JP2017084013A - Power management program, power management method, and power management apparatus - Google Patents

Abstract

An object of the present invention is to provide a power management program capable of realizing power control with reduced influence of processing efficiency of a data center. When a DR request representing a power saving request is received, a stop in each group among a plurality of operating server devices managed in a group for each usage type in a data center that executes power control A determination step for determining a possible server device and a first stop step for stopping the server device determined in the determination step are executed by a computer. [Selection] Figure 1

Description

  This case relates to a power management program, a power management method, and a power management apparatus.

  There is known a mechanism called Demand Response (Demand Response: DR) that demands power saving from consumers such as homes and factories without adjusting the power generation amount when power supply is tight, and adjusts the power supply and demand. Further, Open Automated Demand Response (OpenADR) is also known as a standard (protocol) for automatically controlling DR (see, for example, Patent Document 1).

  When power saving is requested to the consumer by OpenADR, the power consumption of the air-conditioning equipment and lighting equipment installed in the customer's facility is automatically controlled according to a predetermined power saving condition. Thereby, power saving is implemented.

Japanese Patent Laid-Open No. 2015-22553

  As described above, there is a mechanism for automatically controlling the power consumption of air-conditioning equipment and lighting equipment using OpenADR, but a mechanism for automatically controlling the power consumption of a data center that centrally manages server devices is known. Not.

  For example, in the United States, low-rise data centers are often built in the suburbs, and if the construction site is cold, the temperature inside the data center is adjusted using the outside air, so the power consumption of air conditioning equipment can be reduced. . Therefore, it is assumed that power supply tightness itself does not occur so much. However, in Japan, data centers are often built in urban areas, and it is required to make the data center a multi-storey building. In this case, high-performance air conditioners other than the lighting equipment must be installed in the building for the number of floors, and a large amount of power is consumed only for controlling the lighting equipment and the air conditioning equipment. In other words, a mechanism for simply controlling air-conditioning equipment and lighting equipment is insufficient, and a new mechanism specific to the data center is required for the data center.

  In view of this, an object of one aspect is to provide a power management program, a power management method, and a power management apparatus capable of realizing power control with reduced influence of data center processing efficiency.

  The power management program disclosed in the present specification has a plurality of operating server devices that are managed by being grouped for each use type in a data center that executes power control when a DR request representing a power saving request is received. The power management program which makes a computer perform the determination step which determines the server apparatus which can be stopped in each group, and the 1st stop step which stops the server apparatus determined by the said determination step.

  In the power management method disclosed in the present specification, when a DR request indicating a power saving request is received, a plurality of operating server devices that are managed by being grouped for each use type in a data center that executes power control A power management method in which a computer executes a determination step of determining a server device that can be stopped in each group, and a first stop step of stopping the server device determined in the determination step.

  The power management apparatus disclosed in this specification has a plurality of operating server apparatuses that are managed by being grouped for each use type in a data center that executes power control when a DR request representing a power saving request is received. The power management device includes a determination unit that determines a server device that can be stopped in each group, and a control unit that stops the server device determined by the determination unit.

  According to the power management program, the power management method, and the power management apparatus disclosed in the present specification, it is possible to realize power control in which the influence of the processing efficiency of the data center is reduced.

FIG. 1 is a diagram for explaining an example of an ADR system. FIG. 2 is an example of a hardware configuration of the power management apparatus. FIG. 3 is an example of a functional block diagram of the power management apparatus. FIG. 4 is an example of a server information table. FIG. 5 is an example of a server status table. FIG. 6 is an example of a DR contract information table. FIG. 7 is a flowchart illustrating an example of the operation of the power management apparatus. FIG. 8 is an example of a graph modeling the relationship between the CPU usage rate and the power consumption. FIG. 9 is another example of a graph showing the relationship between the CPU usage rate and the power consumption. FIG. 10 is a flowchart illustrating another example of the operation of the power management apparatus. FIG. 11 is a flowchart illustrating a part of the operation of the power management apparatus.

  Hereinafter, an embodiment for carrying out this case will be described with reference to the drawings.

  FIG. 1 is a diagram for explaining an example of the ADR system S. As shown in FIG. 1, the ADR system S includes a node 100 managed by the power company EP, a node 200 managed by the aggregator AGR, a data center management device 300 installed in the data center DC, a power management device 400, and a plurality of A server device 500 is included. The node 100 and the node 200 are connected via the Internet NW1. The node 200 and the data center management device 300 are connected via the Internet NW2. Although not shown in the figure, in addition to the data center DC, other data centers, and consumers such as homes and factories are connected to the Internet NW2.

  When the power supply is tight, the power company EP transmits a DR request indicating a power saving request to the data center DC. Specifically, the node 100 transmits a DR request according to OpenADR. The DR request is sometimes called an event or a DR event, for example. The DR request includes, as power saving conditions, a power saving contract date and time indicating the date and time when the power saving contract is started, a power saving end date and time indicating the date and time when the power saving contract is ended, a power reduction amount, and the like.

  The DR request transmitted from the power company EP is received by the aggregator AGR before reaching the data center DC. The aggregator AGR is an operator that mediates between the electric power company EP, the data center DC, and the above-described consumer. The aggregator AGR collects in advance data centers DC and consumers that can provide power in order to respond to DR requests from the power company EP, and when receiving the DR request, the aggregator AGR select. When the selection is completed, the aggregator AGR reduces and distributes the power reduction amount to the selected data center DC or customer, and determines the contract power consumption after the data center DC or customer performs power saving. When the aggregator AGR determines the contract power consumption, it transmits a DR request. When the aggregator AGR selects the data center DC, the node 200 transmits a DR request to the data center management apparatus 300 according to OpenADR. The DR request transmitted by the node 200 includes the determined contract power consumption.

  The data center DC receives the DR request transmitted from the aggregator AGR. Specifically, the data center management device 300 that manages the data center DC receives the DR request. That is, the data center management apparatus 300 corresponds to a window that first receives a DR request transmitted from the aggregator AGR. When the data center management apparatus 300 receives the DR request, it issues a power saving execution request and transmits it to the power management apparatus 400.

  The power management apparatus 400 manages the power consumption of the plurality of server apparatuses 500. In addition, the power management apparatus 400 manages the operating states of the plurality of server apparatuses 500. When a communication failure occurs between the server apparatuses 500, the location of the communication failure is displayed on the monitor monitor 410 based on an alarm issued from the server apparatus 500. Display the time, contents, etc. When the power management apparatus 400 receives the power saving execution request from the data center management apparatus 300, the power management apparatus 400 stops the server apparatus 500 according to the power saving conditions. Thereby, power saving of the data center DC is performed.

  The power management apparatus 400 periodically transmits the reduced power amount to the data center management apparatus 300. The node 200 collects the amount of power received by the data center management device 300 and transmits a DR representing a power saving response to the node 100. That is, the aggregator AGR reports the power saving performance to the power company EP. When receiving the DR, the power company EP pays the aggregator AGR a reward for responding to the DR request. When the aggregator AGR receives a reward from the power company EP, the aggregator AGR pays the data center DC. That is, the ADR system S according to the present embodiment is an incentive type system.

  The power management apparatus 400 stops the server apparatus 500 and then starts the server apparatus 500 in consideration of the influence of the processing efficiency of the data center DC. As a result, the quality of services provided by the data center DC to the users who use the data center DC is not reduced. Hereinafter, a detailed configuration and operation of the power management apparatus 400 will be described.

  Next, the hardware configuration of the power management apparatus 400 described above will be described with reference to FIG. Note that the nodes 100 and 200, the data center management device 300, and the server device 500 described above have basically the same hardware configuration as that of the power management device 400, and a description thereof will be omitted.

  FIG. 2 is an example of a hardware configuration of the power management apparatus 400. As shown in FIG. 2, the power management apparatus 400 includes at least a central processing unit (CPU) 400A, a random access memory (RAM) 400B, a read only memory (ROM) 400C, and a network I / F (interface) 400D. . The power management apparatus 400 may include at least one of a hard disk drive (HDD) 400E, an input I / F 400F, an output I / F 400G, an input / output I / F 400H, and a drive apparatus 400I as necessary. The CPU 400A, RAM 400B, ROM 400C, network I / F 400D, HDD 400E, input I / F 400F, output I / F 400G, input / output I / F 400H, and drive device 400I are connected to each other via an internal bus 400J. At least the CPU 400A and the RAM 400B cooperate to realize a computer.

An input device 710 is connected to the input I / F 400F. Examples of the input device 710 include a keyboard and a mouse.
A display device 720 is connected to the output I / F 400G. Examples of the display device 720 include a liquid crystal display and a monitoring monitor 410.
A semiconductor memory 730 is connected to the input / output I / F 400H. Examples of the semiconductor memory 730 include a universal serial bus (USB) memory and a flash memory. The input / output I / F 400H reads a program and data stored in the semiconductor memory 730.
The input I / F 400F and the input / output I / F 400H include, for example, a USB port. The output I / F 400G includes a display port, for example.

A portable recording medium 740 is inserted into the drive device 400I. Examples of the portable recording medium 740 include a removable disk such as a Compact Disc (CD) -ROM and a Digital Versatile Disc (DVD). The drive device 400I reads a program and data recorded on the portable recording medium 740.
The network I / F 400D includes, for example, a port and a physical layer chip (PHY chip). The power management apparatus 400 is connected to a network NW3 such as a local area network (LAN) via a network I / F 400D. A data center management device 300 is connected to the network NW3.

  In the RAM 400B described above, the program stored in the ROM 400C or the HDD 400E is stored by the CPU 400A. In the RAM 400B, the program recorded on the portable recording medium 740 is stored by the CPU 400A. When the stored program is executed by the CPU 400A, various functions to be described later are realized, and various operations to be described later are executed. In addition, what is necessary is just to make a program according to the flowchart mentioned later.

  Next, functions of the above-described power management apparatus 400 will be described with reference to FIGS.

  FIG. 3 is an example of a functional block diagram of the power management apparatus 400. FIG. 4 is an example of the server information table T1. FIG. 5 is an example of the server state table T2. FIG. 6 is an example of the DR contract information table T3.

  As shown in FIG. 3, the power management apparatus 400 includes a server information storage unit 401, a server state storage unit 402, a contract information storage unit 403, a server load monitoring unit 404 as a monitoring unit, and a stopped server determination unit 405 as a determination unit. , And a server control unit 406 as control means. The power management apparatus 400 manages a plurality of server apparatuses 500 by grouping for each set usage type (for example, usage).

  Specifically, the plurality of server devices 500 belonging to the A group realize DB access and calculation functions. The plurality of server devices 500 belonging to the B group realize a communication control function. A plurality of server devices 500 belonging to the C group realize a WEB server function. The plurality of server devices 500 belonging to the D group realize a backup function. Note that a plurality of server devices 500 may be grouped from the A group into a group other than the D group according to the usage type.

  The server information storage unit 401 stores server information that is information related to the server device 500. More specifically, as shown in FIG. 4, the server information storage unit 401 manages server information using a server information table T1. The server information includes a server host name, an IP address, and a belonging group as components. The server host name is a host name assigned to each of the plurality of server devices 500. The IP address is an IP address assigned to each of the plurality of server devices 500. The belonging group is a group to which a plurality of server devices 500 belong. According to the server information table T1, the power management apparatus 400 manages X server apparatuses 500 as an A group. Similarly, the power management apparatus 400 manages Y server apparatuses 500 as a B group, Z server apparatuses 500 as a C group, and W server apparatuses 500 as a D group.

  The server state storage unit 402 stores a server state representing the state of the server device 500. More specifically, as shown in FIG. 5, the server state storage unit 402 manages the server state using the server state table T2. The server status includes IP address, CPU usage rate, disk I / O, memory swap, current status, and schedule control.

  Although the IP address is as described above, the CPU usage rate is the usage rate of the CPU included in each of the plurality of server apparatuses 500. The disk I / O is an average value for a certain time (for example, 5 minutes) of the number of processes in the sleep state waiting for disk input / output. Since the number of processes moves from 0 to 2 in a steady state, the average value is between 0 and 2. However, at high load, the number of processes is, for example, 5 or more, and the average value is also 5 or more. The memory swap is a rate at which the capacity per second (Kbytes / s) of the memory swapped out to the disk within a certain time (for example, 5 minutes) is other than zero. If there is enough memory, the value of memory swap will be almost 0 and the rate will be 0%. However, if the value of memory swap is other than 0 in bursts, the rate may exceed 20%.

  The current state represents the current state of the plurality of server devices 500, and either an operating state or a stopped state (more specifically, an idle state) is registered. The schedule control represents the state of the server device 500 scheduled for the next time of the current state. For example, when the current state of the server device 500 is a stop state, if there is a plan to start the server device 500 next time, the start is registered in the planned state. Conversely, when the server device 500 is stopped at the next time, the scheduled state is unregistered. Further, for example, when the current state of the server device 500 is an operating state, if there is a plan to stop the server device 500 next time, the stop is registered in the planned state. On the other hand, when the server device 500 is to be operated at the next time, the scheduled state is unregistered.

  The contract information storage unit 403 stores contract information. More specifically, as shown in FIG. 6, the contract information storage unit 403 manages contract information using a DR contract information table T3. The contract information includes the contract start date and time, the contract end date and time, and the contract power consumption. The contract information is included in the DR request transmitted by the aggregator AGR. The data center management apparatus 300 extracts the contract information included in the DR request and includes it in the power saving execution request. When receiving the power saving execution request, the power management apparatus 400 (more specifically, the server load monitoring unit 404) registers the contract information in the contract information storage unit 403 and notifies the stopped server determination unit 405 of the reception of the power saving execution request.

  The server load monitoring unit 404 uses the server information stored in the server information storage unit 401 to monitor the load on the server device 500. Specifically, the server load monitoring unit 404 periodically acquires a CPU usage rate, a first numerical value related to disk I / O, and a second numerical value related to memory swap for each server device 500 to which an IP address is assigned. To do. The server load monitoring unit 404 calculates the average value of the number of processes described above using the acquired first numerical value, and calculates the ratio described above using the second numerical value. The server load monitoring unit 404 registers the CPU usage rate, the average value of the number of processes, and the ratio in the server state storage unit 402 as loads. In addition, the server load monitoring unit 404 controls an alarm issued from the server device 500.

  When the data center DC receives the DR request, the stop server determination unit 405 determines a stopable server device 500 in each of the groups A to D from the plurality of operating server devices 500. More specifically, when the power management apparatus 400 receives a power saving execution request, the stopped server determination unit 405 selects each group from a plurality of operating server apparatuses 500 based on the server information, the server state, and the contract information. The number of server apparatuses 500 that can be stopped in A to D is determined. In addition, the stopped server determination unit 405 calculates the power consumption for each group of the server devices 500 and the power consumption of the entire data center DC based on the number of operating server devices 500 and the CPU usage rate. The detailed operation of the stop server determination unit 405 will be described later.

  The server control unit 406 controls the operation of the server device 500. Specifically, the server control unit 406 stops the server device determined by the stop server determination unit 405. In addition, the server control unit 406 reduces the load from the stopped server devices 500 when the load stored in the server state storage unit 402 exceeds a predetermined first threshold (hereinafter referred to as an upper limit threshold). At least one possible server device 500 is activated. Furthermore, the server control unit 406 stops the activated server device 500 when the load stored in the server state storage unit 402 falls below a predetermined second threshold (hereinafter referred to as a lower limit threshold).

  Next, the operation of the power management apparatus 400 when the data center DC receives the DR request will be described with reference to FIG.

  FIG. 7 is a flowchart illustrating an example of the operation of the power management apparatus 400. As shown in FIG. 7, the server load monitoring unit 404 continues to wait until it receives a power saving execution request (step S101: NO). When the server load monitoring unit 404 receives the power saving execution request (step S101: YES), the stopped server determining unit 405 selects an unselected group K (K = A, B, C, or D) ( Step S102). More specifically, the server load monitoring unit 404 registers the contract information included in the power saving execution request in the contract information storage unit 403, and the stopped server determination unit 405 performs subsequent processing in the group stored in the server information storage unit 401. Select one group that is not running.

  When the processing in step S102 is completed, the stopped server determination unit 405 then acquires and averages the CPU usage rates of the server devices 500 belonging to the selected group (hereinafter referred to as a selected group) K (step S103). More specifically, the stopped server determination unit 405 extracts the IP address of the server device 500 belonging to the selected group K from the server information storage unit 401, and uses the extracted IP address to calculate the CPU usage rate from the server state storage unit 402. get. Then, the stop server determination unit 405 averages the acquired CPU usage rate. This is because it is assumed that the load of each server device 500 is equally distributed.

  When the process of step S103 is completed, the stop server determination unit 405 then increments the stoptable number Nk of the K group (step S104). For example, when 0 is set as the initial value of Nk, the stoppable number Nk becomes 1 by increment. Note that the increment may be every plural units (for example, five units).

  When the process of step S104 is completed, the stopped server determination unit 405 then calculates a predicted value V1 of the power consumption after stopping the server device 500 in the selected group K (step S105). Specifically, the stop server determination unit 405 calculates the predicted value V1 using the following calculation formula (1).

V1 = {(Pmax−Pidle) × m / 100 + Pidle} × (Xk−Nk)
... (1)
Pmax: Maximum power consumption of one server device 500 Pidle: Power consumption when one server device 500 is stopped m: CPU usage rate changed by reducing the number of operating devices Xk: Servers operating in the K group Number of devices 500 Nk: Number of K groups that can be stopped

  When the process of step S105 is completed, the stop server determination unit 405 then determines whether there is an unselected group (step S106). Here, when the stop server determination unit 405 determines that there is an unselected group (step S106: YES), the processing of steps S102 to S105 described above is repeated. Thereby, the predicted value V1 of another group that has not been selected is also calculated. On the other hand, when the stop server determination unit 405 determines that there is no unselected group (step S106: NO), the predicted value V1 of each group is added up to calculate the predicted value V2 of the power consumption of the entire data center DC. (Step S107). For example, when the stoppable number NA = 1, NB = 1, NC = 1, and ND = 1, the stop server determination unit 405 predicts the predicted value V2 when the server devices 500 belonging to the groups A to D are stopped one by one. Is calculated.

  When the process of step S107 is completed, the stopped server determination unit 405 then determines whether or not the predicted value V2 is below the contract power consumption (step S108). More specifically, the stop server determination unit 405 acquires and determines the contract power consumption from the contract information storage unit 403. When the stop server determination unit 405 determines that the predicted value V2 is not less than the contract power consumption (step S108: NO), that is, when it is expected that power saving is insufficient, the processing of steps S102 to S107 is repeated. As a result, the stoppable number NA, NB, NC, ND of each group A to D is increased, and a new predicted value V2 is calculated.

  On the other hand, when the stop server determination unit 405 determines that the predicted value V2 is less than the contract power consumption (step S108: YES), that is, when it is expected that power saving is sufficient, the stop server determination unit 405 can be stopped. The number of server devices 500 that are different is determined (step S109). That is, the stop server determination unit 405 determines the stoppable numbers NA, NB, NC, and ND that have calculated the predicted value V2 that is lower than the contract power consumption.

  Then, when the process of step S109 is completed, the server control unit 406 stops the server devices 500 for the determined number (step S110). More specifically, the server control unit 406 transmits a command for stopping the server device 500 to the server device 500 for each group. As a result, the power consumption of the entire data center DC falls below the contract power consumption, and power saving is completed. In particular, when the server device 500 is randomly stopped, for example, the stop of the server devices 500 belonging to the group A may be concentrated, so that the service provided by the data center DC may be reduced. However, in the present embodiment, since the stop of the server device 500 is distributed in units of groups, the possibility that the service provided by the data center DC is reduced is avoided.

  Here, the reduction of the power consumption will be specifically described with reference to FIG.

  FIG. 8 is an example of a graph modeling the relationship between the CPU usage rate and the power consumption. As shown in FIG. 8, the power consumption of one server device 500 changes linearly in proportion to the CPU usage rate. For example, when the CPU usage rate is 100%, the maximum power consumption Pmax of the server device 500 is 300 W. Conversely, when the CPU usage rate is 0%, that is, when the server device 500 is stopped, the power consumption Pidle is 200 W. If the number XA of the server devices 500 belonging to the A group that is operating without stopping is 50 and the average value of the CPU usage rate of the 50 server devices 500 is 50%, the stopable number NA is calculated as follows. Assuming 0 units, the power consumption of 50 units belonging to the A group is 12,500 W (watts) using the above-described calculation formula (1).

  Similarly to the A group, the B group, the C group, and the D group have an XB = XC = XD = 50 units, an average value of the CPU usage rate of 50%, and an NB = NC = ND = 0 unit. The power consumption of 50 units belonging to each of the D groups is 12,500 W. Therefore, the power consumption of the entire data center DC before stopping the server device 500 is 50,000 W, which is the sum of the power consumption of the A to D groups. Therefore, when the contract power consumption is 40,000 W, it is required to stop the server device 500 in order to save power.

  On the other hand, if NA = 15 units are stopped from XA = 50 units, if the average value of CPU usage increases from 50% to 70%, the power consumption of 35 units in operation belonging to Group A Is 9,450 W (watts) using the above-described calculation formula (1). Similarly to the A group, the B group, the C group, and the D group have XB = XC = XD = 50 units, an average value of 70% of the CPU usage rate, and NB = NC = ND = 15 units. The power consumption of the 35 operating units belonging to each of the D groups is 9,450 W. Therefore, the power consumption of the entire data center DC after the server apparatus 500 is stopped is 37,800 W by adding the power consumption of the A to D groups. Therefore, when the contract power consumption is 40,000 W, this means that the power saving request has been met.

FIG. 9 is another example of a graph showing the relationship between the CPU usage rate and the power consumption.
As described above, the stop server determination unit 405 calculates the power consumption using the graph shown in FIG. Instead of the graph shown in FIG. 8, as shown in FIG. 9, a graph in which the increase in power consumption per server device 500 is geometrical with respect to the increase in the CPU usage rate or this is modeled. The stop server determination unit 405 may calculate the power consumption using the graph.

  The graph shown in FIG. 9 uses a function called a performance state (P-state) to reduce power consumption. P-state is a function that reduces power consumption by changing the combination of the operating voltage and clock of the CPU.

  Next, the operation of the power management apparatus 400 while the data center DC is in the power saving period will be described with reference to FIG.

  FIG. 10 is a flowchart illustrating another example of the operation of the power management apparatus 400. As shown in FIG. 10, the power management apparatus 400 repeats the processing from steps S201 to S209 for each set period (for example, every 5 to 10 minutes) during the power saving period. Specifically, when the processing of step S107 shown in FIG. 7 is completed and the data center DC shifts to the power saving period, the server load monitoring unit 404 monitors the load of the server device 500 (step S201).

  More specifically, the server load monitoring unit 404 acquires the CPU usage rate, the first numerical value related to the disk I / O, and the second numerical value related to the memory swap from each of the plurality of server apparatuses 500. The server load monitoring unit 404 calculates the average value of the number of processes described above using the acquired first numerical value, and calculates the ratio described above using the second numerical value. The server load monitoring unit 404 registers the CPU usage rate, the average value of the number of processes, and the ratio in the server state storage unit 402 as loads.

  When the process of step S201 is completed, the stop server determination unit 405 determines whether or not the load exceeds the upper limit threshold (step S202). More specifically, the stop server determination unit 405 determines whether any of the CPU usage rate, the average value of the number of processes, or the ratio exceeds an upper limit threshold corresponding to each. For example, 85% is set as the upper limit threshold for the CPU usage rate, 5 processes are set as the upper limit threshold for the average value of the number of processes, and 20% is set as the upper limit threshold for the ratio. These upper thresholds may be changed as appropriate according to the quality of service provided by the data center DC.

  Here, when the stopped server determination unit 405 determines that the load exceeds the upper limit threshold (step S202: YES), the server device 500 capable of reducing the load is started from the stopped server devices 500 (step S203). . More specifically, the stop server determination unit 405 exceeds the upper limit threshold when it is determined that the CPU usage rate, the average value of the number of processes, or the ratio exceeds the corresponding upper limit threshold value, thereby causing a high load. The server device 500 is activated. By starting the server device 500, the load is distributed and the load is reduced. When the stop server determination unit 405 determines that the load does not exceed the upper limit threshold (step S202: NO), the process of step S203 is skipped.

  When the process of step S202 or S203 is completed, the stop server determination unit 405 then calculates the power consumption V3 of the entire data center DC (step S204), and whether or not the power consumption V3 is less than the contract power consumption. Is determined (step S205). That is, when the server apparatus 500 is activated by the process of step S203, the power consumption V3 of the entire data center DC changes. Therefore, the power consumption V3 after starting the server device 500 is calculated, and it is determined whether or not the power saving condition is satisfied. In addition, when calculating the power consumption V3, the stopped server determination unit 405 uses the number of operating devices after the server device 500 is started, the CPU usage rate that is changed by increasing the number of operating devices, and the calculation formula (1) described above. To do.

  If the stop server determination unit 405 determines that the power consumption V3 is not less than the contract power consumption (step S205: NO), the stop server determination unit 405 determines the number of server devices 500 that can be stopped (step S206). The control unit 406 stops the server devices 500 for the determined number (step S207). In other words, when the server stop unit 405 determines that the server device 500 has been started so that the power saving condition is not satisfied, the server control unit 406 stops the server device 500 again. Thereby, the power saving condition is satisfied again. Note that when the stop server determination unit 405 determines the number of server devices 500 that can be stopped, the processing of steps S102 to S108 described above is used.

  When the process of step S205 or S207 is completed, the stop server determination unit 405 determines whether or not the load falls below the lower limit threshold (step S208). More specifically, the stop server determination unit 405 determines whether any of the CPU usage rate, the average value of the number of processes, or the ratio is below the corresponding lower threshold. For example, 5% is set as the lower limit threshold for the CPU usage rate, 2 processes are set as the lower limit threshold for the average value of the number of processes, and 5% is set as the lower limit threshold for the ratio. These lower thresholds may be appropriately changed according to the power consumption of the entire data center DC.

  Here, when the stop server determination unit 405 determines that the load falls below the lower limit threshold (step S208: YES), the server control unit 406 stops the activated server device 500 (step S209). More specifically, when the stop server determination unit 405 determines that the load is low because any one of the CPU usage rate, the average value of the number of processes, or the ratio falls below the corresponding lower limit threshold value, the stop server determination unit 405 is below the lower limit threshold value. The server device 500 is stopped. By stopping the server device 500, wasteful power consumption can be suppressed while maintaining the service provided by the data center DC. When the stop server determination unit 405 determines that the load does not fall below the lower limit threshold (step S208: NO), the process of step S209 is skipped. When the processing from step S201 to S209 described above is completed, the processing from step S201 to S209 is repeated again after the set period has elapsed. When the power saving period ends, the power management apparatus 400 stops the process shown in the flowchart shown in FIG.

  As described above, according to the power management apparatus 400 according to the present embodiment, when the data center DC that executes power control receives the DR request, the stop server determination unit 405 The server devices 500 that can be stopped in each group are determined from among a plurality of operating server devices 500 that are managed in groups. Then, the server control unit 406 stops the server device determined by the stop server determination unit 405. Thereby, the power control which reduced the influence of the processing efficiency of data center DC is realizable.

  Furthermore, with reference to FIG. 11, a process for distinguishing between an alarm based on a communication failure and an alarm based on the stop of the server device 500 according to the present embodiment will be described.

  FIG. 11 is a flowchart illustrating a part of the operation of the power management apparatus 400. First, when the server control unit 406 stops the server device 500 by the process of step S <b> 110 described with reference to FIG. 7, the stopped server device 500 issues an alarm toward the server load monitoring unit 404. However, the alarm issued by the server device 500 is not an alarm based on a communication failure but an alarm based on a stop for power saving. For this reason, as shown in FIG. 11A, when the process of step S110 is completed, the server load monitoring unit 404 excludes the alarm based on the stop according to the present embodiment from the alarm based on the communication failure (step S301). Thereby, the alarm based on the stop according to the present embodiment is excluded from the alarm displayed on the monitor monitor 410 and displayed. As a result, the manager of the data center DC can determine that the alarm is based on a communication failure when the alarm is displayed.

  The server load monitoring unit 404 does not exclude the alarm based on the stop according to the present embodiment, but displays the alarm based on the communication failure and the alarm display based on the stop according to the present embodiment (for example, display color or display size). You may make it display by changing. Thereby, the administrator of the data center DC can distinguish the alarm based on the communication failure and the alarm based on the stop according to the present embodiment.

  Similarly, as illustrated in FIG. 11B, when the process of step S207 is completed, the server load monitoring unit 404 excludes the alarm based on the stop according to the present embodiment from the alarm based on the communication failure (step S302). When the process of step S209 is completed, the server load monitoring unit 404 excludes the alarm based on the stop according to the present embodiment from the alarm based on the communication failure (step S303). Also by these processes, the same effects as those described with reference to FIG. 11A can be obtained.

  The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to the specific embodiments according to the present invention, and various modifications are possible within the scope of the gist of the present invention described in the claims.・ Change is possible.

In addition, the following additional notes are disclosed regarding the above description.
(Supplementary note 1) When a DR request representing a power saving request is received, a stop in each group from among a plurality of operating server devices managed and grouped for each use type in a data center that executes power control The power management program which makes a computer perform the determination step which determines a possible server apparatus, and the 1st stop step which stops the server apparatus determined by the said determination step.
(Supplementary Note 2) A monitoring step of monitoring the load of the server device that has not been stopped in the first stop step, and when the load exceeds a predetermined first threshold, the load is reduced from among the stopped server devices. A starting step of starting at least one server device that can be reduced, and a second stopping step of stopping the server device started in the starting step when the load falls below a predetermined second threshold. The power management program according to Supplementary Note 1, which is characterized.
(Supplementary Note 3) The monitoring step periodically acquires the number of processes related to disk input / output waiting and the ratio related to memory swap for each of the server devices, and the starting step includes either the number of processes or the ratio The power management program according to appendix 2, wherein the at least one server device is activated when the first threshold value is exceeded.
(Supplementary Note 4) The server device is obtained by adding the power consumption amount at the time of stoppage to the multiplication result of the difference value between the maximum power consumption amount of the server device and the power consumption amount when the server device is stopped and the CPU usage rate. By calculating the power consumption per unit for each group and multiplying the power consumption per unit by the number of server devices for each group operating in the data center, The power management program according to any one of appendices 1 to 3, further comprising a calculation step of calculating a power consumption amount consumed in the data center.
(Supplementary Note 5) Using a predetermined graph in which the increase in power consumption per unit of the server device is geometrical with respect to the increase in the CPU usage rate and the CPU usage rate, the server device 1 By calculating the power consumption per unit for each group and multiplying the power consumption per unit by the number of the server devices for each group operating in the data center, The power management program according to any one of supplementary notes 1 to 3, further comprising a calculation step of calculating a power consumption amount consumed at the center.
(Supplementary Note 6) When the power consumption calculated in the calculation step exceeds the contract power consumption included in the DR request, use the predicted value of the power consumption after stopping at least one of the server devices The power management program according to appendix 4 or 5, further comprising a third stop step of stopping the server device that can be stopped.
(Supplementary note 7) Any one of Supplementary notes 1 to 6, further comprising: an exclusion step of excluding an alarm that is issued based on a stop of the server device from an alarm that is issued when the server device is faulty The power management program according to item 1.
(Supplementary note 8) When a DR request indicating a power saving request is received, a stop in each group among a plurality of operating server devices managed in a group for each use type in a data center that executes power control A power management method in which a computer executes a determination step of determining a possible server device and a stop step of stopping the server device determined in the determination step.
(Supplementary note 9) When a DR request representing a power saving request is received, a stop in each group among a plurality of operating server devices managed and grouped for each use type in a data center that executes power control A power management apparatus comprising: a determination unit that determines a possible server device; and a control unit that stops the server device determined by the determination unit.
(Additional remark 10) It has a monitoring means which monitors the load of the server apparatus which the said control means did not stop, and when the said load exceeds the predetermined 1st threshold value, the said control means is a server apparatus which has stopped The power according to appendix 9, wherein at least one server device capable of reducing the load is activated from among the servers, and the activated server device is stopped when the load falls below a predetermined second threshold value. Management device.
(Additional remark 11) The said monitoring means acquires regularly the number of processes regarding disk input / output waiting for each said server apparatus, and the ratio regarding a memory swap, The said control means is either the said number of processes or the said ratio. The power management apparatus according to appendix 10, wherein the at least one server apparatus is activated when a first threshold value is exceeded.
(Additional remark 12) The said determination means adds the power consumption at the time of the said stop to the multiplication result of the difference value of the maximum power consumption of the said server apparatus and the power consumption at the time of the said server apparatus stop, and CPU usage rate Then, the power consumption per unit of the server device is obtained for each group, and the power consumption per unit is multiplied by the number of the server devices for each group operating in the data center. The power management apparatus according to any one of appendices 9 to 11, wherein the power consumption consumed in the data center is calculated by adding together.
(Additional remark 13) The said determination means utilizes the predetermined | prescribed graph that the increase in the power consumption per 1 unit | set of the said server apparatus is geometric series with respect to the increase in the CPU utilization rate and the CPU utilization rate, The power consumption per unit of the server device is obtained for each group, and the power consumption per unit is multiplied by the number of the server devices for each group operating in the data center. The power management device according to any one of appendices 9 to 11, wherein the power consumption consumed in the data center is calculated.
(Supplementary Note 14) The power consumption after the control unit stops at least one of the server devices when the power consumption calculated by the determination unit exceeds the contract power consumption included in the DR request. 14. The power management apparatus according to appendix 12 or 13, wherein the server apparatus that can be stopped is stopped using the predicted value.
(Supplementary note 15) Any one of Supplementary notes 10 to 14, wherein the monitoring unit excludes an alarm issued based on a stop of the server device from an alarm issued when the server device fails. The power management apparatus according to item 1.

S ADR system EP Electric power company AGR Aggregator DC Data center 100, 200 Node 300 Data center management device 400 Power management device 401 Server information storage unit 402 Server status storage unit 403 Contract information storage unit 404 Server load monitoring unit 405 Stopped server determination unit 406 Server controller

Claims (9)

When a DR request representing a power saving request is received, a server device that can be stopped in each group among a plurality of operating server devices that are managed by being grouped for each use type in a data center that executes power control A decision step to determine,
A first stop step of stopping the server device determined in the determination step;
Power management program that causes a computer to execute. A monitoring step of monitoring a load of the server device that has not been stopped in the first stop step;
An activation step of activating at least one server device capable of reducing the load from among the stopped server devices when the load exceeds a predetermined first threshold;
A second stop step of stopping the server device started in the start step when the load falls below a predetermined second threshold;
The power management program according to claim 1, comprising: The monitoring step periodically obtains the number of processes related to disk input / output waiting for each server device and the ratio related to memory swap,
The starting step starts the at least one server device when any one of the number of processes and the ratio exceeds the first threshold value.
The power management program according to claim 2. The power consumption at the time of stop is added to the multiplication result of the difference between the maximum power consumption of the server device and the power consumption at the time of stopping the server device and the CPU usage rate. Power consumption for each group, and by multiplying the power consumption per unit by the number of server devices for each group operating in the data center, A calculation step for calculating power consumption consumed;
4. The power management program according to claim 1, comprising: Using the CPU usage rate and a predetermined graph in which the increase in power consumption per server device is geometrical with respect to the increase in the CPU usage rate, the power per server device The consumption amount is obtained for each group, and the power consumption per unit is multiplied by the number of the server devices for each group operating in the data center, and the total is consumed at the data center. Calculating step for calculating power consumption,
4. The power management program according to claim 1, comprising: If the power consumption calculated in the calculation step exceeds the contract power consumption included in the DR request, stop using at least one predicted power consumption value after stopping at least one of the server devices. A third stop step of stopping possible server devices;
The power management program according to claim 4 or 5, characterized by comprising: An excluding step of excluding an alarm that is issued based on a stop of the server device from an alarm that is issued when the server device fails;
The power management program according to any one of claims 1 to 6, characterized by comprising: When a DR request representing a power saving request is received, a server device that can be stopped in each group among a plurality of operating server devices that are managed by being grouped for each use type in a data center that executes power control A decision step to determine,
A stopping step of stopping the server device determined in the determining step;
A power management method in which the computer executes. When a DR request representing a power saving request is received, a server device that can be stopped in each group among a plurality of operating server devices that are managed by being grouped for each use type in a data center that executes power control A determination means for determining
Control means for stopping the server device determined by the determination means;
A power management device.

Images