JP2015021661A - Air conditioning control method and air conditioning control system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a virtual machine optimum cooling method capable of reducing power consumption of an entire system while maintaining service quality.SOLUTION: An information processing apparatus includes: a first information processing apparatus that includes first hardware resources; and a second information apparatus. An air conditioning control method includes: a first step of determining the second information processing apparatus as an activated information processing apparatus candidate if a value obtained by dividing a used amount of the first hardware resources by an upper limit of the first hardware resources exceeds a predetermined threshold; a second step of determining the second information processing apparatus as an information processing apparatus candidate cooled in advance on the basis of a preset number of information processing apparatuses; and a third step of determining an information apparatus to be cooled in advance and to be activated from the activated information processing apparatus candidate and the information processing apparatus candidate cooled in advance so that a sum of power consumption of the first information processing apparatus and the second information processing apparatus and power consumption of the air conditioner is a minimum.

Description

  The present invention relates to an air conditioning control method and an air conditioning control system, and more particularly to an air conditioning control method and an air conditioning control system according to a cooling method for an information processing apparatus.

  In recent years, with the progress of the information society, there has been a demand for power saving of information processing apparatuses installed in communication facilities and data centers and cooling air conditioners that cool the information processing apparatuses. The information processing apparatus refers to, for example, an ICT (Information and Communication Technology) apparatus, a router, or a server. Various attempts have been made to save power in the information processing apparatus and the cooling air conditioner that cools the information processing apparatus.

  In accordance with the operation status of the information processing apparatus, efforts are being made to save power of the information processing apparatus by concentrating loads on some information processing apparatuses using virtual machines. The virtual machine is software that executes a virtualized computer after the resources such as a CPU (Central Processing Unit) and a memory of the computer are virtualized by flexible division and integration regardless of a physical configuration. If a virtual machine is used, an OS (Operating System) and applications can be executed by the same operation as a physical computer (physical machine).

  For example, at night and on weekends, the operation rates of CPUs and memories of a plurality of information processing apparatuses installed in communication facilities and data centers are low. It is possible to reduce the power consumption of the entire information processing apparatus by consolidating the load on a small number of information processing apparatuses using a virtual machine and powering off the power of the information processing apparatus that does not aggregate the load. For example, in Non-Patent Document 1, a technique for selecting and determining an information processing apparatus to be activated in consideration of resource increase / decrease of the information processing apparatus has been put into practical use.

  When operating an actual service, it is often necessary to limit information processing apparatuses that operate virtual machines depending on service requirements and license requirements.

  In the technique of Non-Patent Document 1, an information processing device to be activated is determined according to a resource change of the information processing device while satisfying a rule regarding the correspondence relationship between the virtual machine and the information processing device. However, there is only a viewpoint of reducing the number of activated information processing apparatuses, and reduction of power consumption of the entire system including the air conditioner cannot be expected.

  In addition, when the information processing device located far from the air conditioner is activated and when the information processing device located near the air conditioner is activated, the power required for cooling the information processing device varies. This is because, when trying to cool an information processing device far from the air conditioner, the temperature of the air increases due to the air transport cost (fan power) and the heat transfer from the walls and floor when the air moves. This is because it is necessary to set the temperature of the air to be low. In addition, since the momentum of the air is weakened away from the air conditioner and hot air is likely to flow around, the set temperature of the air conditioner needs to be lowered in advance. When the information processing apparatus to be operated is located close to the air conditioner, it is advantageous in terms of power saving. However, whether or not the information processing apparatus is close to the air conditioner depends on the equipment conditions. Therefore, by appropriately arranging the information processing apparatus to be activated and operating only the air conditioners close to the information processing apparatus to be activated, there is a possibility that power can be reduced without lowering the service quality level.

VMware Distributed Resource Management: Design, Implementation, and Lessons Learned, http://labs.vmware.com/download/157/ Hitachi, “Computer-friendly power saving solution for computer equipment (floor blowing air conditioning system)”, http://www.hitachi-pt.co.jp/ref/ref_it.html

  In general, air conditioners are provided in areas such as the information processing equipment room, communication equipment room, and data center, and the set temperature of the air conditioner is controlled to keep the suction temperature of the information processing equipment below a predetermined value. . With respect to the air conditioner cooling setting for cooling the entire room, the power consumption of the air conditioner can be greatly reduced by partially cooling only the periphery of the activated information processing apparatus. For example, in Non-Patent Document 2, when the operation rate of the entire information processing apparatus is low, the information processing apparatuses are aggregated in a specific range, and the number of air conditioners operated is controlled while avoiding temperature deviation due to temperature rise at the time of aggregation. Thus, a method for reducing the power consumption of the information processing apparatus and the air conditioner has been proposed.

  However, while it takes 1 to 5 minutes from the start command of the information processing device to the actual startup, it takes 10 to change the setting for the air conditioner and the air conditioner exhibits the cooling capacity. It takes about 30 minutes. That is, if the setting of the air conditioner is changed after the information processing apparatus needs to be started due to a change in the information processing apparatus resource, the cooling cannot be performed in time and temperature deviation occurs. Therefore, there is a problem that the service quality may be greatly reduced. On the other hand, if the activation of the information processing apparatus is delayed in order to avoid temperature deviation, there is a problem that service provision is delayed.

  The present invention has been made in view of such problems, and an object of the present invention is to provide an air conditioning control method and an air conditioning control system that reduce power consumption of the entire system while maintaining service quality. is there.

