software execution so that the predicted power consumption by the first or second software execution is less than a predetermined value. Identify the information processing device to be
The software execution control function unit stops execution of the second software by the information processing apparatus that executes the first and second software, and assigns the second software to the identified information processing apparatus The temperature control device according to appendix 1 or 2.
[Appendix 4]
When there is no information processing apparatus that processes the predicted processing amount so that the predicted power consumption is less than a specified value, the software execution control function unit stops the software having a low execution priority. The temperature control apparatus according to any one of appendices 1 to 3.
[Appendix 5]
The temperature control device according to any one of appendices 1 to 4, wherein the time history of the processing amount is obtained by statistically processing the time history of the actually measured processing amount.
[Appendix 6]
The time history of the processing amount shows a time history of the processing amount of software for one day, and shows the time history of the processing amount of software for one day classified by day of the week or date, and
The power consumption prediction function unit selects a time history of the software processing amount for the day classified by the day of the week or date according to a prediction time, and the processing is performed from the time history of the selected processing amount. The temperature control device according to any one of appendices 1 to 5, which predicts an amount.
[Appendix 7]
When there is no information processing apparatus that processes the predicted processing amount so that the predicted power consumption of the information processing apparatus that executes the software is less than a specified value, the software execution control function unit executes The temperature control device according to any one of appendices 1 to 5, wherein software having a low priority is stopped.
[Appendix 8]
The device specifying function unit specifies an information processing device that processes the predicted processing amount so that the power consumption predicted when the stopped or unassigned software is executed is less than a specified value,
The temperature control device according to appendix 7, wherein the software execution control function unit assigns the stopped software or unassigned software to the identified information processing device.
[Appendix 9]
A temperature control device connected to a plurality of information processing devices and controlling the temperature of the information processing device,
Predicting the amount of processing from the time history of the amount of processing per unit time processed by the information processing device by executing software, and from the power consumption specified for each information processing device and changing according to the amount of processing per unit time A power consumption prediction function for predicting power consumption of the information processing apparatus when processing at least a part of the predicted processing amount;
A device specifying function for specifying an information processing device to which the predicted processing amount is allocated, and causing the specified processing amount to cause the predicted processing amount so that the predicted power consumption is less than a specified value. Software execution control function to assign software,
A program to realize
[Appendix 10]
The information processing apparatus is stored in a plurality of areas for each predetermined number,
The program according to appendix 9, wherein the specified value is a threshold value for making a total value of the predicted power consumption in an information processing device arranged in the area less than a specified value determined in the area.
[Appendix 11]
The information processing apparatus simultaneously executes first software and second software having a lower execution priority than the first software,
The power consumption that varies depending on the processing amount is the power consumption that varies depending on the processing amount per unit time by the execution of the first and second software,
The power consumption prediction function predicts the processing amount of the plurality of software from the processing amount time history, and processes at least a part of the predicted processing amount of the plurality of software executions. Predicting the power consumption of the information processing device from the power consumption that varies depending on the amount of processing,
The device specifying function is information for processing at least a part of the processing amount of the predicted plurality of software such that the predicted power consumption of the information processing device executing the plurality of software is less than a predetermined value. Identify the processing device, and
The program according to appendix 9 or 10, wherein the software execution control function assigns the plurality of software to the identified information processing apparatus.
[Appendix 12]
At least one of the plurality of information processing apparatuses simultaneously executes first software and second software having a lower execution priority than the first software;
The power consumption information indicates power consumption that varies depending on the amount of processing per unit time by execution of the first and second software,
The power consumption prediction function predicts the processing amount of the first and second software from the time history of the processing amount, respectively, and at least of the predicted processing amount due to the execution of the first and second software Predicting the power consumption of the information processing device when processing a part from the power consumption information,
The device specifying function processes at least a part of a processing amount predicted by the second software execution so that the predicted power consumption by the first or second software execution is less than a predetermined value. Identify the information processing device,
The software execution control function stops execution of the second software by an information processing device that executes the first and second software, and assigns the second software to the identified information processing device. The program according to appendix 9 or 10.
