Power efficiency calculation device, power efficiency calculation method, power efficiency calculation system, and program

Abstract

Provided is a power efficiency calculation device (1) comprising: a metrics collection unit (101) that acquires metrics measurements including a task execution volume and power consumption from a physical server group (30); a metrics measurement imputation determination unit (102) that determines whether the amount of change in metrics measurements is equal to or larger than a first specified value to determine a specific imputation method; a power efficiency calculation frame imputation determination unit (103) that determines whether the value of a deviation between a power efficiency calculation time frame and a metrics acquisition interval is equal to or larger than a second specified value to determine a specific imputation method; a metrics imputation unit (104) that performs imputation by using the determined specific imputation methods; and a power efficiency calculation unit (105) that calculates a value of integration over the power efficiency calculation time frame to calculate a power efficiency.

Description

[Technical Field]

[0001] The present invention relates to a power efficiency calculation device, a power efficiency calculation method, a power efficiency calculation system, and a program for calculating power efficiency for application processing. [Background technology]

[0002] To date, various definitions have been proposed for data center power efficiency indicators. Among them, the only indicator that can measure and calculate in real time how power-efficiently a specific application is being operated is DCeP (Data Center energy Productivity) (see Non-Patent Document 1). DCeP is defined by the following formula (1):

[0003]

number

[0004] DCeP (Power Efficiency) is defined as "Useful Work Produced" (referred to herein as "Work Output") divided by "Total Energy Consumed to Perform that Work" (referred to herein as "Power Consumption"). In addition, "Useful Work Produced" is defined by the following formula (2).

[0005]

number

[0006] where "M" is the number of tasks started during the evaluation window, "V i ” is the normalization coefficient obtained by summing the task values, “U i“(t,T)” is the time-based utility function for each task, “t” is the elapsed time from the start to the completion of the task, and “T” is the absolute time of task completion.

[0007] Furthermore, Non-Patent Document 2 defines "Useful Work Produced" (amount of work produced) as the energy consumed when various applications are processed in an HPC (High Performance Computing) data center, and describes the results of calculating DCeP (power efficiency). However, Non-Patent Document 2 does not go into the details of the power calculation (amount of power consumption), and only reports the calculation results.

[0008] Furthermore, Non-Patent Document 3 defines "Useful Work Produced" as the number of times two types of applications are executed within a certain period of time (Useful Computational Units), and normalizes the number of times each type is executed using a weighting coefficient (1:0.08). However, Non-Patent Document 3 also does not go into details about power calculations (amount of power consumption), and only reports the calculation results.

[0009] It is assumed that the actual measured value of "Total Energy Consumed to Perform that Work" (power consumption) will be obtained by using metrics collection software (resource monitoring software) to obtain metrics, which are evaluation indicators of performance, etc., from the system ("physical server group 30" described below). Prometheus, one of the de facto standards for metrics collection software, can obtain power [W] by using an IPMI (Intelligent Platform Management Interface) exporter. However, because this power (power consumption) is an instantaneous value, it is necessary to calculate the integral value to obtain the power consumption [Wh or J]. Similarly, it is assumed that metrics that require the calculation of an integral value will be used for "Useful Work Produced" (amount of work results). [Prior art documents] [Non-patent literature]

[0010] [Non-Patent Document 1] D. Anderson, et al., “A framework for data center energy productivity.” The Green Grid, 2008. [Non-patent document 2] A. Grishina, et al., “DC energy data measurement and analysis for productivity and waste energy assessment,” 2018 IEEE International Conference on Computational Science and Engineering (CSE), IEEE, 2018. [Non-patent document 3] Landon H. Sego, et al., “Implementing the data center energy productivity metric,” ACM Journal on Emerging Technologies in Computing Systems (JETC) 8.4 (2012): 1-22. [Non-patent document 4] “prometheus-community / ipmi_exporter,” GitHub, Inc., [online], [Retrieved May 9, 2022], Internet <URL:https: / / github.com / prometheus-community / ipmi_exporter / blob / master / docs / metrics.md#power-consumption> Summary of the Invention [Problem to be solved by the invention]

[0011] When calculating an integral value such as the amount of power consumption, if the following two points are not handled properly, a large discrepancy will occur between the actual value and the calculated result. (1) Complementation between metric measurements (2) Correction (complementation) of the discrepancy between the time frame for power efficiency calculation (Assessment Window) and the metric acquisition interval These two points are bound to occur with metrics collection tools like Prometheus.

[0012] FIG. 7 is a diagram showing metrics (power [W] or number of accesses) data collected from a system ("physical server group 30" described later) in time series (horizontal axis: time [s]). In the example of Figure 7, the amount of power (energy consumption) to be calculated is the Assessment Window period, which is the calculation period for DCeP (power efficiency). However, the metrics collection tool only holds data for the shaded area, and cannot calculate the areas before and after the shaded area (referenced P1 and P2). Therefore, it is necessary to correct the discrepancy between the time frame for power efficiency calculation (Assessment Window) and the metrics acquisition interval. In addition, it is necessary to decide whether to use linear interpolation between the data in the shaded areas (instantaneous power values ​​indicated by circles), i.e., between the metric measurement values, as shown in Figure 7, or other interpolation methods.

[0013] In this way, without (1) interpolating between metric measurements and (2) correcting (interpolating) the discrepancy between the time frame for power efficiency calculation (Assessment Window) and the metric acquisition interval, it is impossible to accurately calculate integrated values ​​such as power consumption.Furthermore, DCeP (power efficiency), which is calculated using power consumption, cannot be calculated accurately because power consumption cannot be calculated accurately.

[0014] The present invention has been made in view of the above points, and an object of the present invention is to accurately calculate power efficiency by using the interpolated integral value. [Means for solving the problem]

[0015] The power efficiency calculation device according to the present invention is a power efficiency calculation device that calculates the power efficiency resulting from the execution of an application installed on a group of physical servers, and includes a metrics collection unit that acquires, from the group of physical servers, metrics measurements that are evaluation indices necessary for power efficiency calculation and that include task execution amounts of the applications and power consumption of the group of physical servers; a metrics measurement value complementation determination unit that determines whether or not complementation between metrics measurements is necessary depending on whether or not a change in the metrics measurement value is equal to or greater than a first specified value, and determines a predetermined complementation method depending on whether or not complementation is necessary; and a power efficiency calculation unit that determines whether or not a difference between a time frame for power efficiency calculation and a metrics acquisition interval is equal to or greater than a second specified value. a power efficiency calculation frame complementation determination unit that determines whether or not complementation is required for the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval, and determines a predetermined complementation method depending on whether complementation is required; a metrics complementation unit that complements between the metrics measurement values ​​using the predetermined complementation method determined by the metrics measurement value complementation determination unit, and complements the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval using the predetermined complementation method determined by the power efficiency calculation frame complementation determination unit; and a power efficiency calculation unit that calculates the power efficiency by calculating an integral value over the time frame for power efficiency calculation for the task execution amount and the power consumption of the group of physical servers, which are the metrics measurement values ​​complemented by the metrics complementation unit. [Effects of the Invention]

[0016] According to the present invention, the interpolated integral value can be used to calculate the accurate power efficiency. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a diagram showing the overall configuration of a power efficiency calculation system including a power efficiency calculation device according to an embodiment of the present invention. [Figure 2] 10 is a diagram for explaining an error "α" associated with a first specified value according to the present embodiment. FIG. [Figure 3]10 is a flowchart showing the flow of processing executed by the power efficiency calculation device according to the present embodiment. [Figure 4] FIG. 10 is a diagram illustrating an example in which the data transfer volume of UPF is collected as a metric measurement value. [Figure 5] FIG. 10 is a diagram illustrating an example in which power is collected as a metric measurement. [Figure 6] FIG. 2 is a hardware configuration diagram illustrating an example of a computer that realizes the functions of the power efficiency calculation device according to the present embodiment. [Figure 7] This is a diagram showing chronological data of metrics measurements (power consumption or number of accesses) collected from the system. DETAILED DESCRIPTION OF THE INVENTION

[0018] Next, an embodiment of the present invention (hereinafter referred to as "the present embodiment") will be described. FIG. 1 is a diagram showing the overall configuration of a power efficiency calculation system 1000 including a power efficiency calculation device 1 according to this embodiment. The power efficiency calculation system 1000 includes a group of physical servers 30 configured by a data center or the like, and a power efficiency calculation apparatus 1 connected to the group of physical servers 30 for communication.

[0019] The physical server group 30 is operated by, for example, building a virtualization infrastructure on a physical server, and one or more applications 3 are installed in a virtual machine (VM) or container on a virtual OS to execute processing. Each application 3 realizes a service by executing one or more tasks.

[0020] The power efficiency calculation device 1 collects metrics (such as "task execution amount" and "power consumption," which will be described later) that are evaluation indices required for calculating DCeP (power efficiency) from the physical server group 30, determines whether interpolation between metric measurement values ​​is necessary, and whether correction (interpolation) of the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval is necessary, and determines the interpolation method for each. Then, the power efficiency calculation device 1 calculates DCeP (power efficiency) after performing the necessary interpolation.

[0021] As a result, the power efficiency calculation device 1 determines whether or not it is necessary to interpolate between the metrics measurement values ​​and to correct (interpolate) the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval, and then calculates an integral value using the data of the interpolated metrics measurement values, thereby enabling accurate calculation of power efficiency. Here, by determining whether or not interpolation is necessary, the power efficiency calculation device 1 does not perform linear interpolation or polynomial interpolation, which will be described later, and can reduce overhead due to unnecessary calculations. Furthermore, by readjusting this complementation process at predetermined time intervals, the power efficiency calculation device 1 can determine whether complementation is necessary and perform calculation of DCeP (power efficiency) according to the amount of change over time in the metrics measurement values.

[0022] Next, the power efficiency calculation device 1 will be described in detail. As shown in FIG. 1, the power efficiency calculation device 1 includes a control unit 10, an input / output unit 11, and a storage unit 12.

[0023] The input / output unit 11 inputs and outputs information to and from each server of the physical server group 30. The input / output unit 11 is composed of a communication interface that transmits and receives information via a communication line, and an input / output interface that inputs and outputs information to and from an input device such as a keyboard and an output device such as a monitor (not shown).

[0024] The storage unit 12 is configured by a hard disk, a flash memory, a RAM (Random Access Memory), or the like. The storage unit 12 temporarily stores programs for executing the functions of the control unit 10 and information necessary for the processing of the control unit 10. Furthermore, the storage unit 12 also stores metrics required for calculating DCeP (power efficiency) collected from each physical server in the physical server group 30, an OS (virtual OS), a VM / container, an application, etc. in a metrics collection database (DB: DataBase) 100.

[0025] The control unit 10 is responsible for all the processing executed by the power efficiency calculation device 1, and as shown in FIG. 1, includes a metrics collection unit 101, a metrics measurement value complementation determination unit 102, a power efficiency calculation frame complementation determination unit 103, a metrics complementation unit 104, and a power efficiency calculation unit 105.

[0026] The metrics collection unit 101 acquires metrics (evaluation indexes for performance, etc.) required for calculating power efficiency from the physical server group 30 and stores them in the metrics collection DB 100. The metrics collector 101 collects metrics from the physical servers, OSs (virtual OSs), VMs / containers, applications, etc. that make up the physical server group 30 using existing metrics collection software (for example, Prometheus, etc.). For example, the metrics collector 101 collects information on task execution volume (number of requests processed by applications, amount of data transferred, etc.) and power consumption [W] as metrics measurement values ​​from the physical server group 30 and stores it in the metrics collection DB 100. The metrics collector 101 collects metrics from the physical server group 30 in advance of calculating DCeP (power efficiency) ("preparation stage" described below). The metrics collector 101 also collects metrics when calculating DCeP (power efficiency) ("operation stage" described below) and when readjusting the complementation method ("readjustment stage" described below).

[0027] Furthermore, in the advance preparation stage and the readjustment stage, the metrics collection unit acquires the measured value (power) of the power consumption of the entire physical server group 30 from a power meter (not shown) provided in the physical server group 30.

[0028] The metrics measurement value complementation determination unit 102 determines whether or not to complement between metrics measurement values ​​depending on changes in the metrics measurement values ​​(discrete values) collected by the metrics collection unit 101, and if it determines that complementation should be performed, it determines the complementation method (linear complementation, polynomial complementation, etc.). If the change in the discrete value is smaller than a specified value (first specified value) for interpolation between metrics measurement values, the metrics measurement value interpolation determination unit 102 determines that the change is small and the deviation of the integral value is small, and therefore determines not to perform interpolation.On the other hand, if the change in the discrete value is equal to or greater than the first specified value, the metrics measurement value interpolation determination unit 102 determines that the deviation of the integral value will be large, and therefore determines to perform interpolation. This first specified value is readjusted by comparing it with the measured value of the wattmeter at predetermined intervals. This keeps the deviation of the integrated value below a certain percentage. Details will be described later.

[0029] Specifically, the metrics measurement value complementation determination unit 102 acquires metrics measurement values ​​for a predetermined period from the metrics collection DB 100. Then, the metrics measurement value complementation determination unit 102 calculates the magnitude D of the time difference of the metrics measurement values, as shown in the following equation (3). Here, x i indicates the i-th metric measurement value (discrete value), and N is the number of measurement timings (number of measurements) in a given period.

[0030]

number

[0031] The metrics measurement value complementation determination unit 102 may use the following formula (4) instead of formula (3). In formula (3), by using a square root, it is possible to evaluate the future frequency of error occurrence due to load fluctuations without omission. On the other hand, by allowing formula (4) to take positive and negative values, it is possible to take into account the cancellation of overshooting and undershooting of errors. The metrics measurement value complementation determination unit 102 is set in advance to use either formula (3) or formula (4).

[0032]

number

[0033] Then, the metrics measurement value complementation determination unit 102 determines to perform complementation when the magnitude D of the time difference is equal to or greater than a first specified value shown in the following equation (5). Then, the metrics measurement value complementation determination unit 102 determines, for example, linear complementation or polynomial complementation as the complementation method. Linear interpolation is a method of interpolating metric measurements by linearly connecting a value at a given time with the value at the next time point. Polynomial interpolation is a method of interpolating metric measurements by polynomial approximation using multiple past values. When the metrics measurement value complementation determination unit 102 determines that complementation should be performed, it may preset the complementation method to be used as either linear complementation or polynomial complementation, or it may determine the complementation method using logic such as setting a threshold value according to the magnitude D of the time difference, and adopting polynomial complementation if the magnitude is equal to or greater than the predetermined threshold, and adopting linear complementation if the magnitude is less than the predetermined threshold.

[0034] Furthermore, the metrics measurement value complement determination unit 102 determines not to perform complementation if the magnitude D of the time difference is less than a first specified value shown in the following equation (5). If it is determined that no interpolation is to be performed, the metrics measurement value interpolation determination unit 102 interpolates the metrics measurement value by assuming that it will remain constant (constant value) from the previous measurement point until the next measurement timing (constant value interpolation). That is, the metrics measurement value complement determination unit 102 determines whether or not complementation between metrics measurement values ​​is necessary, and then determines a predetermined complement method (linear complement, polynomial complement, constant value complement, etc.).

[0035] The metrics measurement value complement determination unit 102 determines to perform complementation (linear complementation, polynomial complementation, etc.) when the magnitude D of the time difference is equal to or greater than a first specified value shown in the following equation (5). Note that "α" is an error value set so as to satisfy a preset allowable error rate, and details will be described later.

[0036]

number

[0037] If the magnitude of the time difference D is equal to or greater than a first specified value, the metrics measurement value complementation determination unit 102 may instruct the metrics collection unit 101 to narrow the interval between metrics measurements. This makes it possible to reduce the magnitude of the deviation between metrics measurement values.

[0038] <Method for determining the error "α" related to the first specified value> Here, the error "α" relating to the first specified value will be explained. FIG. 2 is a diagram for explaining the error "α" associated with the first specified value. Reference numeral 21 in FIG. 2 shows a time series of power values ​​of the physical server group 30 that are the subject of DCeP (power efficiency) calculation. Here, the power value indicated by W1 is the power value of the physical server group 30 measured by a power meter. The power value indicated by W2 is a value calculated based on the power metric measurement values ​​using linear interpolation or polynomial interpolation to interpolate between the metric measurement values. Note that the following description will be given assuming that the calculation is performed using linear interpolation. The power value indicated by W3 is a value calculated without interpolating the metric measurement values, i.e., calculated using the constant value interpolation described above.

[0039] Here, the value P calculated by linear interpolation of the power metric measurement values ​​is linear and the value P obtained from the power meter measureand the error α base =|1-P linear / P measure Then, the error tolerance that determines whether or not the interpolation is necessary is determined, and the error "α" is defined as shown in the following equation (6).

[0040]

number

[0041] As shown by reference numeral 22 in FIG. 2, the error (α base ) and calculate the calculated error (α base The error "α" is determined as a value that decreases as the difference between the predetermined error tolerance and the error "α" decreases. Then, the first specified value is set using this error "α" (see equation (5)). In equation (6), the error "α" is determined by calculating the percentage of the integral value of constant value interpolation calculated without interpolating the metric measurement value (the total of the triangles that are added or subtracted when linearly interpolated).

[0042] Returning to FIG. 1, the power efficiency calculation window complementation determination unit 103 determines whether to correct (complement) the discrepancy between the time window (Assessment Window) for power efficiency calculation and the metrics acquisition interval, and if it is determined that complementation should be performed, determines the complementation method (linear complementation, polynomial complementation, etc.).

[0043] When the start time and end time of the time frame (Assessment Window) for power efficiency calculation do not match the metrics acquisition timing, the power efficiency calculation window complementation determination unit 103 determines whether the difference between the start and end times should be complemented. Note that when the difference value is smaller than a specified value (the "second specified value" described later), the power efficiency calculation window complementation determination unit 103 determines not to complement (constant value complementation) because the load (resource consumption) of the complementary calculation process would be large.

[0044] The power efficiency calculation window complement determination unit 103 calculates the maximum time difference between the time window (Assessment Window) for power efficiency calculation and the metrics acquisition timing based on the time granularity (e.g., 1 second) for DCeP (power efficiency) calculation. Note that this maximum time difference is the sum of the difference on the start side and the difference on the end side.

[0045] Then, the power efficiency calculation window complement determination unit 103 determines whether the time difference is equal to or greater than a preset specified value (second specified value). This second specified value may be set to, for example, n% of the time window (Assessment Window) for power efficiency calculation, or may be set to a fixed value such as m seconds. This second specified value is set in the power efficiency calculation frame complement determining unit 103 in advance.

[0046] If the maximum value of the time difference is equal to or greater than the second specified value, the power efficiency calculation frame complement determination unit 103 determines to perform complementation. Then, the power efficiency calculation frame complement determination unit 103 determines, for example, linear complementation or polynomial complementation as the complementation method. The power efficiency calculation frame complement determining unit 103 is set in advance to adopt either linear complement or polynomial complement as the complement method when it is determined that complement is to be performed.

[0047] Furthermore, if the maximum value of the time difference is less than the second specified value, the power efficiency calculation frame complement determining unit 103 determines not to perform complement. If it is determined that no interpolation is to be performed, the power efficiency calculation frame interpolation determination unit 103 performs interpolation by assuming that the metrics measurement value (measurement value at the start and end of the metrics measurement) is constant during the time difference (constant value interpolation). In other words, the power efficiency calculation window complement determination unit 103 determines whether or not it is necessary to correct (complement) the discrepancy between the time window (Assessment Window) for power efficiency calculation and the metrics acquisition interval, and then determines a predetermined complement method (linear complement, polynomial complement, constant value complement, etc.).

[0048] The metrics complement unit 104 complements the metrics measurement values ​​using the complement method (linear complement, polynomial complement, constant value complement, etc.) determined by the metrics measurement value complement determination unit 102 for the metrics collected by the metrics collection unit 101. The metrics complement unit 104 also corrects (complements) the discrepancy between the time frame (assessment window) for power efficiency calculation and the metrics acquisition interval using the complement method (linear complement, polynomial complement, constant value complement, etc.) determined by the power efficiency calculation window complement determination unit 103.

[0049] The power efficiency calculation unit 105 calculates an integral value using the interpolated value of "Useful Work Produced" (amount of work product) and "Total Energy Consumed to Perform that Work" (amount of power consumption) shown in the above formula (1). The power efficiency calculation unit 105 calculates the power efficiency by dividing the calculated "Useful Work Produced" (amount of work product) by the calculated "Total Energy Consumed to Perform that Work" (amount of power consumption).

[0050] 1, the power efficiency calculation device 1 may not include the power efficiency calculation frame complement determination unit 103 in the control unit 10, and may only perform complementation between metric measurement values ​​by the metrics measurement value complement determination unit 102. Furthermore, the power efficiency calculation device 1 may not include the metrics measurement value complement determination unit 102 in the control unit 10, and may only perform correction (complementation) of the discrepancy between the time frame of the power efficiency calculation and the metrics acquisition interval by the power efficiency calculation frame complement determination unit 103.

[0051] <Processing of power efficiency calculation device> Next, the flow of processing executed by the power efficiency calculation device 1 will be described. FIG. 3 is a flowchart showing the flow of processing executed by the power efficiency calculation device 1 according to this embodiment. The processing executed by the power efficiency calculation device 1 can be roughly divided into a preparatory stage, an operational stage, and a readjustment stage.

[0052] In the advance preparation stage, the power efficiency calculation device 1 sets a first specified value using metrics collected in advance from the physical server group 30 and values ​​acquired from the power meter, and determines whether or not interpolation between metrics measurement values ​​is necessary, and determines the interpolation method. The power efficiency calculation device 1 also determines whether or not correction (interpolation) of the discrepancy between the time frame (Assessment Window) for power efficiency calculation and the metrics acquisition interval is necessary, and determines the interpolation method.

[0053] In addition, in the operation stage, the power efficiency calculation device 1 uses the interpolation method determined in the advance preparation stage to interpolate between the metrics measurement values ​​and correct (interpolate) the discrepancy between the time frame (Assessment Window) for power efficiency calculation and the metric acquisition interval, then calculates the integral value and calculates DCeP (power efficiency).

[0054] In the readjustment stage, the power efficiency calculation device 1 updates the first specified value using the metrics and power meter values ​​acquired at that time. Then, it again determines whether or not interpolation between metric measurements is necessary, and whether or not correction (interpolation) of the discrepancy between the time frame (Assessment Window) for power efficiency calculation and the metric acquisition interval is necessary, and determines the interpolation method. Then, the power efficiency calculation device 1 returns to the operation stage and repeats the process of calculating DCeP (power efficiency).

[0055] It is assumed that the second specified value is set in advance in the power efficiency calculation frame complement determining unit 103. As an application example, an example of monitoring DCeP (power efficiency) in the operating environment of a UPF (User Plane Function) application that transfers user data packets from a user terminal to a data network (such as the Internet) will be described.

[0056] First, as a preliminary preparation stage, the metrics collection unit 101 of the power efficiency calculation device 1 acquires power metrics and power meter values ​​from the physical server group 30 (step S10).

[0057] Next, the metrics measurement value complementation determination unit 102 extracts the power metrics for a predetermined period (a period longer than the Assessment Window) and the power meter values ​​during that period, and calculates the error "α" to be used as the first specified value for complementation between the metrics measurement values ​​(step S11).

[0058] Next, the metrics collector 101 acquires metrics required for power efficiency calculation from the physical server group 30 (for example, data transfer volume as task execution volume, power consumption [W]) and stores them in the metrics collection DB 100 (step S12). The collection of metrics required for power efficiency calculation in step S12 is performed continuously at predetermined intervals (assessment windows). Here, for example, the metrics measurement interval is set to every 60 minutes, and metrics are collected for 24 hours. FIG. 4 shows an example in which the UPF data transfer volume [Gbps] is collected as a metrics measurement value, and FIG. 5 shows an example in which power is collected.

[0059] Then, the metrics measurement value complement determining unit 102 determines whether or not to complement the metrics measurement values ​​for each of the acquired metrics (data transfer amount, power consumption [W]), and determines the complement method (step S13).

[0060] Specifically, the metrics measurement value complementation determination unit 102 calculates the magnitude D of the time difference shown in equation (3) or (4) for each metrics measurement value, and compares it with a first specified value (see equation (5)) calculated using the error "α" calculated in step S11. If the magnitude D of the time difference is equal to or greater than the first specified value, it determines that complementation should be performed, and determines linear complementation, polynomial complementation, or the like (linear complementation is used here) as the complementation method. On the other hand, if the magnitude D of the time difference is less than the first specified value, it determines that complementation should not be performed, and determines constant value complementation as the complementation method.

[0061] Here, for example, the first specified value of the data transfer rate (metric measurement amount) is calculated as 28.8 [Gbps] from the data shown in Figure 4 and the error α, and the first specified value of the power (metric measurement amount) is calculated as 2640 [W] from the data shown in Figure 5 and the error α, and these are compared with the magnitude of the time difference D. If the magnitude D of the time difference is calculated as, for example, 33 [Gbps] from the data transfer amount shown in FIG. 4, then 33 > 28.8, and the magnitude D of the time difference is equal to or greater than the first specified value. Therefore, it is determined that interpolation should be performed, and linear interpolation is selected as the interpolation method. On the other hand, if the magnitude D of the time difference is calculated as, for example, 2200 [W] from the power shown in FIG. 5, then 2200 < 2460, and the magnitude D of the time difference is less than the first specified value. Therefore, it is determined that interpolation should not be performed, and constant value interpolation is selected as the interpolation method.

[0062] Next, the power efficiency calculation window complement determining unit 103 determines whether to complement the difference between the time window (Assessment Window) for power efficiency calculation and the metrics acquisition interval, and determines a complement method (step S14).

[0063] Specifically, the power efficiency calculation window complement determination unit 103 calculates the maximum time difference between the time window (Assessment Window) for power efficiency calculation and the metrics acquisition timing based on the time granularity (1 minute here) for DCeP (power efficiency) calculation. If the metrics measurement interval is 60 minutes, the difference (maximum value) on the start side is 59 minutes and the difference (maximum value) on the end side is 59 minutes, so the maximum time difference is 59 + 59 = 118 minutes. Here, if the second specified value is set to 10% of the time frame (Assessment Window) for calculating the power efficiency, and the time frame (Assessment Window) for calculating the power efficiency is 360 minutes, the second specified value is 360×0.1=36 minutes. Therefore, the power efficiency calculation frame complement determination unit 103 determines that complement should be performed because the maximum time difference (118 minutes) is greater than or equal to the second specified value (36 minutes), and determines that linear complement or polynomial complement, etc. (here, linear complement is used) should be used as the complement method.

[0064] Here, if the maximum value of the time difference is equal to or greater than the second specification, the power efficiency calculation frame complement determination unit 103 may set the metrics measurement interval to be shorter. For example, the metrics measurement interval may be changed from 60 minutes to 30 minutes. This allows the maximum value of the time difference to be reduced. The processes from steps S10 to S14 are carried out in the advance preparation stage.

[0065] Next, the processing in the operation stage will be described. In the operation stage, the following steps S15 to S17 are repeatedly executed every Assessment Window (for example, every 360 minutes).

[0066] First, the metrics collection unit 101 of the power efficiency calculation device 1 collects metrics (task execution amount (data transfer amount), power) for a predetermined period (step S15).

[0067] Next, the metrics complement unit 104 of the power efficiency calculation device 1 executes metrics complementation processing (step S16). Specifically, in step S13, the metrics interpolation unit 104 performs interpolation processing between the metrics measurement values ​​using the interpolation method (linear interpolation, polynomial interpolation, constant value interpolation, etc.) determined by the metrics measurement value interpolation determination unit 102. Note that in the example of step S13, the metrics interpolation unit 104 determines linear interpolation as the interpolation method for data transfer volume and constant value interpolation as the interpolation method for power consumption, and therefore performs interpolation between the metrics measurement values ​​using the interpolation method determined for each. In step S14, the metrics complementing unit 104 performs a process of complementing the discrepancy between the time frame (Assessment Window) for power efficiency calculation and the metrics acquisition interval, using the complementing method (linear complementing, polynomial complementing, constant value complementing, etc.) determined by the power efficiency calculation window complementing determining unit 103. Note that in the example of step S14, the metrics complementing unit 104 determines linear complementing as the complementing method, and therefore complements the discrepancy between the time frame (Assessment Window) for power efficiency calculation and the metrics acquisition interval by linear complementing.

[0068] Then, the power efficiency calculation unit 105 calculates the integral of "Useful Work Produced" (amount of work product) and "Total Energy Consumed to Perform that Work" (amount of power consumption) shown in the above formula (1) by using the interpolated values. Specifically, the power efficiency calculation unit 105 calculates the integral of the task execution amount and the power metric measurement value using the interpolated data. Then, the power efficiency calculation unit 105 calculates the power efficiency by dividing the calculated "Useful Work Produced" (amount of work product) by the calculated "Total Energy Consumed to Perform that Work" (amount of power consumption) (step S17). This is the processing in the operation stage.

[0069] Next, the power efficiency calculation device 1 executes a readjustment stage process at predetermined time intervals (step S18). The readjustment stage process is the same as the processes from steps S10 to S14 in the advance preparation stage. In the readjustment stage, the power efficiency calculation device 1 executes an interpolation process at predetermined time intervals to again determine whether interpolation between metrics measurement values ​​is necessary and whether correction (interpolation) of the discrepancy between the time frame (Assessment Window) for power efficiency calculation and the metrics acquisition interval is necessary, and determines the interpolation method. By providing this readjustment stage, it becomes possible to execute an optimal interpolation process according to the change in metrics measurement values ​​over time. Then, the power efficiency calculation device 1 returns to the operation stage from step S15 and repeats the process of calculating DCeP (power efficiency).

[0070] In this way, the power efficiency calculation device 1 determines whether or not it is necessary to interpolate between the metrics measurement values ​​and to correct (interpolate) the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval, and then calculates an integral value based on the interpolated data, thereby enabling accurate calculation of power efficiency. Furthermore, by readjusting this complementation process at predetermined time intervals, the power efficiency calculation device 1 can determine whether complementation is necessary and perform calculation of DCeP (power efficiency) according to the amount of change in the metrics value over time.

[0071] <Hardware configuration> The power efficiency calculation device 1 according to this embodiment is realized by a computer 900 having a configuration as shown in FIG. 6, for example. 6 is a hardware configuration diagram showing an example of a computer 900 that realizes the functions of the power efficiency calculation device 1 according to this embodiment. The computer 900 includes a CPU (Central Processing Unit) 901, a ROM (Read Only Memory) 902, a RAM 903, an HDD (Hard Disk Drive) 904, an input / output I / F (Interface) 905, a communication I / F 906, and a media I / F 907.

[0072] The CPU 901 operates based on a program stored in the ROM 902 or the HDD 904, and performs control by the control unit 10 (FIG. 1). The ROM 902 stores a boot program executed by the CPU 901 when the computer 900 is started up, programs related to the hardware of the computer 900, and the like.

[0073] The CPU 901 controls an input device 910 such as a mouse or keyboard, and an output device 911 such as a display or printer, via an input / output I / F 905. The CPU 901 acquires data from the input device 910 via the input / output I / F 905, and outputs generated data to the output device 911.

[0074] The HDD 904 stores programs executed by the CPU 901 and data used by the programs. The communication I / F 906 receives data from other devices via a communication network (e.g., NW (Network) 920) and outputs the data to the CPU 901, and also transmits data generated by the CPU 901 to other devices via the communication network.

[0075] The media I / F 907 reads a program or data stored in the recording medium 912 and outputs it to the CPU 901 via the RAM 903. The CPU 901 loads a program related to a target process from the recording medium 912 onto the RAM 903 via the media I / F 907, and executes the loaded program. The recording medium 912 is an optical recording medium such as a DVD (Digital Versatile Disc) or a PD (Phase Change Rewritable Disc), a magneto-optical recording medium such as an MO (Magneto Optical Disk), a magnetic recording medium, a semiconductor memory, or the like.

[0076] For example, when a computer 900 functions as the power efficiency calculation device 1 according to this embodiment, a CPU 901 of the computer 900 executes a program loaded onto a RAM 903, thereby realizing the functions of the power efficiency calculation device 1. Furthermore, data in the RAM 903 is stored in an HDD 904. The CPU 901 reads and executes a program related to a target process from a recording medium 912. Alternatively, the CPU 901 may read a program related to a target process from another device via a communication network (NW 920).

[0077] <Effects> The effects of the power efficiency calculation device 1 and the like according to the present invention will be described below. The power efficiency calculation device according to the present invention is a power efficiency calculation device 1 that calculates the power efficiency resulting from the execution of an application 3 installed on a physical server group 30, and includes a metrics collection unit 101 that acquires, from the physical server group 30, metrics measurements that are evaluation indexes required for power efficiency calculation and that include the task execution amount of the application 3 and the power consumption of the physical server group 30; a metrics measurement value complementation determination unit 102 that determines whether or not complementation between the metrics measurements is required based on whether or not a change in the metrics measurement value is equal to or greater than a first specified value, and determines a predetermined complementation method depending on whether or not complementation is required; and a power efficiency calculation unit 103 that determines whether or not a difference between the time frame for power efficiency calculation and the metrics acquisition interval is equal to or greater than a second specified value. a power efficiency calculation frame complement determination unit (103) that determines whether or not complement is required for the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval, and determines a predetermined complement method depending on whether complement is required; a metrics complement unit (104) that complements the metrics measurement values ​​using the predetermined complement method determined by the metrics measurement value complement determination unit (102), and complements the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval using the predetermined complement method determined by the power efficiency calculation frame complement determination unit (103); and a power efficiency calculation unit (105) that calculates the power efficiency by calculating an integral value over the time frame for power efficiency calculation of the task execution amount and the power consumption of the physical server group (30), which are the metrics measurement values ​​complemented by the metrics complement unit (104).

[0078] By doing this, the power efficiency calculation device 1 determines whether or not it is necessary to interpolate between the metrics measurement values ​​and to correct (interpolate) the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval, and then calculates an integral value using the data of the interpolated metrics measurement values, thereby enabling accurate calculation of power efficiency.

[0079] The power efficiency calculation device according to the present invention is a power efficiency calculation device 1 that calculates the power efficiency resulting from the execution of an application 3 installed on a physical server group 30, and is characterized by comprising: a metrics collection unit 101 that acquires, from the physical server group 30, metrics measurements that are evaluation indices necessary for power efficiency calculation and that include the task execution amount of the application and the power consumption of the physical server group 30; a metrics measurement value complementation determination unit 102 that determines whether or not interpolation between the metrics measurement values ​​is necessary based on whether or not a change in the metrics measurement value is greater than or equal to a first specified value, and determines a predetermined interpolation method depending on whether or not interpolation is necessary; a metrics complementation unit 104 that complements the metrics measurement values ​​using the predetermined interpolation method determined by the metrics measurement value complementation determination unit 102; and a power efficiency calculation unit 105 that calculates the power efficiency by calculating an integral value over a time frame of power efficiency calculation of the task execution amount and the power consumption of the physical server group 30, which are the metrics measurement values ​​complemented by the metrics complementation unit 104.

[0080] In this way, the power efficiency calculation device 1 can determine whether or not interpolation between the metrics measurement values ​​is necessary, and then calculate an integral value using the data of the interpolated metrics measurement values, thereby calculating accurate power efficiency.

[0081] Furthermore, in the power efficiency calculation device 1, the metrics measurement value complementation determination unit 102 is characterized in that, when the amount of change in the metrics measurement value is equal to or greater than a first specified value, it determines the predetermined complementation method between the metrics measurement values ​​to be linear complementation or polynomial complementation, and when the amount of change in the metrics measurement value is less than the first specified value, it determines the predetermined complementation method between the metrics measurement values ​​to be constant value complementation, assuming that the value between the metrics measurement values ​​is constant.