  The present invention for solving the above problem is a data center management system for optimizing business arrangement and air conditioning setting of a plurality of information processing devices arranged in a target area air-conditioned by a plurality of air conditioners, Load balancing in a cluster taking into account rules between information processing device groups and virtual machine groups, rules between information processing device groups, information storage unit that saves power and temperature, and rules And the information processing device to be started / stopped so that the total value of the power consumption of the information processing device and the air conditioner is minimized under the condition not exceeding the set value according to the increase / decrease of the resource amount And a preliminary cooling information processing apparatus decision to determine a list of minimum information processing apparatuses to be cooled in advance based on a predetermined condition The information processing device that minimizes the power consumption of the air conditioner and the information processing device, the information processing device that performs the preliminary cooling, and the set temperature are determined from the information processing device that is running, the information processing device that is activated, and the information processing device that is to be precooled The system includes an overall optimization unit and an information collection unit that collects information necessary for an operation for optimizing the business arrangement and the air conditioning setting of the information processing device from the air conditioning and the information processing device.

  In order to achieve such an object, the present invention provides an air conditioning control method by an information processing apparatus and a system for managing an air conditioner that cools the information processing apparatus, wherein the information The processing apparatus includes a first information processing apparatus having a first hardware resource and a second information processing apparatus having a second hardware resource, and includes at least the first information processing apparatus and the second information processing apparatus. When a value obtained by dividing an upper limit value of the first hardware resource from a usage amount of the first hardware resource exceeds a predetermined threshold in the group including the information processing device, the second information Based on a first step for determining a candidate of an information processing device for starting a processing device and a preset number of units, the second information processing device is determined as a candidate for an information processing device for cooling in advance. Candidate information processing devices to be activated so that the total value of the second step and the power consumption of the first information processing device and the second information processing device and the power consumption of the air conditioner is minimized. And a third step of determining an information processing device to be cooled and activated in advance from the candidates of information processing devices to be cooled in advance.

  As described above, according to the present invention, it is possible to realize power saving of the entire system while maintaining service quality.

It is a lineblock diagram showing the air-conditioning control system concerning one embodiment of the present invention. It is a figure which shows the aspect of a virtual machine group and an information processing apparatus group. It is a figure which shows the power consumption value of information processing apparatus when the various resource values of information processing apparatus are each allocated to a virtual machine. It is a figure which shows the data showing the change of the power consumption of an air conditioner according to the blowing temperature of an air conditioner, the return temperature of an air conditioner, external temperature, and preset temperature. Corresponding discharge temperature of the air conditioner which is a diagram showing a variation of the suction temperature of the information processing apparatus I _{1 ~} information processing apparatus I ₁₀ when the changed 1 ° C.. Is a diagram illustrating the suction temperature variation of the information processing apparatus I _{1 ~} information processing apparatus I ₁₀ when the heating value of the corresponding information processing apparatus is 100W changed. Corresponding discharge temperature of the air conditioner which is a graph showing a change amount of the return temperature of the air conditioner A _{1 ~} air conditioner A ₃ when the changed 1 ° C.. Is a graph showing a change amount of the return temperature of the air conditioner A _{1 ~} air conditioner A ₃ when the heating amount is 100W corresponding changes in the information processing apparatus. It is a flowchart which shows the air-conditioning control method using the virtual machine concerning one Embodiment of this invention. Is a diagram showing an example of the relationship from the information processing apparatus I ₁ and CPU utilization and CPU utilization across the cluster of the information processing apparatus I _5. It is a flowchart which shows the load distribution method concerning one Embodiment of this invention. It is a flowchart which shows the search method of the starting / stop candidate of information processing apparatus concerning one Embodiment of this invention. It is a flowchart which shows the search method of the information processing apparatus which pre-cools concerning one Embodiment of this invention. It is a figure which shows the example which determines the number of the minimum information processing apparatuses preliminarily cooled for every virtual machine group and every information processing apparatus group. It is a flowchart which shows the information processing apparatus which performs preliminary cooling, the information processing apparatus which starts and stops, and the search method of the preset temperature of an air conditioner concerning one Embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(Structure of the invention)
FIG. 1 shows three configurations of an air conditioning control system, an information processing apparatus, and an air conditioner according to an embodiment of the present invention. The information processing apparatus I _i acquires information measured by the virtual machine management unit 11 that operates the virtual machine, the temperature sensor unit 12, the power sensor unit 13, and the temperature sensor unit 12 and the power sensor unit 13. And a device management unit 14. The virtual machine management unit 11 includes virtual machines 111, 112,. The air conditioner A _j includes a temperature sensor unit 21, a power sensor unit 22, and an air conditioning management unit 23 that acquires information measured by the temperature sensor unit 21 and the power sensor unit 22.

Air conditioning control system 1000 according to an embodiment of the present invention, a power control unit 5 for controlling the power of the information processing apparatus _{I i} and the air conditioner _{A j,} the information processing apparatus _{I i} of the virtual machines 111 and 112, ..., X And a virtual machine control unit 6 for controlling the above. The air conditioning control system 1000 according to an embodiment of the present invention includes an air conditioning control unit 8 that controls the air conditioner A _j , and an optimization unit 7 that performs load balancing of the virtual machine and determination of an information processing device for preliminary cooling. Is provided. Furthermore, the air conditioning control system 1000 according to an embodiment of the present invention includes an information collecting unit 4 that collects power consumption information of the information processing device I _i and the air conditioner A _j , suction temperature information of the information processing device I _i , and the like. And an information storage unit 3 for storing the information. The optimization unit 7 starts and stops a load distribution unit 71 that performs load distribution described later, an information processing device power supply optimization unit 72 (first determination unit) that searches for start and stop candidates for the information processing device, and the like. Overall optimization unit 73 (third determination unit) that determines the information processing device, the information processing device that performs preliminary cooling, and the set temperature of the air conditioner, and the preliminary cooling information processing device that searches for candidates for information processing devices that perform preliminary cooling A determination unit 74 (second determination unit). Information storage unit 3, information collection unit 4, power supply control unit 5, virtual machine control unit 6, optimization unit 7, air conditioning control unit 8, virtual machine management unit 11, and information processing device management unit 14 The air conditioning management unit 23 is connected via the network 9. The network 9 is configured by a LAN (Local Area Network), for example.

  The information storage unit 3 stores cluster information that determines ranges of a plurality of information processing apparatuses that perform load distribution and power control for transferring a virtual machine to the information processing apparatus. In this specification, a cluster refers to a collection of information processing apparatuses that can be controlled by the air conditioning control system, and the cluster includes a plurality of groups of information processing apparatuses having different purposes. The information storage unit 3 stores information on a virtual machine group that is a set of a plurality of virtual machines and an information processing apparatus group that is a set of a plurality of information processing apparatuses. The information storage unit 3 stores rules for the virtual machine and the information processing apparatus. For example, since a license is tied to a specific information processing apparatus, there is a case where it is desired to move a certain virtual machine only by a specific information processing apparatus. As a rule, a rule for moving a certain virtual machine group by a certain information processing apparatus group, a rule for not moving a certain virtual machine group by a certain information processing apparatus, or the like can be considered.

  The upper limit value of resources (hardware resources such as CPU, memory, network, and I / O) that can be used by the virtual machine is determined for each information processing apparatus, and information related to the upper limit value of the resource is also stored in the information storage unit 3 in advance. ing.

  When there is a temperature sensor unit that measures the suction temperature and a power sensor unit that measures the power of the entire information processing device for each information processing device, the information collection unit 4 sends an SNMP (Simple Network Management) from the information processing device management unit 14. Information measured by the temperature sensor unit and the power sensor unit is collected by Protocol) or IPMI (Intelligent Platform Management Interface). Note that the temperature sensor unit and the power sensor unit do not necessarily have to be inside the information processing apparatus, and may be external sensors.

  When the information processing apparatus is in the activated state, the suction temperature of the information processing apparatus must be maintained below a certain value. The upper limit value of the suction temperature of the information processing apparatus is determined in advance for the device, and information regarding the upper limit value of the suction temperature determined in advance is also stored in the information storage unit 3 in advance. The information storage unit 3 can also collect information on the upper limit value of the suction temperature by SNMP, IPMI, or the like.

  The air conditioner management unit 23 manages the air outlet temperature information, the return temperature information of the air conditioner, and the power consumption information of the air conditioner. The blowout temperature of the air conditioner is the temperature in the vicinity of the cold air outlet of the indoor unit, and is measured by a temperature sensor installed in the vicinity of the outlet. The return temperature of the air conditioner is the temperature near the air inlet of the indoor unit, and is measured by a temperature sensor installed near the air inlet. The information collecting unit 4 collects the air temperature blower temperature information, the air conditioner return temperature information, and the air conditioner power consumption information from the air conditioning management unit 23, and the information storage unit 3 collects the information collected by the information collecting unit 4. save.

  The information processing device power supply optimization unit 72 searches for a stop of the information processing device when the processing amount of the entire cluster decreases and the processing amount of each information processing device in the cluster decreases. The resource usage rate of the entire cluster that triggers a search for stopping the information processing apparatus is set in advance and stored in the information storage unit 3. The resource usage rate of the entire cluster is the sum of the resource usage of all virtual machines on all information processing devices in the cluster divided by the sum of the upper limit values of resources of all information processing devices in the cluster. Value.

  The information processing apparatus power supply optimization unit 72 searches for the stop of some information processing apparatuses in the cluster when the processing amount of the entire cluster decreases and there are remaining information processing apparatuses that are active in the entire cluster.

  On the other hand, when the processing amount of the entire cluster increases and the processing amount of each information processing device in the cluster increases, the information processing device power supply optimization unit 72 searches for activation of the information processing device. The resource usage rate of the entire cluster that is a trigger for searching for activation of the information processing apparatus is set in advance and stored in the information storage unit 3.

  The information processing device power supply optimization unit 72 searches for activation of several information processing devices in the cluster when the processing amount of the entire cluster increases and there are insufficient information processing devices that are active in the entire cluster. .

  Further, when the additional information processing apparatus is activated due to an increase in resources, the stopped information processing apparatus is pre-cooled so that the information processing apparatus can be activated immediately. The number of preliminary cooling is set for each information processing device group and stored in the information storage unit 3. When the resource increases, the pre-cooled information processing apparatus is activated based on the information of the pre-cooled information processing apparatus stored in the information storage unit 3, and the information processing is promptly adapted to the resource fluctuation.

The virtual machine control unit 6 can transfer a virtual machine between information processing apparatuses in response to a command from the load distribution unit 71. The CPU, memory, and network I / O of the information processing apparatus are consumed for each virtual machine, and the virtual machine management unit 11 manages information related to the CPU, memory, and network I / O consumption of the information processing apparatus. . The information collecting unit 4 periodically collects information related to consumption of the CPU, memory, and network I / O of the information processing apparatus I _i from the virtual machine management unit 11.

The air conditioning control unit 8 can change the set temperature of the air conditioner A _j in response to a command from the overall optimization unit 73. The set temperature of the air conditioner referred to in this specification may be either the return temperature of the air conditioner or the outlet temperature of the air conditioner, and the air conditioner is controlled so that the return temperature or the outlet temperature follows the set temperature.

  FIG. 2 shows aspects of the virtual machine group and the information processing apparatus group. In FIG. 2, the first group 210 includes a first information processing device 211, a second information processing device 212, and a third information processing device 213. The second group 220 includes a fourth information processing device 224, a fifth information processing device 225, and a sixth information processing device 226. The third group 230 includes a second information processing device 212, a third information processing device 213, a fifth information processing device 225, and a sixth information processing device 226. The fourth group 240 includes a seventh information processing device 247. Like the second information processing device 212, the third information processing device 213, the fifth information processing device 225, and the sixth information processing device 226, a group is overlapped with the information processing device or the virtual machine. May be set. Further, there may be an information processing apparatus that does not belong anywhere, such as an eighth information processing apparatus 258.

When assigned respectively various resource value of the information processing apparatus I _i to a virtual machine 3 shows the power consumption value of the information processing apparatus I _i a (W). Various resource values include the CPU operating frequency (MHz), memory capacity (MB), and network I / O speed (Mbps). The power consumption of the information processing apparatus I _i changes according to various resource values respectively assigned to the virtual machines. The information storage unit 3 stores the relationship table information shown in FIG. The relation table information shown in FIG. 3 is used when determining an information processing apparatus to be started / stopped, which will be described later. The relation table information may be stored in the information storage unit 3 in advance, or the relation table information may be acquired while operating the air conditioning control system of the present embodiment. Further, instead of the relationship table, the relationship with the power consumption may be stored as a function such as a linear function regarding parameters of the CPU operating frequency, memory capacity, and network I / O speed.

A linear function related to parameters of the CPU operating frequency, memory capacity, and network I / O speed is expressed by the following (Equation 1). (Expression 1) represents how much the power consumption of the information processing apparatus changes when the CPU operating frequency, memory capacity, and network I / O speed of each information processing apparatus are changed.
ΔH _i = f _i (ΔD ₁ , ΔD ₂ ,..., ΔD _n , ΔM ₁ , ΔM ₂ ,..., ΔM _n , ΔN ₁ , ΔN ₂ ,..., ΔN _n ) (Formula 1)

In (Equation 1), i = 1~n, n is the number of information processing apparatus, [Delta] H _i is the power consumption of the amount of change in the information processing apparatus _{I i,} [Delta] D _i is the change in the operating frequency of the CPU of the information processing apparatus _{I i} ΔM _i is the amount of change in the memory capacity of the information processing device I _i , ΔN _i is the amount of change in the network I / O speed of the information processing device I _i , and the amount of change f _i is the suction temperature of the i th information processing device Represents a function related to temperature change. The function f _i can be given by a linear expression of ΔD _i , ΔM _i , ΔN _i in (Equation 2), for example.
f _i (ΔD ₁ , ΔD ₂ , ..., ΔD _n , ΔM ₁ , ΔM ₂ , ..., ΔM _n , ΔN ₁ , ΔN ₂ , ..., ΔN _n ) = a ₁ ΔD ₁ + a ₂ ΔD ₂ + ... + a _n ΔD _n + b ₁ ΔM ₁ + b ₂ ΔM ₂ +… + b _n ΔM _n + c ₁ ΔN ₁ + c ₂ ΔN ₂ +… + c _n ΔN _n (Formula 2)

In (Equation 2), a _i represents the coefficient power consumption H _i of the information processing apparatus I _i is set in consideration that the influence by the change in operating frequency D _i of the CPU of the information processing apparatus I _i. b _i represents the coefficient power consumption H _i of the information processing apparatus I _i is set in consideration that the influence by the change in memory capacity M _i of the information processing apparatus I _i. c _i denotes a coefficient power H _i of the information processing apparatus I _i is set in consideration that the influence by the change of the network I / O speed N _i of the information processing apparatus I _i. (Expression 1) and (Expression 2) are used when determining an information processing apparatus to be activated / deactivated, which will be described later.

  FIG. 4 shows data representing changes in the power consumption of the air conditioner according to the air outlet temperature, the return temperature of the air conditioner, the outside air temperature, and the set temperature. The power consumption of the air conditioner varies depending on the blowout temperature, return temperature, outside air temperature, and set temperature. The term “outside temperature” as used herein refers to the temperature outside where the outdoor unit of the air conditioner is installed. The information storage unit 3 stores information on the relationship between the blowout temperature of the air conditioner, the return temperature of the air conditioner, the outside air temperature, the set temperature of the air conditioner, and the change in power consumption of the air conditioner. The relationship information may be calculated from the parameter sheet in advance, or the relationship information may be acquired while operating. Further, the information storage unit 3 determines the relationship between the power consumption, not the table, according to the air temperature blowout temperature, the air conditioner return temperature, the outside air temperature, the air conditioner set temperature, and the change in the air conditioner power consumption. You may preserve | save as functions, such as a linear function regarding a parameter.

The information processing apparatus I _i differs in ease of cooling depending on the position where it is installed, and the amount of change in the suction temperature differs for each air conditioner. In addition, due to heat generated by the power consumption of the information processing apparatus, the suction temperature changes according to the power consumption of the information processing apparatus. The information storage unit 3 stores the relationship between the change in the blowout temperature of the air conditioner and the return temperature of the air conditioner, and the relationship between the heat generation amount of the information processing device and the return temperature of the air conditioner.

5 shows the amount of change in the suction temperature of the information processing device I ₁ to the information processing device I ₁₀ when the air temperature of the air conditioner corresponding to FIG. 5 changes by 1 ° C., and the heat generation amount of the information processing device corresponding to FIG. The amount of change in the suction temperature of the information processing device I ₁ to the information processing device I _{10 when} changed is shown. The information shown in FIGS. 5 and 6 may be stored in the information storage unit 3 in advance, or the relationship may be acquired while operating. In addition, the relationship between the amount of change in the suction temperature of the information processing device instead of the table may be stored as a function such as a linear function related to the parameters of the blowout temperature of the air conditioner and the power consumption of the information processing device. The two tables shown in FIGS. 5 and 6 are collectively referred to as the information processing device suction temperature influence degree. The function of the information processing device suction temperature influence degree is expressed by the following (Equation 3). (Expression 3) represents how much the suction temperature of the information processing device changes when the blowout temperature of each air conditioner and the power consumption of each information processing device are changed.
ΔT _i = f ′ _i (ΔS ₁ , ΔS ₂ ,..., ΔS _m , ΔH ₁ , ΔH ₂ ,..., ΔH _n ) (Formula 3)

In (Expression 3), i = 1 to n, n is the number of information processing devices, j = 1 to m, m is the number of air conditioners A _j , ΔT _i is the amount of change in the suction temperature of the information processing device I _i , [Delta] S _j is outlet temperature variation of an air conditioner a _j located j-th, [Delta] H _i is the power consumption of the change amount of the i-th information processing apparatus, f _i is the suction temperature of the i-th information processing apparatus Represents a function related to temperature change. The function f _i can be given by a linear expression of ΔS _j and ΔH _i in (Expression 4), for example.
f ' _i (ΔS ₁ , ΔS ₂ ,…, ΔS _m , ΔH ₁ , ΔH ₂ ,…, ΔH _n ) = d ₁ ΔS ₁ + d ₂ ΔS ₂ +… + d _m ΔS _m + e ₁ ΔH ₁ + e ₂ ΔH ₂ +… + e _n ΔH _n (Formula 4)

In (Expression 4), d _j is the air conditioner A _j in which the suction temperature T _i of the information processing device located at the i-th (i = 1 to n) is located at the j-th (j = 1 to m). A coefficient set in consideration of the influence of the change in the blowing temperature S _j is expressed. e _i is determined by a change in power consumption H _i of the information processing device in which the suction temperature T _i of the information processing device located at the i-th (i = 1 to n) is located at the i-th (i = 1 to n). This represents the coefficient set in consideration of the influence.

The coefficient d _j measures the change in the suction temperature T _i of the information processing device when the air-conditioning blowout temperature S _j and the power consumption H _i of the information processing device are changed for a certain period, and uses multiple regression analysis as learning data You may obtain | require by etc. Or you may obtain | require by changing the power consumption of an air conditioning machine and the power consumption of information processing apparatus by fluid simulation.

7 shows the amount of change in return temperature of the air conditioners A ₁ to A ₃ when the air temperature of the air conditioner corresponding to FIG. 7 changes by 1 ° C., and the heat generation amount of the information processing device corresponding to FIG. The change amount of the return temperature of the air conditioner A ₁ to the air conditioner A ₃ at the time is shown.

  The return temperature of the air conditioner is also different in the amount of change with respect to the air temperature of the air conditioner. The information storage unit 3 stores the relationship between the change in the blow-off temperature of the air conditioner and the return temperature of the air conditioner, and the relationship information between the heat generation amount of the information processing device and the return temperature of the air conditioner. The relationship information may be entered in advance, or the relationship information may be acquired while operating. Further, not the table but the relationship with the return temperature of the air conditioner may be stored as a function such as a linear function related to the parameters of the air outlet temperature and the power consumption of the information processing apparatus.

(Basic operation of the invention)
FIG. 9 is a flowchart showing an air conditioning control method according to an embodiment of the present invention. The optimization unit 7 acquires the configuration information of the information processing apparatus and the configuration information of the virtual machine (S901). The information processing apparatus configuration information and the virtual machine configuration information include information processing apparatus arrangement information and virtual machine allocation information in the information processing apparatus. The optimization unit 7 monitors the operation information of the information processing apparatus and the power consumption information of the air conditioner (S902). The operation information of the information processing apparatus includes the usage amounts of various resources shown in FIG. 3, the power consumption amount with respect to the usage amount of various resources, and the relational expressions (Equation 1) and (Equation 2) of the power consumption amount with respect to the usage amounts of various resources. )including. The optimization unit 7 calculates the resource deviation amount (S903). The resource deviation is the resource usage rate of the information processing device obtained by dividing the resource usage of all the virtual machines on one specific information processing device by the upper limit value of the resource of one information processing device. Is the absolute value of the value obtained by subtracting the resource usage rate of the entire cluster.

CPU utilization of the information processing apparatus I ₅ from the information processing apparatus I ₁ and FIG. 10 shows an example of the relationship between the CPU utilization of the entire cluster. The CPU usage rate of the entire cluster is set to 20% (A). As illustrated in FIG. 10, the CPU usage rate of the information processing device I ₁ is 20%, the CPU usage rate of the information processing device I ₂ is 40%, the CPU usage rate of the information processing device I ₃ is 10%, and the information processing device I The CPU usage rate of No. ₄ is 30%, and the CPU usage rate of the information processing apparatus I ₅ is 15%. Therefore, compared with the CPU usage rate of the entire cluster, the CPU deviation of the information processing apparatus I ₁ is 0%, the CPU deviation of the information processing apparatus I ₂ is 20%, and the CPU deviation of the information processing apparatus I ₃ is The CPU deviation of the information processing apparatus I ₄ is 10%, and the CPU deviation of the information processing apparatus I ₅ is 5%. The load distribution unit 71 performs load distribution when the CPU deviation amount of the information processing apparatus I _i exceeds the first set value. In FIG. 10, the first set value is set to 15%, which is represented by a line segment of ± 15% with reference to the CPU usage rate 20% (A) of the entire cluster, and a line segment of + 15% is represented by a line segment B, A -15% line segment is represented by line segment B '. Deviations of the CPU 10 in the information processing apparatus I ₂ is 20%, the excess first set value (B), a small amount of deviation of the CPU using virtual machines, for example, the information processing apparatus I ₁ Distribute the load.

The optimization unit 7 determines whether or not there is an information processing apparatus having a deviation amount larger than the first set value (B, B ′) (S904). If there is an information processing apparatus I _i having a larger deviation amount than the first set value (B, B ′), the process proceeds to step S905. If there is no information processing apparatus having a larger deviation amount than the first set value (B, B ′), the process proceeds to step S906. The load distribution unit 71 performs load distribution described in detail in FIG. 10 (S905). The optimization unit 7 calculates the resource usage rate of the entire cluster (S906). The optimization unit 7 determines whether the resource usage rate of the entire cluster is outside the allowable range (S907). If the resource usage rate of the entire cluster is outside the allowable range, the process proceeds to step S908. If the resource usage rate of the entire cluster is not outside the allowable range, the process returns to step S902. The information processing apparatus power supply optimization unit 72 searches for candidates for starting / stopping the information processing apparatus (S908) (first step). The preliminary cooling information processing device determination unit 74 searches for an information processing device to be preliminary cooled (S909) (second step). The overall optimization unit 73 searches for information processing devices that perform preliminary cooling, information processing devices that are started and stopped, and set temperatures of air conditioners (S910) (third step). The air conditioning control unit 8 changes the set temperature of the air conditioner, and the power supply control unit 5 starts and stops the information processing apparatus (S911). Step S908 will be described in detail with reference to FIG. 12, step S909 with reference to FIG. 13, and step S910 with reference to FIG.

(load distribution)
FIG. 11 is a flowchart showing a load distribution method according to an embodiment of the present invention. The load distribution unit 71 acquires operation information of the information processing apparatus (S1001). The load distribution unit 71 searches for information processing apparatuses whose resource deviation exceeds the first set value in the group in the cluster. The load distribution unit 71 extracts one virtual machine from the searched information processing apparatuses (S1002). The load distribution unit 71 extracts an information processing apparatus having the smallest deviation amount from all the information processing apparatuses permitted by the rule as the migration destination of the selected virtual machine (S1003). The load distribution unit 71 determines whether all virtual machines in the cluster have been searched (S1004). When searching for all virtual machines in the cluster, the process advances to step S1005. If all virtual machines in the cluster have not been searched, the process returns to step S1002. The load distribution unit 71 moves the virtual machine to the information processing apparatus so that the current deviation amount is minimized (S1005). The load distribution unit 71 determines whether or not the deviation amount is equal to or less than the first set value (S1006). If the deviation amount is not less than or equal to the first set value, the process returns to step S1001. When the deviation amount is equal to or less than the first set value, the load distribution is terminated and the process proceeds to step S905 shown in FIG.

  The load distribution unit 71 periodically monitors the load in the cluster, and migrates the virtual machine so that the deviation amount of the resources in the cluster is uniform. The deviation amount of all information processing devices in the cluster is calculated. Deviation amounts of all information processing apparatuses in the cluster are calculated for each resource CPU, memory, and network I / O. The load distribution unit 71 determines whether or not the deviation amount of the resource exceeds the first set value set in advance for the CPU, the memory, and the network I / O.

  When there is a deviation amount that exceeds the first set value that has been determined in advance, among all virtual machines in the cluster, the deviation amount of the current resource is the smallest and the migration destination of the virtual machine that does not violate the rules The information processing apparatus is searched.

  The load distribution unit 71 performs steps S1001 to S1005 until the resource deviation amount is equal to or less than a first set value determined in advance.

  The load distribution unit 71 uses steps S1001 to S1006 to select the position of the information processing device based on the operating rate of the information processing device and the increase or decrease of the information processing device resource, and to which information processing device the virtual machine is assigned to Decide whether to migrate.

(Search for information processing device start / stop candidates)
FIG. 12 is a flowchart showing a method for searching for candidates for starting / stopping the information processing apparatus according to an embodiment of the present invention. The information processing device power supply optimization unit 72 extracts one combination of information processing devices in the cluster (S1101). The information processing apparatus power supply optimization unit 72 calculates the resource usage rate and the deviation amount of the entire cluster of the information processing apparatus when the load is distributed for the selected combination of information processing apparatuses (S1102). The information processing apparatus power supply optimization unit 72 determines whether the resource usage rate of the entire cluster falls within the second set value (S1103). If the resource usage rate of the entire cluster falls within the second set value, the process proceeds to step S1104. If the resource usage rate of the entire cluster does not fall within the second set value, the process returns to step S1101. The information processing device power supply optimization unit 72 determines whether or not the deviation amounts of all the information processing devices are within the second set value (S1104). When the deviation amounts of all the information processing apparatuses are within the second set value, the process proceeds to step S1105. If the deviation amount of the information processing apparatus does not fall within the second set value, the process returns to step S1101. The information processing apparatus power supply optimization unit 72 adds the selected combination of information processing apparatuses to the combination list of information processing apparatus candidates to be activated / deactivated (S1105). The information processing device power supply optimization unit 72 determines whether or not all combinations of information processing devices have been tried (S1106). In step S1106, the search for candidates for starting / stopping the information processing apparatus is terminated, and the process proceeds to step S908 shown in FIG.