[Appendix 13]
The program according to any one of appendices 9 to 12, wherein the time history of the processing amount is obtained by statistically processing the time history of the actually measured processing amount.
[Appendix 14]
The time history of the processing amount shows a time history of the processing amount of software for one day, and shows the time history of the processing amount of software for one day classified by day of the week or date, and
The power consumption prediction function selects a time history of the processing amount of the software for one day classified by the day of the week or date according to a prediction time, and the processing amount from the time history of the selected processing amount The program according to any one of appendices 9 to 12, which predicts.
[Appendix 15]
When there is no information processing device that processes the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than a specified value, the software execution control function performs priority of execution. The program according to any one of appendices 9 to 14, which stops software having a lower rank.
[Appendix 16]
The device specifying function specifies an information processing device that processes the predicted processing amount so that the power consumption predicted when the stopped or unassigned software is executed is less than a specified value,
The program according to appendix 15, wherein the software execution control function assigns the stopped software or unassigned software to the identified information processing apparatus.
[Appendix 17]
A temperature control method for controlling a temperature of the information processing apparatus connected to a plurality of information processing apparatuses,
Predicting the processing amount from the time history of the processing amount per unit time processed by the information processing apparatus by executing the software;
Predicting the power consumption of the information processing device when processing at least a part of the predicted processing amount from the power consumption specified for each information processing device and changing according to the processing amount per unit time;
Identifying an information processing device to which the predicted processing amount is allocated so that the predicted power consumption is less than a specified value;
Assigning software that causes the predicted processing amount to the identified information processing apparatus;
A temperature control method comprising:
[Appendix 18]
The information processing apparatus is stored in a plurality of areas for each predetermined number,
The temperature control method according to appendix 17, wherein the specified value is a threshold value for making a total value of the predicted power consumption in an information processing device arranged in the area less than a specified value determined in the area. .
[Appendix 19]
The information processing apparatus executes a plurality of software simultaneously,
The power consumption that varies depending on the amount of processing indicates the power consumption that varies depending on the amount of processing per unit time by execution of a plurality of software,
The power consumption prediction step predicts the processing amount of the plurality of software from the processing amount time history information, respectively, and processes at least a part of the predicted processing amount of the plurality of software. Predicting the power consumption of the information processing device from the power consumption information,
In the device specifying step, information for processing at least a part of the processing amount of the predicted plurality of software so that the predicted power consumption of the information processing device executing the plurality of software is less than a predetermined value. Identify the processing device, and
The temperature control method according to appendix 17 or 18, wherein the software execution control step assigns the plurality of software to the specified information processing apparatus.
[Appendix 20]
At least one of the plurality of information processing apparatuses simultaneously executes first software and second software having a lower execution priority than the first software;
The power consumption information indicates power consumption that varies depending on the amount of processing per unit time by execution of the first and second software,
The power consumption prediction step predicts the processing amount of the first and second software from the time history information of the processing amount, respectively, and the processing amount of the predicted execution of the first and second software Predicting the power consumption of the information processing apparatus when processing at least a part from the power consumption that varies depending on the processing amount,
The device specifying step processes at least a part of the processing amount predicted by the second software execution so that the predicted power consumption by the first or second software execution is less than a predetermined value. Identify the information processing device,
The software execution control step stops the execution of the second software by the information processing apparatus that executes the first and second software, and assigns the second software to the specified information processing apparatus; The temperature control method according to appendix 17 or 18.
[Appendix 21]
The temperature control method according to any one of appendices 17 to 20, wherein the time history of the processing amount is obtained by statistically processing the time history of the actually measured processing amount.
[Appendix 22]
The time history of the processing amount shows a time history of the processing amount of software for one day, and shows the time history of the processing amount of software for one day classified by day of the week or date, and
The power consumption prediction step selects a time history of the software processing amount for the day classified by the day of the week or date according to a prediction time, and the processing amount from the time history of the selected processing amount The temperature control method according to any one of appendices 17 to 21, which predicts.