[0082] In this way, when the power efficiency calculation device 1 determines that interpolation is necessary, it can perform appropriate interpolation using linear interpolation or polynomial interpolation. On the other hand, when the power efficiency calculation device 1 determines that interpolation is unnecessary, it can reduce overhead due to unnecessary calculation by using constant value interpolation.

[0083] Furthermore, in the power efficiency calculation device 1, the metrics collection unit 101 acquires the values ​​of the power meters of the physical server group 30, and the metrics measurement value complementation determination unit 102 calculates an error based on the power meter value and a value obtained by linearly complementing or polynomially complementing the power, which is the metrics measurement value, and sets the first specified value as a value that becomes smaller as the difference between the calculated error and a preset error tolerance rate becomes smaller.

[0084] In this way, the power efficiency calculation device 1 can set an appropriate first specified value according to the amount of change in the metrics measurement value, which allows the power efficiency calculation device 1 to appropriately determine whether or not interpolation between metrics measurement values ​​is necessary and the interpolation method.

[0085] Furthermore, in the power efficiency calculation device 1, the metrics collection unit 101 acquires the power meter values ​​of the physical server group 30 and the power consumption, which is a metrics measurement value, at predetermined intervals, and the metrics measurement value complementation determination unit 102 updates the first specified value using the acquired power meter values ​​and the power consumption, which is a metrics measurement value.

[0086] In this way, by updating the first specified value using the power meter value and the power consumption, which is a metric measurement value, obtained at specified intervals, it is possible to appropriately determine whether or not supplementation is necessary and decide on a supplementation method based on the change in the metric measurement value over time.

[0087] The power efficiency calculation device according to the present invention is a power efficiency calculation device 1 that calculates the power efficiency resulting from the execution of an application 3 installed on a physical server group 30, and is characterized by comprising: a metrics collection unit 101 that acquires from the physical server group 30 metrics measurement values ​​that are evaluation indices required for power efficiency calculation and include the task execution amount of the application 3 and the power consumption of the physical server group; a power efficiency calculation frame completion determination unit 103 that determines whether or not the difference between the time frame of the power efficiency calculation and the metrics acquisition interval needs to be complemented based on whether or not the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is equal to or greater than a second specified value, and determines a predetermined complement method depending on whether or not the complement is necessary; a metrics completion unit 104 that complements the difference between the time frame of the power efficiency calculation and the metrics acquisition interval using the predetermined complement method determined by the power efficiency calculation frame completion determination unit 103; and a power efficiency calculation unit 105 that calculates the power efficiency by calculating an integral over the power efficiency calculation time frame of the task execution amount and the power consumption of the physical server group 30, which are the metrics measurement values ​​complemented by the metrics completion unit 104.

[0088] By doing this, the power efficiency calculation device 1 can determine whether or not it is necessary to correct (complement) the discrepancy between the time frame for power efficiency calculation and the metrics acquisition interval, and then calculate an integral value using the data of the complemented metrics measurement values, thereby calculating accurate power efficiency.

[0089] Furthermore, in the power efficiency calculation device 1, the power efficiency calculation frame complement determination unit 103 is characterized in that, when the value of the deviation between the time frame of the power efficiency calculation and the metrics acquisition interval is equal to or greater than a second specified value, it determines the predetermined complement method to be linear complement or polynomial complement, and when the value of the deviation between the time frame of the power efficiency calculation and the metrics acquisition interval is less than the second specified value, it determines the predetermined complement method for the deviation between the time frame of the power efficiency calculation and the metrics acquisition interval to be constant value complement, assuming that the metric measurement value of the deviation between the time frame of the power efficiency calculation and the metrics acquisition interval can be regarded as a constant value.

[0090] In this way, when the power efficiency calculation device 1 determines that interpolation is necessary, it can perform appropriate interpolation using linear interpolation or polynomial interpolation. On the other hand, when the power efficiency calculation device 1 determines that interpolation is unnecessary, it can reduce overhead due to unnecessary calculation by using constant value interpolation.

[0091] The present invention is not limited to the above-described embodiments, and many modifications can be made by a person having ordinary skill in the art within the technical concept of the present invention. [Explanation of symbols]

[0092] 1. Power efficiency calculation device 3. Application 10 Control Unit 11 Input / output section 12 Storage section 30 physical servers 100 Metrics Collection DB 101 Metrics Collection Department 102 Metrics measurement value complementation judgment unit 103 Power efficiency calculation frame completion judgment unit 104 Metrics Complementary Section 105 Power Efficiency Calculation Unit 1000 Power Efficiency Calculation System

Claims

1. A power efficiency calculation device that calculates power efficiency by executing an application installed on a group of physical servers, a metrics collection unit that acquires, from the group of physical servers, metrics measurement values ​​that are evaluation indices necessary for calculating power efficiency, and that include task execution amounts of the applications and power consumption of the group of physical servers; a metrics measurement value complementation determination unit that determines whether or not complementation between metrics measurement values ​​is necessary based on whether or not the amount of change in the metrics measurement value is equal to or greater than a first specified value, and determines a predetermined complementation method depending on whether or not complementation is necessary; a power efficiency calculation frame complement determination unit that determines whether or not complementation of the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is necessary based on whether or not the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is equal to or greater than a second specified value, and determines a predetermined complement method depending on whether or not complementation is necessary; a metrics complementing unit that complements between metrics measurement values ​​using a predetermined complementing method determined by the metrics measurement value complementing determination unit, and complements a discrepancy between a time frame for power efficiency calculation and a metrics acquisition interval using the predetermined complementing method determined by the power efficiency calculation frame complementing determination unit; a power efficiency calculation unit that calculates an integral value of the task execution amount and the power consumption of the group of physical servers, which are the metrics measurement values ​​complemented by the metrics complement unit, over a time frame for the power efficiency calculation, thereby calculating power efficiency; A power efficiency calculation device comprising:

2. A power efficiency calculation device that calculates power efficiency by executing an application installed on a group of physical servers, a metrics collection unit that acquires, from the group of physical servers, metrics measurement values ​​that are evaluation indices necessary for calculating power efficiency, and that include task execution amounts of the applications and power consumption of the group of physical servers; a metrics measurement value complementation determination unit that determines whether or not complementation between metrics measurement values ​​is necessary based on whether or not the amount of change in the metrics measurement value is equal to or greater than a first specified value, and determines a predetermined complementation method depending on whether or not complementation is necessary; a metrics complementation unit that complements the metrics measurement values ​​using the predetermined complementation method determined by the metrics measurement value complementation determination unit; a power efficiency calculation unit that calculates an integral value of the task execution amount and the power consumption of the group of physical servers, which are the metrics measurement values ​​complemented by the metrics complement unit, over a time frame for the power efficiency calculation, thereby calculating power efficiency; A power efficiency calculation device comprising:

3. The metrics measurement value complementation determination unit When the amount of change in the metric measurement value is equal to or greater than the first specified value, the predetermined interpolation method between the metric measurement values ​​is determined to be linear interpolation or polynomial interpolation, and when the amount of change in the metric measurement value is less than the first specified value, the predetermined interpolation method between the metric measurement values ​​is determined to be constant value interpolation, assuming that the value between the metric measurement values ​​is constant.

3. The power efficiency calculation device according to claim 1 or 2, wherein:

4. the metrics collection unit acquires values ​​from the power meters of the group of physical servers; The metrics measurement value complementation determination unit calculates an error based on the value of the power meter and a value obtained by linearly complementing or polynomially complementing the power, which is the metrics measurement value, and sets the first specified value as a value that decreases as the difference between the calculated error and a preset error tolerance rate decreases.

4. The power efficiency calculation device according to claim 3,

5. the metrics collection unit acquires power meter values ​​of the group of physical servers and power consumption, which is a metrics measurement value, for each predetermined period; The metrics measurement value complementation determination unit updates the first specified value using the acquired value of the power meter and the power consumption that is the metrics measurement value.

5. The power efficiency calculation device according to claim 4,

6. A power efficiency calculation device that calculates power efficiency by executing an application installed on a group of physical servers, a metrics collection unit that acquires, from the group of physical servers, metrics measurement values ​​that are evaluation indices necessary for calculating power efficiency, and that include task execution amounts of the applications and power consumption of the group of physical servers; a power efficiency calculation frame complement determination unit that determines whether or not complementation of the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is necessary based on whether or not the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is equal to or greater than a second specified value, and determines a predetermined complement method depending on whether or not complementation is necessary; a metrics complementing unit that complements a discrepancy between a time frame for power efficiency calculation and a metrics acquisition interval using a predetermined complementing method determined by the power efficiency calculation frame complementing determination unit; a power efficiency calculation unit that calculates an integral value of the task execution amount and the power consumption of the group of physical servers, which are the metrics measurement values ​​complemented by the metrics complement unit, over a time frame for the power efficiency calculation, thereby calculating power efficiency; A power efficiency calculation device comprising:

7. The power efficiency calculation frame complement determination unit When the value of the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is equal to or greater than the second specified value, the predetermined interpolation method is determined to be linear interpolation or polynomial interpolation, and when the value of the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is less than the second specified value, the predetermined interpolation method for the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is determined to be constant value interpolation, assuming that the metric measurement value for the difference between the time frame of the power efficiency calculation and the metrics acquisition interval can be regarded as a constant value.

7. The power efficiency calculation device according to claim 1, wherein:

8. A power efficiency calculation method for a power efficiency calculation device that calculates power efficiency by executing an application installed on a group of physical servers, comprising: The power efficiency calculation device a metrics collection step of acquiring, from the group of physical servers, metrics measurement values ​​which are evaluation indices necessary for calculating power efficiency and include task execution amounts of the applications and power consumption of the group of physical servers; a metrics measurement value complementation determination step for determining whether or not complementation between metrics measurement values ​​is necessary based on whether or not the amount of change in the metrics measurement value is equal to or greater than a first specified value, and for determining a predetermined complementation method depending on whether or not complementation is necessary; a power efficiency calculation frame complementation determination step for determining whether or not complementation of the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is necessary based on whether or not the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is equal to or greater than a second specified value, and for determining a predetermined complementation method depending on whether or not complementation is necessary; a metrics interpolation step of interpolating between metrics measurements by the predetermined interpolation method determined in the metrics measurement value interpolation determination step, and interpolating a difference between a time frame for power efficiency calculation and a metrics acquisition interval by the predetermined interpolation method determined in the power efficiency calculation frame interpolation determination step; a power efficiency calculation step of calculating an integral value of the task execution amount and the power consumption of the group of physical servers, which are the metric measurement values ​​interpolated in the metric interpolation step, over a time frame of the power efficiency calculation, to calculate power efficiency; A power efficiency calculation method comprising:

9. A power efficiency calculation system including: a group of physical servers on which applications that execute one or more tasks are installed; and a power efficiency calculation device that calculates power efficiency resulting from execution of the applications installed on the group of physical servers, The power efficiency calculation device a metrics collection unit that acquires, from the group of physical servers, metrics measurement values ​​that are evaluation indices necessary for calculating power efficiency, and that include task execution amounts of the applications and power consumption of the group of physical servers; a metrics measurement value complementation determination unit that determines whether or not complementation between metrics measurement values ​​is necessary based on whether or not the amount of change in the metrics measurement value is equal to or greater than a first specified value, and determines a predetermined complementation method depending on whether or not complementation is necessary; a power efficiency calculation frame complement determination unit that determines whether or not complementation of the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is necessary based on whether or not the difference between the time frame of the power efficiency calculation and the metrics acquisition interval is equal to or greater than a second specified value, and determines a predetermined complement method depending on whether or not complementation is necessary; a metrics complementing unit that complements between metrics measurement values ​​using a predetermined complementing method determined by the metrics measurement value complementing determination unit, and complements a discrepancy between a time frame for power efficiency calculation and a metrics acquisition interval using the predetermined complementing method determined by the power efficiency calculation frame complementing determination unit; a power efficiency calculation unit that calculates an integral value of the task execution amount and the power consumption of the group of physical servers, which are the metrics measurement values ​​complemented by the metrics complement unit, over a time frame for the power efficiency calculation, thereby calculating power efficiency; A power efficiency calculation system comprising:

10. A program for causing a computer to function as the power efficiency calculation device according to any one of claims 1 to 6.

Images