  The information processing device power supply optimization unit 72 determines a candidate information processing device to be activated when the resource usage rate of the entire cluster exceeds the third set value (predetermined threshold). Further, when the resource usage rate of the entire cluster falls below the fourth set value, a candidate information processing apparatus to be stopped is determined.

  When it is determined that the information processing apparatus is to be stopped, the virtual machines in all the information processing apparatuses activated in the cluster are assigned to other information processing apparatuses for all the information processing apparatuses activated in the cluster. And the load distribution unit 71 distributes the load. The information processing apparatus power supply optimization unit 72 calculates the resource usage rate and the resource deviation amount of the entire cluster when the load is distributed. Then, the resource usage rate of the entire cluster of information processing devices falls within the second set value (S1103), and the resource deviation amount does not exceed the first set value in all the information processing devices (S1104). A list of combinations of information processing devices is listed (S1105).

  When it is determined that the information processing device is to be activated, all the information that is stopped in the cluster is compared to all the information processing devices that are stopped in the cluster among the information processing devices that are precooled. The processing device is activated and the load distribution unit 71 distributes the load. The information processing apparatus power supply optimization unit 72 calculates the resource usage rate and the resource deviation amount of the entire cluster when the load is distributed. Then, the resource usage rate of the entire cluster of the information processing apparatus falls within the second set value (S1103), and the resource deviation amount does not exceed the first set value in all the information processing apparatuses (S1104). Lists combinations of processing devices (S1105).

  If there is no pre-cooled information processing device, the pre-cooled information processing device is read as all the information processing devices that have stopped.

  The information processing apparatus power supply optimization unit 72 uses, in steps S1101 to S1106, resources of the information processing apparatuses that are operating (at least in groups including information processing apparatuses that are operating and information processing apparatuses that are not operating). Information for starting an information processing device that is not operating when a value obtained by dividing the upper limit value of the resource of the information processing device that is operating from the usage amount of the first hardware resource) exceeds a predetermined threshold value It is determined as a candidate for the processing device.

(Search for information processing equipment to be precooled)
FIG. 13 is a flowchart showing a method for searching for an information processing apparatus to be precooled according to an embodiment of the present invention. The preliminary cooling information processing device determination unit 74 acquires the number information of information processing devices to be preliminary cooled for each group (S1201). The preliminary cooling information processing device determination unit 74 selects one combination from the list of information processing device combinations that are candidates for starting and stopping the information processing device (S1202). The preliminary cooling information processing device determination unit 74 selects information processing device candidates corresponding to the number of information processing devices to be preliminary cooled for each group (S1203). The preliminary cooling information processing device determination unit 74 stores the combination having the smallest number of information processing devices to be preliminary cooled (S1204). The preliminary cooling information processing device determination unit 74 determines whether or not all combinations of information processing devices have been tried (S1205). If all combinations of information processing apparatuses have been tried, the process advances to step S1206. If all combinations of information processing apparatuses have not been tried, the process returns to step S1203. The preliminary cooling information processing device determination unit 74 adds the combination of information processing devices to be precooled to the combination list of information processing devices to be precooled together with the information processing devices to be started / stopped, and information processing device candidates to be started / stopped. (S1206). The preliminary cooling information processing device determination unit 74 determines whether or not all combinations of information processing device candidates to be activated / stopped have been tried (S1207). If all the start / stop candidate host combinations have not been tried, the process returns to step S1202. When all combinations of start / stop candidate hosts have been tried, the search for the information processing apparatus to be precooled is terminated, and the process proceeds to step S909 shown in FIG.

  The preliminary cooling information processing device determination unit 74 searches for a candidate for the preliminary cooling information processing device for all combinations of information processing devices listed in the information processing device power supply optimization unit 72. When searching, an information processing device to be pre-cooled is determined in consideration of rules determined for the information processing device group. Set in advance how many or how many resources to pre-cool for each group. The preliminary cooling information processing device determination unit 74 calculates combinations of information processing devices to be precooled for each group, and extracts a list of combinations with the smallest number of units to be activated.

  The preliminary cooling information processing device determination unit 74 uses steps S1201 to S1207 to determine information processing devices that are not operating as candidates for information processing devices that cool in advance based on the number of units determined in advance in the group. To do.

  FIG. 14 shows an example in which the minimum number of information processing devices to be precooled is determined for each virtual machine group and each information processing device group. In FIG. 14, the first group 310 includes a first information processing device 311, a second information processing device 312, and a third information processing device 313. The second group 320 includes a fourth information processing device 324, a fifth information processing device 325, and a sixth information processing device 326. The third group 330 includes a second information processing device 312, a third information processing device 313, a fifth information processing device 325, and a sixth information processing device 326.

  The relationship between the number of preliminary cooling units for each group and the minimum number of preliminary cooling units for the entire system will be described. For example, in the case of FIG. 14, the number of information processing apparatuses to be pre-cooled for each group is determined, the first group 310 and the second group 320 are one, and the third group 330 is two. . As a combination that minimizes the number of preliminary cooling units, the second information processing device 312 and the fifth information processing device 325, the third information processing device 313, the fifth information processing device 325, and the second information processing device 312 There are four combinations of the sixth information processing device 326, the third information processing device 313, and the sixth information processing device 326. From the four combinations, the combination that minimizes the power consumption of the air conditioner and the information processing device is found brute force.