[Appendix 23]
In the software execution control step, execution priority is given when there is no information processing device that processes the predicted processing amount so that the predicted power consumption of the information processing device that executes the software is less than a specified value. The temperature control method according to any one of appendices 17 to 22, wherein software having a lower rank is stopped.
[Appendix 24]
The device identification control step identifies an information processing device that processes the predicted processing amount so that the power consumption predicted when the stopped or unassigned software is executed is less than a specified value. And
24. The temperature control method according to appendix 23, wherein the software execution control step assigns the stopped software or unassigned software to the identified information processing apparatus.

DESCRIPTION OF SYMBOLS 10 Temperature control apparatus 50 Power supply 110-240 Information processing apparatus 410 Power monitoring apparatus 420 Computer 430 Load distribution apparatus 800 Processing time history information 850 Power consumption information 900 Apparatus setting information 950 Apparatus management information 980 Software execution priority information 985 Queue information 990 program

Claims (7)

A temperature control device for controlling the temperature of a plurality of information processing devices,
Power consumption that varies depending on the processing amount per unit time that is predicted for the processing amount from the time history of the processing amount per unit time that is processed by the information processing device by executing software and that is specified for each information processing device From the power consumption prediction function unit that predicts the power consumption of the information processing apparatus when processing at least a part of the predicted processing amount,
An apparatus specifying function unit that specifies an information processing apparatus that processes the predicted processing amount, and the specified processing amount in the specified information processing apparatus so that the predicted power consumption is less than a specified value. A processing unit that implements a software execution control function unit that assigns software that generates
A temperature control device comprising: The information processing apparatus is stored in a plurality of areas for each predetermined number,
2. The temperature control device according to claim 1, wherein the specified value is a threshold value that makes a total value of the predicted power consumption in an information processing device arranged in the area less than a specified value determined in the area. . The information processing apparatus simultaneously executes first software and second software having a lower execution priority than the first software,
The power consumption that varies depending on the processing amount is the power consumption that varies depending on the processing amount per unit time by the execution of the first and second software,
The power consumption prediction function unit predicts the processing amount of the first and second software from the time history of the processing amount, and the processing amount of the predicted execution of the first and second software Predicting the power consumption of the information processing apparatus when processing at least a part of it from the power consumption that varies depending on the processing amount,
The device specifying function unit processes at least a part of a processing amount predicted by the second software execution so that the predicted power consumption by the first or second software execution is less than a predetermined value. Identify the information processing device to be
The software execution control function unit stops execution of the second software by the information processing apparatus that executes the first and second software, and assigns the second software to the identified information processing apparatus The temperature control device according to claim 1 or 2.   When there is no information processing apparatus that processes the predicted processing amount so that the predicted power consumption is less than a specified value, the software execution control function unit stops the software having a low execution priority. The temperature control apparatus of any one of Claims 1-3. A temperature control device connected to a plurality of information processing devices and controlling the temperature of the information processing device,
Predicting the amount of processing from the time history of the amount of processing per unit time processed by the information processing device by executing software, and from the power consumption specified for each information processing device and changing according to the amount of processing per unit time A power consumption prediction function for predicting power consumption of the information processing apparatus when processing at least a part of the predicted processing amount;
A device specifying function for specifying an information processing device to which the predicted processing amount is allocated, and causing the specified processing amount to cause the predicted processing amount so that the predicted power consumption is less than a specified value. Software execution control function to assign software,
A program to realize The plurality of information processing devices are stored in a plurality of areas for each predetermined number,
6. The prescribed value is for making a total value of the predicted power consumption of a predetermined number of information processing devices arranged in the area less than a prescribed value defined in the area. Program. A temperature control method for controlling a temperature of the information processing apparatus connected to a plurality of information processing apparatuses,
A step of predicting the processing amount from a time history of a processing amount per unit time processed by an information processing device that executes software; and power consumption that is specified for each information processing device and varies depending on the processing amount per unit time Predicting power consumption of the information processing apparatus when processing at least a part of the predicted processing amount;
Identifying an information processing device to which the predicted processing amount is allocated so that the predicted power consumption is less than a specified value;
Assigning software that causes the predicted processing amount to the identified information processing apparatus;
A temperature control method comprising: