Information processing system and power supply device

Abstract

The present invention comprises: a power facility that supplies power of each phase of a three-phase AC power supply to a plurality of information processing devices disposed in a rack in a server room; and an operation management unit that controls the information processing devices and the power facility. The operation management unit allocates a job to the information processing devices. At least two power supply devices that switch the phase of the power supplied to the information processing devices in order to eliminate a three-phase load imbalance occurring during the operation of the information processing devices are disposed in the rack. The power supply devices include an input unit to which the power of each phase of the three-phase AC power supply is input, and an output unit that is connected to the information processing devices and supplies power. The power phases are circulated so that power phases in which, among the two power supply devices, the power phase of a first output unit of a first power supply device and the power phase of a second output unit disposed at the same position as the first output unit of a second power supply device are different can be selected. This makes it possible to perform power saving management.

Description

Information processing system and power supply device

[0001] The present disclosure relates to an information processing system and a power supply device.

[0002] With the development of information and communications technologies, such as cloud computing, there is a demand for more advanced and diverse information processing devices, which are the infrastructure of the system. Meanwhile, to prevent global warming, reducing not only the power consumed by information processing devices but also the overall power consumption, including the power consumed by the equipment that supplies power to each information processing device and the equipment that cools each information processing device, has become an important issue. For example, in Patent Document 1, when a workload is assigned to each information processing device, the sum of the power consumption of the information processing device, power supply equipment, and air conditioning equipment is calculated for each assignment combination, and all possible assignment combinations for the total number of information processing devices are calculated, and the combination that minimizes total power consumption is extracted. Furthermore, in Patent Document 2, operational information of the information processing device, power supply equipment, and air conditioning equipment is acquired for each hierarchical group, and control is performed to optimize performance relative to power consumption for each hierarchical group.

[0003] Patent Document 1: JP 2009-252056 A, International Publication No. 2011 / 030469

[0004] However, the systems described in Patent Documents 1 and 2 do not implement power saving initiatives led by the power and cooling equipment. Furthermore, neither of these patent documents discloses power saving initiatives that resolve three-phase load imbalances on the equipment side when allocating jobs.

[0005] The present disclosure has been made to solve the above-mentioned problems, and aims to provide an information processing system and a power supply device that saves power by eliminating three-phase load imbalance on the equipment side when allocating jobs.

[0006] The information processing system disclosed herein comprises power equipment that supplies power of each phase of a three-phase AC power supply to a plurality of information processing devices arranged in racks in a server room, and an operation management unit that controls the information processing devices and the power equipment, wherein the operation management unit allocates jobs to the information processing devices, and at least two power supply devices that switch the phase of the power supplied to the information processing devices are arranged in the rack to eliminate three-phase load imbalances that occur when the information processing devices are operating, and the power supply devices have input units that input power of each phase of the three-phase AC power supply and output units that are connected to the information processing devices and supply power, and the power phase is rotated so that the power phase of a first output unit of a first power supply device and the power phase of a second output unit arranged in the same position as the first output unit of a second power supply device can be different from each other.

[0007] According to the information processing system of the present disclosure, power saving can be achieved by eliminating the three-phase load imbalance on the equipment side when allocating jobs.

[0008] 1 is a block configuration diagram of an information processing system according to a first embodiment. FIG. 2 is a diagram illustrating an example of hardware of a first control unit, a second control unit, and a linking unit of the information processing system according to the first embodiment. FIG. 3 is a functional block diagram of the first control unit, the second control unit, and the linking unit of the information processing system according to the first embodiment. FIG. 4 is a diagram illustrating an example of the layout of a server room, power equipment, and cooling equipment in the information processing system according to the first embodiment. FIG. 5 is a flowchart illustrating job allocation in the information processing system according to the first embodiment. FIG. 6 is a flowchart illustrating job allocation in the information processing system according to the first embodiment. FIG. 7 is a diagram illustrating an example of a connection configuration of PSUs in the information processing system according to the first embodiment. FIG. 8 is a diagram illustrating an efficiency curve of a PSU connected to an information processing device constituting the information processing system according to the first embodiment, and a diagram illustrating power loss of the PSU relative to a predicted power consumption value of the information processing device. FIG. 9 is a diagram illustrating imbalance in three-phase loads. FIG. 10 is a diagram illustrating imbalance in three-phase loads. FIG. 11 is a diagram illustrating an example of an operation procedure for switching power to an information processing device in the information processing system according to the first embodiment. FIG. 12 is a diagram illustrating switching of power to an information processing device in the information processing system according to the first embodiment. FIG. 13 is a diagram illustrating an idle state of an information processing device. FIG. 14 is a diagram illustrating the arrangement in a rack of iPDUs connecting information processing devices that constitute the information processing system according to the first embodiment. FIG. 15 is a diagram illustrating an example of the circuit configuration of an iPDU mounted on the information processing system according to the first embodiment. FIG. 16 is a diagram illustrating the connection of multiple iPDUs described in FIG. 16 mounted on the information processing system according to the first embodiment to information processing devices. FIG. 17 is a diagram illustrating an example of the circuit configuration of an iPDU mounted on the information processing system according to the first embodiment. FIG. 18 is a diagram illustrating an example of contacts of an iPDU mounted on the information processing system according to the first embodiment. FIG. 19 is a diagram illustrating the connection of multiple iPDUs described in FIG. 19 mounted on the information processing system according to the first embodiment to information processing devices.

[0009] Hereinafter, preferred embodiments of the information processing system according to the present application will be described with reference to the drawings. Note that the same reference numerals are used to designate the same contents and corresponding parts, and detailed descriptions thereof will be omitted. Similarly, in the following embodiments, redundant descriptions of components with the same reference numerals will be omitted.

[0010] 1 is a block diagram of an information processing system 10 according to embodiment 1. The information processing system 10 includes information processing devices 1-1 to 1-n, power facilities 11a and 11b that supply power to the information processing devices 1-1 to 1-n, cooling facilities 12a and 12b that cool the information processing devices 1-1 to 1-n, and an operation management device 20 that manages the information processing devices 1-1 to 1-n, the power facilities 11a and 11b, and the cooling facilities 12a and 12b.

[0011] Jobs, which are workloads, are assigned to the information processing devices 1-1 to 1-n. A job refers to a set of tasks for which instructions are given to an information processing device, such as "browse the web." One job may be assigned to each of the information processing devices 1-1 to 1-n, multiple jobs may be assigned, or no jobs may be assigned to each of the information processing devices 1-1 to 1-n.

[0012] The operation management device 20 comprises a first control unit 15, a second control unit 16, and a linking unit 19 that links the first control unit 15 and the second control unit 16. The first control unit 15 is connected to the information processing devices 1-1 to 1-n, has location and operation information for the information processing devices 1-1 to 1-n, and calculates the power consumption of each of the information processing devices 1-1 to 1-n. The first control unit 15 also allocates jobs to the information processing devices 1-1 to 1-n. The second control unit 16 acquires operation information for the power equipment 11a, 11b and the cooling equipment 12a, 12b from the power monitoring unit 13, the cooling monitoring unit 14, and the like, and calculates the power loss of the power equipment 11a, 11b and the cooling power of the cooling equipment 12a, 12b based on the predicted power consumption of the information processing devices.

[0013] The information processing devices 1-1 to 1-n, the power equipment 11b, and the cooling equipment 12b are disposed in a server room 30. The power monitoring unit 13 collects power information for each device, such as the individual information processing devices 1-1 to 1-n, as well as power information and operation data for the power equipment, including the transformers, UPS, and cooling equipment, in the server room 30. The cooling monitoring unit 14 collects temperature information and operation data for the cooling equipment in the server room 30. The data from the power monitoring unit 13 and the cooling monitoring unit 14 is transmitted to a first control unit 15 and a second control unit 16 via a general-purpose network (hereinafter, referred to as a general-purpose NW) 17. A management device 18 that displays information from the power monitoring unit 13 and the cooling monitoring unit 14 and information from the first control unit 15 and the second control unit 16 may be connected to the general-purpose NW 17. Furthermore, information processing systems 10a and 10b having a configuration similar to that of the information processing system 10 may be connected to the general-purpose NW 17, and the management device 18 may exchange information between the information processing systems 10, 10a, and 10b, as described below.

[0014] The information processing devices 1-1 to 1-n are server devices, storage devices, network (NW) devices, etc., and the server devices include general-purpose servers, dedicated servers, mainframes, parallel computers, supercomputers, embedded computers, personal computers, etc. The storage devices include magnetic disks, solid-state disks, optical disks, tapes, etc. The NW devices include routers, switches, hubs, etc., and also peripheral devices such as printers, copiers, modems, and displays. The information processing devices 1-1 to 1-n are stored in multiple racks 101.

[0015] The power equipment 11a is installed outside the server room 30 and includes, for example, a transformer, an uninterruptible power supply (UPS), an air circuit breaker (ACB), a molded case circuit breaker (MCCB), a miniature circuit breaker (MCB), a vacuum circuit breaker (VCB), etc. The power equipment 11b is installed inside the server room 30 and includes, for example, a distribution board, a power distribution unit (PDU), a power receiving board, a direct digital controller (DDC), a power supply unit (PSU) mounted on the information processing devices 1-1 to 1-n, etc. However, the classification of the models of the power equipment 11a and 11b is not limited to this. Three-phase power, namely, U-phase, V-phase, and W-phase, is supplied to the information processing devices 1-1 to 1-n via power facilities 11a and 11b.

[0016] The cooling equipment 12a is installed outside the server room 30 and includes, for example, a refrigerator, a cooling tower, an intake and exhaust port, a fan, a duct, refrigerant piping, a liquid cooling device, a power panel, an air handling unit (AHU), etc. The cooling equipment 12b is installed inside the server room 30 and includes, for example, an air handling unit (AHU), an intake and exhaust port, a fan, a duct, refrigerant piping, a direct digital controller (DDC), a variable air volume control device (VAV), a multi-remote control, a chiller, a local cooling device for the front and rear of a fan coil rack row or a rack, etc. However, the classification of the models of the cooling equipment 12a and 12b is not limited to this.

[0017] The power monitoring unit 13 includes a power sensor, a current / voltage sensor, a leakage current sensor, etc., which are installed inside or outside the server room 30. It may also include a branch circuit power meter (BCPM: a distribution board load monitoring device) or an intelligent power distribution unit (iPDU) with a power monitoring function, which is installed in each rack 101. The cooling monitoring unit 14 also includes a temperature sensor, a humidity sensor, a flow rate and flow direction sensor, etc., which are installed inside or outside the server room 30.

[0018] The first control unit 15 is configured as a computer, and the second control unit is configured as a computer or a programmable logic controller (PLC). The control method can be implemented as software such as middleware, an application, embedded control software, or firmware. The cooperation unit 19 cooperates between the control of the first control unit 15 and the control of the second control unit 16. The cooperation unit 19 selects and receives information necessary for the first control unit 15 to perform control, such as operational information and power consumption, from the second control unit 16, and receives information necessary for the second control unit 16 to perform control, such as predicted power consumption of the information processing device, from the first control unit 15. Furthermore, the cooperation unit 19 may construct an interface for transmission and reception between the control units 15 and 16. The cooperation unit 19 may be implemented as a function within the first control unit 15 or the second control unit 16, or may be newly configured as hardware such as a computer and a PLC.

[0019] FIG. 2 shows an example of hardware for the first control unit 15, the second control unit 16, and the link unit 19. The system is composed of a processor 100 and a storage device 200. Although not shown, the storage device includes a volatile storage device such as a random access memory and a non-volatile auxiliary storage device such as a flash memory. A hard disk auxiliary storage device may also be used instead of the flash memory. The processor 100 executes a program input from the storage device 200 to perform the control described below. The program is input from the auxiliary storage device to the processor 100 via the volatile storage device. The processor 100 may output data such as calculation results to the volatile storage device of the storage device 200, or may store the data in the auxiliary storage device via the volatile storage device. Data input / output to / from the processor 100 and the storage device 200 may be performed via wired or wireless communication, and a transceiver may be provided for input / output.

[0020] FIG. 3 is an example of a functional block diagram of the first control unit 15, second control unit 16, and linking unit 19 according to the first embodiment. Means 151 to 154 and 161 to 164 shown in the diagram are examples of functions performed by each control unit. These may be realized as hardware or as a program. The linking unit 19 manages the linking of information between the means 152 and 153 of the first control unit 15 and the means 162 and 163 of the second control unit. FIG. 3 will be described later together with the explanation of FIG. 5.

[0021] Each device in the server room 30, the power equipment 11a, 11b, the cooling equipment 12a, 12b, the power monitoring unit 13, the cooling monitoring unit 14, the first control unit 15, the second control unit 16, and the linking unit 19, are each equipped with a transceiver, and transmit and receive various information, such as control information, power information, and temperature information, to and from the target devices via the general-purpose network 17, thereby executing the power-saving control shown in the flowcharts in Figures 5, 6, 18, and 19 (described below). The management device 18 may be configured, for example, with a computer and a display, and may extract data from multiple information processing systems 10, 10a, and 10b connected to the general-purpose network 17, analyze and display the progress and trends of power saving in each information processing system, and store information on the power and cooling equipment used by each information processing system. When installing or switching equipment, the management device 18 may review the selection of items A to G (described below) and instruct the second control unit 16 to switch from previously selected items to new items to promote power saving.

[0022] FIG. 4 is a diagram showing an example of the connection arrangement of power equipment 11a, 11b and cooling equipment 12a, 12b disposed inside and outside a server room 30. In FIG. 4, the server room 30 is viewed from above, with multiple racks 101 (30 in the figure) disposed inside the server room 30, each housing an information processing device. The racks 101 are connected to PDUs 111 for each set of multiple racks. In FIG. 4, PDUs 1 to 3 are connected to transformers A to C via UPS_A, UPS_B, and UPS_C. As a result, power is supplied from transformers A to C to racks 1 to 30.

[0023] A plurality of air conditioners 121 (three on each side in the figure) are arranged on opposing walls of the server room 30, and can cool the entire rack 101. The air conditioners 121 also require power, and are therefore connected to transformers a to f via UPSs a to f. Furthermore, the air conditioners 121 are connected to refrigerators 122 by refrigerant piping for cooling, and the refrigerators 122 are connected to cooling towers 123 that generate refrigerant. Job allocation will now be described for the operation of the information processing system 10, which is made up of the information processing devices, power facilities 11a and 11b, and cooling facilities 12a and 12b arranged in FIG. 4.

[0024] <Job Allocation> Job allocation to the information processing system will be described with reference to FIGS. 5 and 6 . FIG. 5 illustrates the control flow of the first control unit 15 on the information processing device side and the control flow of the second control unit 16 on the power and cooling equipment side when allocating jobs. The right side of the page illustrates the control flow on the information processing device side, which is controlled by the first control unit 15. The left side of the page illustrates the control flow on the power and cooling equipment side, which is controlled by the second control unit 16. The liaison unit 19 coordinates, for example, the exchange of information related to the arrows passing through the dashed area Y between the "power and cooling equipment side" and the "information processing device side" in FIG. 5 . Specifically, the liaison unit 19 manages information coordination between the first control unit 15 and the second control unit 16 (the arrows from the "power and cooling equipment side" to the "information processing device side" and the arrows from the "information processing device side" to the "power and cooling equipment side" in FIG. 5 ) to ensure smooth information coordination. Figure 6 is a control flow that describes only the flow of job execution, and specifically describes an example of candidate items selected in steps 1-3 of Figure 5. In both Figures 5 and 6, the numbers written on the left indicate the step numbers, and the same numbers in Figures 5 and 6 represent the same steps.

[0025] First, the control flow of the second control unit 16 on the power and cooling equipment side will be described. First, items to be considered when allocating a new job are selected (step 1-1 in FIGS. 5 and 6 ). The items to be considered are based on the following perspectives. Two major items to be considered for power supply equipment include, but are not limited to, (1) improving the power efficiency of the equipment that supplies power to each information processing device, and (2) improving the power efficiency of the equipment that supplies power to the air conditioners required to cool each information processing device. Item (1) power efficiency of the equipment that supplies power to each information processing device includes, for example, (a) transformer efficiency, (b) UPS efficiency, (c) PSU efficiency, and (d) power improvement amount resulting from balancing three-phase loads. Furthermore, item (2) power efficiency of the equipment that supplies power to the air conditioners required to cool each information processing device includes, for example, (f) transformer efficiency that supplies power to the air conditioners and (g) UPS efficiency. Therefore, items (a) to (g) are examples of items to be considered.

[0026] Next, current operational information of the cooling equipment 12a, 12b and the power equipment 11a, 11b having the selected items among the items to be considered is obtained from the power monitoring unit 13 and the cooling monitoring unit 14 (step 1-2 in FIG. 5, acquisition means 161 in FIG. 3). The operational information of the cooling equipment includes, for example, the cooling efficiency curves of air conditioners such as AHUs, the cooling efficiency curves of chillers and cooling towers, and the cooling efficiency within the server room 30. The operational information of the power equipment also includes the conversion efficiency of the UPS and transformer for the U, V, and W phases, three-phase load imbalance loss, wiring loss, etc., and information related to the selected items is obtained from these.

[0027] Next, based on the acquired current operational information, the relationship between the amount of power consumption resulting from power loss of the equipment or device related to the selected item and the amount of power consumption of the information processing device is calculated based on the efficiency curve of the equipment or device (step 1-3 in FIGS. 5 and 6, calculation means 162 in FIG. 3). Note that the amount of power consumption increase may be used instead of the amount of power consumption.

[0028] Meanwhile, the first control unit 15 on the information processing device side acquires current information processing device processing information (step 2-1 in FIG. 5, acquisition unit 151 in FIG. 3). This processing information includes the CPU type, clock frequency, number of cores, communication environment, etc. of each information processing device. Then, when allocating jobs, the minimum requirements are determined for each information processing device's CPU's current workload status (load factor) (including operating, idle, stopped, etc.), the clock frequency, number of cores, and communication environment required for processing (step 2-2 in FIG. 5). Based on these minimum requirements, candidate information processing devices capable of processing the job are selected (step 2-3 in FIG. 5, selection unit 152 in FIG. 3). The candidate information processing devices have similar specifications, for example, network specifications or information processing device specifications. A predicted value of the power consumption of candidate information processing devices is calculated (step 2-4 in FIGS. 5 and 6, predicted value calculation means 153 in FIG. 3), and a candidate facility that can minimize the sum of the power loss of the facility and the power consumption of the information processing device is determined based on the calculated predicted power consumption value (step 1-4 in FIG. 5, determination means 163 in FIG. 3). Job allocation is performed to the information processing devices connected to the determined facility (step 2-5 in FIG. 5, job allocation execution means 154 in FIG. 3).

[0029] When there are multiple candidate information processing devices to be connected to the determined facility candidate, the following three methods can be considered for allocating jobs to the information processing devices. These allocation methods may also be combined (steps 1-5 in FIG. 5): (1) Allocate to the information processing device located in the location with the highest cooling efficiency. (2) Allocate to the information processing device with the highest efficiency (lowest power loss). (3) Allocate randomly.

[0030] In this embodiment, the allocation method (1) described above is used, i.e., the cooling efficiency of the candidate information processing device and the cooling efficiency of the rack in which the candidate information processing device is stored are calculated, and the information processing device to be used for final processing is determined.

[0031] Since the allocation of jobs to the corresponding information processing devices results in changes to the control of the power facilities 11a, 11b and the cooling facilities 12a, 12b, these changes are fed back to the operation information (step 1-6 in FIG. 5).

[0032] <Job Allocation Focusing on the Efficiency of the PSU to Which the Information Processing Device is Connected> The above-mentioned job allocation will be described, for example, focusing on the efficiency of the PSU to which the information processing device is connected. As shown in Figures 7(a) and 7(b), the PSU is a power supply unit that is mounted on the information processing device in the rack 101 and converts AC power to DC power, and losses occur during power conversion. Figure 7(a) shows a configuration without a redundant system, and Figure 7(b) shows a configuration with a redundant system. The operation of supplying power to the information processing device is the same whether or not there is a redundant system.

[0033] The current efficiency of each PSU is calculated from the PSU efficiency curve and load information connected to the PSU, as shown in FIG. 8(a), which are part of the current equipment operation information acquired in step 1-2. In FIG. 8(a), the vertical axis of the graph represents the PSU efficiency, and the horizontal axis represents the load on the PSU. PSUs a to c shown in FIG. 8(a) represent the relationship between the current load and efficiency of PSUs a to c shown in FIG. 7. Although efficiency curves differ for each type of PSU, for convenience of explanation, PSUs a to c shown in FIG. 8(a) are shown as PSUs of the same model, and the relationship between the current efficiency and load is plotted on a single efficiency curve. Note that when different models of PSUs are used, multiple efficiency curves will exist for each model. Because the number of operating information processing devices connected to PSUs a to c differs, the loads are different and the positions plotted on the efficiency curves are also different. Each PSU has an efficiency curve for each of the U, V, and W phases, but for the sake of convenience, Figure 8(a) shows the efficiency curve for only the U phase. However, the V and W phases can also be calculated in the same way as above.

[0034] From the graph in Fig. 8(a), a graph of power consumption that takes into account the equipment efficiency of the PSU with respect to the predicted power consumption of the information processing device can be drawn as shown in Fig. 8(b). The horizontal axis of the graph in Fig. 8(b) represents the power consumption value p IT The vertical axis represents the power consumption value p IT The sum of the power loss of the PSU and ploss_PSUIT is P PSUIT (PPSUIT = p IT + ploss_PSUIT). And the current P PSUIT is shown on the vertical axis. If there is no power loss ploss_PSUIT due to the PSU conversion loss, PSUIT = p IT and is located on the solid line extending from the bottom left to the top right of the graph. PSUIT is operating at a low load where the equipment efficiency is poor, and the PSU conversion loss is large in the area where the power consumption of the information processing equipment is low, PSUIT On the other hand, as shown in FIG. 15(a), PSU_b and PSU_c have a higher conversion efficiency than PSU_a, so P PSUIT is the P of PSU_a PSUIT The power consumption value p of the information processing device is smaller than IT As the conversion efficiency of PSU increases, P PSUIT = p IT After approaching the solid line of P, the efficiency gradually decreases. PSUIT = p IT It moves away from the solid line.

[0035] The predicted power consumption value of the candidate information processing device is input from the first control unit 15. This value is shown by a bold line in FIG. 8B. In this predicted power consumption value, the PSU power loss is PSUIT Therefore, the processing of the new job is allocated to a candidate information processing apparatus connected to PSU_b.

[0036] <Switching Power to Information Processing Devices to Improve Three-Phase Load Balancing> As described above, jobs are sequentially assigned to information processing devices. In an information processing system, power from a three-phase power source (U, V, and W phases) connected in star or delta configuration is supplied to information processing devices 1-1 through 1-n via power equipment 11a and 11b. The supplied power powers information processing devices 1-1 through 1-n, enabling them to operate and process jobs. However, power consumption differs depending on the operating state of each information processing device connected to the U, V, and W phases. In the star connection shown in FIG. 9(a), a neutral current flows from the information processing devices connected to the U, V, and W phases to the neutral conductor. In the delta connection shown in FIG. 9(b), a circulating current flows in the direction indicated by the arrow. Such currents cause power loss due to cable resistance and circuit breakers installed in the circuit. Furthermore, the unbalanced power consumption state shown in FIG. 10 reduces the efficiency of equipment such as UPSs, increasing power loss.

[0037] To improve this unbalanced state, attention is focused on switching power to an information processing device that the user has set as switchable. An example of the operational procedure for power switching control is shown in FIG. 11 . Each procedure will be described in detail below. Note that the procedure is an example, and the order of the procedures may be reversed. The switching control may be performed by the second control unit 16 or by a PLC newly installed in the operation management device 20. When a PLC is installed, it is necessary that it be able to process or instruct each step of the control shown in FIG. 11 , obtain power information similar to that obtained by the second control unit, and transmit information about the information processing device to be switched to directly or via the link unit 19 to the first control unit 15.

[0038] First, the power consumptions Pu, Pv, and Pw of the U, V, and W phases are obtained from the power monitoring unit 13 or the second control unit 16 (step 3-1 in FIG. 11). Then, as shown in FIG. 12(a), the average value Pave of the obtained power consumptions Pu, Pv, and Pw is calculated using the following formula (step 3-2 in FIG. 11).

[0039] Then, the difference δ of each phase with respect to the average value Pave is calculated as follows:

[0040] Of the differences δ, the U phase and V phase, for which δu>0 and δv>0, are determined as the phases to be switched from (see step 3-2 in FIG. 11, FIG. 12(b)).

[0041] Next, the current CPU load rate or power information of each information processing device of the U-phase and V-phase from which switching is to be performed is obtained from the first control unit 15 or the power monitoring unit 13 (step 3-3 in FIG. 11). Then, from among the information processing devices of each phase, information processing devices that are switchable are set (step 3-4 in FIG. 11). The switchable information processing devices may be determined in advance by specifying the IP address of a physical server, or may be determined based on the type of job being processed. The maximum switchable power amount Pmove of the set switchable information processing devices is calculated for each phase (step 3-5 in FIG. 11).

[0042] As shown in FIG. 12(b), when the maximum switchable power amount Pmove of each phase exceeds the differences δu and δv, the differences δu and δv are determined as the switching amounts (step 3-6 in FIG. 11), and the second control unit 16 outputs switching commands sequentially to switch the phase to be switched to, with the W phase having a negative difference δw from the average value Pave (step 3-7 in FIG. 11).

[0043] On the other hand, as shown in FIG. 13(b), when the maximum switchable power amount Pmove is smaller than the difference δv, Pmove is determined as the switching amount (step 3-6 in FIG. 11), and the second control unit 16 outputs switching commands sequentially at the following ratio to switch the U phase and W phase, whose differences δu and δw from the average value Pave are negative (step 3-7 in FIG. 11).

[0044]

[0045] Therefore, in the case of FIG. 13(b), the power allocated to the U phase is Pmove×(a / a+b), and the power allocated to the V phase is Pmove×(b / a+b) (see FIG. 13(c)).

[0046] After the switching, the power monitoring unit 13 checks the power waveform to ensure there are no abnormalities, such as an extremely high current value or a halt in current flow. The hardware required for switching is described below. Because inrush current may occur during switching, the switching interval is set to avoid this effect. Furthermore, if a large number of switching operations are performed at once, the inrush current may increase, potentially tripping a circuit breaker connected to the circuit. Furthermore, if an inrush current exceeds the capacity of the iPDU, the circuit breaker within the iPDU may be tripped. Therefore, the number of switching operations is set taking these effects into consideration. If an abnormality occurs, it is detected by the power monitoring unit 13, and the second control unit 16 issues an instruction to the power equipment 11a, 11b to stop the supply of power. Completion of the switching is confirmed based on the power consumption of each information processing device (step 3-8 in FIG. 11 ), and the second control unit 16 updates the connection phase information of the switched information processing device (step 3-9 in FIG. 11 ).

[0047] Information processing devices set as switchable also include information processing devices in an idle state (load rate 0%). This is because power is consumed even in an idle state, as shown in FIG. 14. The advantage of switching an information processing device in an idle state is that energy saving can be achieved with high reliability without affecting information processing devices that are operating for processing. Whether an information processing device is in an idle state is determined by determining, from the CPU load rate among the current processing information of the information processing device acquired in step 2-1 of FIG. 5 described above, an information processing device with a load rate of 0% as shown in FIG. 14 as being in an idle state. Power is consumed even at a load rate of 0%, and the power consumption relative to the load rate differs for each information processing device, so it is necessary to obtain the characteristics of each device.

[0048] If the CPU load factor cannot be acquired, an information processing device that has no fluctuation in power consumption for a certain period of time at a level close to the idle state is determined to be in the idle state based on the power consumption of the information processing device, for example, based on the following formula: ０ ≦t≦t ０ +t ｄ In this case, P ｉｄｌｅ +ΔP ｄ ≧P ｓｅｒｖｅｒ t: a fixed time t ０: Idle state information processing determination start time t ｄ : Determination time of information processing device in idle state P ｉｄｌｅ : Power consumption in idle state P ｓｅｒｖｅｒ : Power consumption of the information processing device at a certain time t ΔP ｄ : Maximum power fluctuation range for determining power consumption in idle state

[0049] <Hardware for Switching> Next, the hardware for power switching will be described. Power switching is performed by the iPDU 141, which has already been described in FIG. 7. FIG. 15 is a diagram illustrating the arrangement of the iPDU within the rack 101. FIG. 15(a) is a diagram illustrating an example of the arrangement of the rack 101, power equipment 11b, and cooling equipment 12b within the server room 30. FIG. 15(b) is a diagram illustrating the internal structure of one rack 101. FIG. 15(c) is a perspective view of the iPDU 141 disposed in the rack 101. The iPDU 141 is typically mounted within the rack 101 and is available in outlet bar types and 1U types that resemble a server in appearance. As shown in FIG. 15, two or more outlet bar types (iPDU-1 and iPDU-2 in FIG. 15(b)) are often arranged within the rack for redundancy. When redundancy is required, the power supplies connected to the iPDU-1 and iPDU-2 may be separate or the same system. In this embodiment, power is distributed from the PDU 111 to the iPDU 141 using a star connection or a delta connection. The iPDU-1 and iPDU-2 each have an outlet (power socket) 141p, which receives power from the PDU 111 at the inlet 141i and supplies power from the outlet 141p to each of the information processing devices 1-1 to 1-n housed in the rack 101. Power information such as current and voltage of the information processing devices connected to the iPDU-1 and iPDU-2 can also be acquired. Furthermore, the power supply to the outlet 141p can be controlled individually or as a group. The information processing device 1-1 shown in FIG. 15(b) is a dual-cord server, and its power supply operates using power distributed via the outlets 141p of the iPDU-1 and iPDU-2.

[0050] <iPDU Circuit Configuration 1> Figure 16 shows an example of the circuit configuration of the iPDU-1. The power source (PDU 111) is connected to the iPDU-1 in a star connection. U-phase, V-phase, W-phase, and neutral wire N are drawn into the terminal block 141d of the inlet 141i. Single-phase power with line voltages (UN, VN, WN) is output from these to each outlet. The output is controlled on and off by a relay 141r in response to a switching command from the second control unit 16. The relay 141r is configured with a 1a or 1b contact that can be switched on and off only. In a delta connection, single-phase power with phase-to-phase voltages (U, V, W) is output, but the other components are the same. With this iPDU-1 configuration, when switching power from the U-phase or V-phase to the W-phase, the U-phase or V-phase relay 141r is disconnected (off) and the W-phase relay 141r is connected (on). However, with this circuit configuration, only one phase of power can be output from one outlet. Therefore, to switch from U phase to W phase, or from V phase to W phase, or from V phase to W phase, or from V phase to U phase, as described in Figures 12 and 13, it is necessary to connect information processing devices to at least three outlets. In Figure 16, four outlets are arranged to form one module. In this case, three modules are required.

[0051] FIG. 17 is a diagram illustrating a configuration example in which the iPDU-1 and iPDU-2 configured as shown in FIG. 16 are arranged in a rack 101. The iPDU-1 and iPDU-2 are connected to information processing devices 1-1 to 1-n in the rack 101, and power is supplied from both the iPDU-1 and iPDU-2 to one information processing device. In FIG. 17, the iPDU-1 outputs single-phase power with line voltages UN, VN, and WN, respectively, as a module consisting of four outlets 141p, in order from top to bottom. Meanwhile, the iPDU-2 rotates the line power phase at the outlet 141p by changing the wiring of the U, V, and W phases in the terminal block 141d. In other words, the power phase is rotated so that the power phase of the output section of the iPDU-1 and the power phase of the output section of the iPDU-2 located in the same position as the output section of the iPDU-1 can be different. As a result, single-phase power with line voltages VN, WN, and UN is output from the outlets 141p in a module-by-module fashion, starting from the top. This configuration facilitates the connection of the outlets 141p that supply different phases of power to the iPDU-1 and iPDU-2 when the information processing devices 1-1 to 1-n are connected to the outlets 141p in a module-by-module fashion from the top. Furthermore, the three-phase load is more easily balanced than when phase rotation is not used, reducing power loss due to three-phase load imbalance. The modularization and module unit determination methods depend on the user's needs. Furthermore, if a malfunction occurs in either iPDU-1 or iPDU-2, the remaining iPDU supplies power to the information processing device, thereby providing redundancy, as shown in FIG. 7(b).

[0052] <iPDU Circuit Configuration 2> Figure 18 shows another example of the circuit configuration of the iPDU-1. The power source (PDU 111) is connected to the iPDU-1 in a star connection. U-phase, V-phase, W-phase, and neutral line N are drawn into the terminal block 141d of the inlet 141i. Single-phase power, representing line voltages (UN, VN, WN), is output from these to each outlet. The relay 141s has a contact 1c for switching between two line voltages and a contact 1a or 1b for switching the electrical circuit on and off, connected in series. This allows switching contact 1c in response to a switching command from the second control unit 16 to switch between line voltages UN and VN, VN and WN, or WN and UN, making it possible to select two voltages of different phases from a single outlet 141p. Furthermore, the electrical circuit can be connected or disconnected by turning contact 1a or 1b on or off.

[0053] In the delta connection, the circuit configuration is as shown in Fig. 19, and the relay 141t has a contact 1c for switching between two interphase voltages and a contact 1a for switching the electric circuit on and off, connected in series. This makes it possible to switch between interphase voltages U and V, V and W, and W and U by switching contact 1c in response to a switching command from the second control unit 16, making it possible to select two voltages of different phases from one outlet. Furthermore, the electric circuit can be connected or disconnected by turning contact 1a or 1b on or off.

[0054] In the circuit configurations shown in Figures 18 and 19, a data processing device only needs to be connected to one outlet to switch from U phase to V phase, and two outlets only need to be connected to switch from U phase to V phase or W phase. While Figures 18 and 19 show a module in which multiple outlets are grouped together to form a source and destination phase, a predetermined number of modules may be configured for each outlet. When connecting data processing devices to modules, from the perspective of three-phase load balancing, it is preferable to connect them evenly to all modules rather than biased to a specific module. Furthermore, relay 141g may be configured using two 1a or 1b contacts (2a and 2b contacts) instead of the 1c contact shown in Figure 20 (2a and 2b contacts). However, this has disadvantages, such as an increased number of parts and the possibility of a phase-to-phase short circuit occurring if one a or b contact is welded and a switching command is input to the other. However, using the 2a or 2b contacts offers cost advantages.

[0055] 21 is a diagram illustrating a configuration example in which the iPDU-1 and iPDU-2 configured as shown in FIG. 18 are arranged in a rack 101. The iPDU-1 and iPDU-2 are connected to information processing devices 1-1 to 1-n in the rack 101, and power is supplied from both the iPDU-1 and iPDU-2 to one information processing device. The iPDU-1 outputs power from multiple phases selectable from line voltages UN or VN, VN or WN, and WN or UN, starting from the top of the outlet 141p. Meanwhile, the iPDU-2 is connected by changing the wiring of the U, V, and W phases in the terminal block 141i, thereby circulating the line power phase at the outlet 141p. In other words, the power phase is circulated so that the power phase of the output section of the iPDU-1 and the power phase of the output section of the iPDU-2 located in the same position as the output section of the iPDU-1 can select different power phases. This allows the outlets 141p to output power from multiple selectable phases, i.e., line voltages VN or WN, WN or UN, and UN or VN, starting from the top. This configuration facilitates the connection of outlets 141p that supply different phases of power to iPDU-1 and iPDU-2 when connecting information processing devices to the outlets 141p in order. This configuration facilitates balancing of three-phase loads compared to when power is not circulated, particularly when multiple phases are drawn. This reduces power loss due to three-phase load imbalance. Furthermore, the increased flexibility in phase switching allows for more effective balancing of three-phase loads than the configuration shown in FIG. 17. Furthermore, if a malfunction occurs in either iPDU-1 or iPDU-2, the remaining iPDU supplies power to the information processing device, providing redundancy similar to that shown in FIG. 7(b).

[0056] Although exemplary embodiments are described in this disclosure, the various features, aspects, and functions described in the embodiments are not limited to the application of a particular embodiment, but may be applied to the embodiments alone or in various combinations. Therefore, countless variations not illustrated are anticipated within the scope of the technology disclosed in this specification. For example, variations in, addition to, or omission of at least one component are included.

[0057] 1-1 to 1-n: information processing device, 10: information processing system, 11a, 11b: power equipment, 12a, 12b: cooling equipment, 13: power monitoring unit, 14: cooling monitoring unit, 15: first control unit, 16: second control unit, 17: general-purpose network, 18: management device, 19: collaboration unit, 30: server room, 100: processor, 101: rack, 111: PDU, 121: air conditioner, 122: refrigerator, 123: cooling tower, 141: iPDU, 200: storage device.

Claims

1. An information processing system comprising: power equipment that supplies power of each phase of a three-phase AC power supply to a plurality of information processing devices arranged in a rack in a server room; and an operation management unit that controls the information processing devices and the power equipment, wherein the operation management unit allocates jobs to the information processing devices, and at least two power supply devices that switch the phase of the power supplied to the information processing devices to eliminate three-phase load imbalance that occurs when the information processing devices are operating are arranged in the rack, and the power supply devices have an input unit that inputs power of each phase of the three-phase AC power supply and an output unit that is connected to the information processing devices and supplies power, and the power phase is rotated so that a different power phase can be selected between a first output unit of a first power supply device and a second output unit arranged in the same position as the first output unit of a second power supply device.

2. An information processing system according to claim 1, characterized in that a switching contact is connected between said input section and said output section for supplying or cutting off the phase power inputted from said input section from said output section.

3. The information processing system according to claim 1, characterized in that the input section is supplied with two different phases of power, and a switching contact for switching the two phases of power is connected between the input section and the output section.

4. An information processing system according to claim 3, further comprising a contact connected in series with said switching contact for supplying or cutting off power to said output section.

5. An information processing system as described in any one of claims 1 to 4, comprising an operation management unit having a first control unit that controls the information processing device and a second control unit that controls the power equipment, wherein the first control unit selects candidate information processing devices that satisfy processing conditions for allocating a job, and the second control unit calculates the amount of power loss of the power equipment relative to a predicted value of power consumption of the candidate information processing device from operation information of the power equipment in operation, and determines, from among the selected candidate information processing devices, an information processing device connected to power equipment that can minimize the sum of the amount of power loss and the power consumption of the selected information processing device as the information processing device to which the job is allocated.

6. A power supply device configured to be arranged in a rack that houses information processing equipment, having an input section that inputs two-phase power of the three-phase AC power supplied to the rack, an output section that is connected to the information processing equipment, and a switching contact between the input section and the output section that switches between the two-phase power.

7. A power supply device according to claim 6, further comprising a contact connected in series with said switching contact for supplying or cutting off power to said output section.

Images