(Search for pre-cooling information processing device, starting / stopping information processing device, and set temperature of air conditioner)
FIG. 15 is a flowchart showing an information processing apparatus for preliminary cooling, an information processing apparatus for starting and stopping, and a method for searching for a set temperature of an air conditioner according to an embodiment of the present invention. The overall optimizing unit 73 extracts one combination from a combination list of information processing device candidates to be precooled and information processing device candidates to be activated / deactivated (S1301). The overall optimizing unit 73 extracts one air conditioning setting candidate for setting the information processing apparatus during startup and preliminary cooling to a set temperature (S1302). The overall optimization unit 73 calculates the power consumption of the air conditioner and the entire information processing device from the set temperature of the air conditioner, the candidate of the information processing device to be precooled, and the candidate of the information processing device to be started / stopped (S1303). When calculating the power consumption of the entire information processing apparatus, the relational expressions (Equation 1) and (Equation 2) of the power consumption with respect to the usage amount of various resources are referred. The overall optimization unit 73 determines whether a combination of set temperatures of all the air conditioners has been tried (S1304). If a combination of set temperatures of all the air conditioners has been tried, the process proceeds to step S1305. When the combination of the set temperatures of all the air conditioners has not been tried, the process returns to step S1302. If the power consumption becomes the smallest among the candidates searched up to now, the overall optimization unit 73 sets the information processing device to be pre-cooled, the information processing device to be started / stopped, and the set temperature of the air conditioner (S1305). The overall optimization unit 73 determines the positions of the information processing device to be pre-cooled and the information processing device to be started / stopped so that the total power consumption of the information processing device and the air conditioner is minimized. The overall optimization unit 73 determines whether all combinations have been tried (S1306). If all combinations have been tried, the process proceeds to step S1305. If not all combinations have been tried, the process returns to step S1301. The overall optimization unit 73 determines the preset temperature of the information processing device to be precooled, the information processing device to be started / stopped, and the air conditioner (S1307). In step S1307, the search for the pre-cooling information processing device, the start / stop information processing device, and the set temperature of the air conditioner is terminated, and the process proceeds to step S910 shown in FIG.

  The overall optimization unit 73 calculates the power consumption of the air conditioner and the power consumption of the information processing device with respect to the list of combinations of the information processing device to be precooled and the information processing device to be activated. In addition, the overall optimization unit 73 satisfies the upper limit temperature determined for the information processing device, and the start-up information processing device and the precooling information processing that minimize the total of the power consumption of the air conditioner and the power consumption of the information processing device. The combination of the set temperature of the device and the air conditioner is determined.

  The air conditioning setting and activation information processing apparatus determined by the overall optimization unit 73 and the virtual machine to be transferred are instructed to the air conditioning control unit 8, the power supply control unit 5, and the virtual machine control unit 6, respectively.

  Therefore, the overall optimization unit 73 uses steps S1301 to S1307 to select information processing device candidates to be activated and advance information so that the total value of the power consumption of the entire information processing device and the power consumption of the air conditioner is minimized. An information processing device to be cooled and activated is determined in advance from candidates of information processing devices to be cooled.

  According to the present embodiment, it is possible to realize power saving of the entire system while maintaining service quality.

DESCRIPTION OF SYMBOLS 3 Information storage part 4 Information collection part 5 Power supply control part 6 Virtual machine control part 7 Optimization part 8 Air-conditioning control part 9 Network 11 Virtual machine management part 12, 21 Temperature sensor part 13, 22 Power sensor part 14 Information processing apparatus management part 23 Air conditioning management unit 71 Load distribution unit 72 Information processing device power supply optimization unit 73 Overall optimization unit 74 Precooling information processing device determination unit 111, 112, X Virtual machine 210, 310 First group 211, 311 First information Processing device 212, 312 Second information processing device 213, 313 Third information processing device 220, 320 Second group 224, 324 Fourth information processing device 225, 325 Fifth information processing device 226, 326 Sixth Information processing device 230, 330 Third group 240 Fourth group 247 Seventh information processing device 2 8 eighth information processing apparatus 1000 air conditioning control system A _j air conditioner I _i information processing apparatus

Claims (2)

An air conditioning control method by a system that manages an information processing apparatus and an air conditioner that cools the information processing apparatus, wherein the information processing apparatus includes a first information processing apparatus having a first hardware resource and a second hardware A second information processing apparatus having a wear resource,
In a group including at least the first information processing apparatus and the second information processing apparatus, a value obtained by dividing the upper limit value of the first hardware resource from the usage amount of the first hardware resource is A first step of determining, as a candidate for an information processing device that activates the second information processing device, when a predetermined threshold is exceeded;
A second step of determining the second information processing apparatus as a candidate of an information processing apparatus that cools the second information processing apparatus in advance based on a preset number of units;
The candidate of the information processing device to be activated and the cooling in advance so that the total value of the power consumption of the first information processing device and the second information processing device and the power consumption of the air conditioner is minimized. An air conditioning control method comprising: a third step of determining an information processing device to be cooled and activated in advance from candidates for the information processing device to perform. An air conditioning control system that manages an information processing device and an air conditioner that cools the information processing device, wherein the information processing device includes a first information processing device having a first hardware resource and a second hardware resource. A second information processing apparatus having
In a group including at least the first information processing apparatus and the second information processing apparatus, a value obtained by dividing the upper limit value of the first hardware resource from the usage amount of the first hardware resource is A first determination unit that determines a candidate of an information processing device that activates the second information processing device when a predetermined threshold is exceeded;
A second determination unit that determines a candidate of an information processing apparatus that cools the second information processing apparatus in advance based on a preset number of units;
The candidate of the information processing device to be activated and the cooling in advance so that the total value of the power consumption of the first information processing device and the second information processing device and the power consumption of the air conditioner is minimized. An air conditioning control system comprising: a third determining unit that determines an information processing device to be cooled and activated in advance from candidates for the information processing device to perform.

Images