Evaluation device and evaluation method

Abstract

In an evaluation device 10, a metrics acquisition unit 11 acquires meta information, power consumption, and an internal state from each of apparatuses used for software development, a metrics storage unit 13 appends time stamps to the power consumption and the internal state and stores the same as metrics, a development project information registration unit 16 acquires information about a software development project, an environmental load calculation unit 18 calculates a power consumption amount on the basis of the power consumption of the apparatuses during a development period, and an environmental load analysis unit 19 analyzes and visualizes the metrics and the meta information of the apparatuses during the development period.

Description

Evaluation device and evaluation method

[0001] The present disclosure relates to an evaluation device and an evaluation method.

[0002] Greenhouse gases (GHGs), including CO2, are believed to be the cause of climate change, and there is a social and global demand to reduce emissions.

[0003] As Information and Communication Technology (ICT) continues to spread and expand, CO2 emissions from electricity consumption are also increasing year by year. Until now, in the ICT field, measures to reduce CO2 emissions have focused on hardware, such as reducing the power consumption of processors.

[0004] On the other hand, software does not directly consume electricity, and as such, not enough efforts have been made to address this issue.However, many ICT devices and equipment are used throughout the software lifecycle (development, operation, disposal, etc.), and they indirectly consume electricity.

[0005] For tangible products, it is common to evaluate the environmental impact at specific stages, such as during manufacturing, based on Life-Cycle Assessment (LCA), but this is not currently possible for intangible software. There has been insufficient discussion about software, and no established methods have been established for measuring, calculating, and analyzing the environmental impact at each stage of the life cycle, including during development. Because no established methods have been established, it is difficult to reduce environmental impacts.

[0006] Additionally, the Pathfinder Framework, a guidance document on calculating GHG emissions per product published by the World Business Council for Sustainable Development (WBSCD), emphasizes the importance of calculating emissions based on actual measurements (primary data). The importance of actual measurements is increasing in order to accurately grasp emissions, and the software lifecycle is no exception.

[0007] There is a demand to make the environmental impact of software observable throughout its life cycle.

[0008] Kei Nishigaki, Takahiro Tsujii, Hideki Sunahara, "A Power Consumption Visualization System for Energy Saving in Computer Rooms," Information Processing Society of Japan Report, 2011, Vol. 2011-IOT-12 No. 35, pp. 1-6. Tomoki Kawaguchi, Shuhei Tajima, Takashi Tomii, "SEE-Con: A Power Consumption Visualization System for Office Spaces with Many Electrical Appliances Considering the Context of Power Usage," DEIM2012, 2012, C6-2.

[0009] The environmental impact of software development can be attributed to the use of ICT equipment, including PCs and servers. Non-Patent Document 1 is a prior art for visualizing the power consumption of ICT equipment. In Non-Patent Document 1, a prototype system is developed to visualize the power consumption of servers in computer rooms, and approaches to reducing power consumption are considered based on the acquired power consumption and temperature data. However, the conclusion is limited to adjusting the temperature.

[0010] Non-Patent Document 2 proposes a system that acquires and visualizes the context of office devices along with their power consumption. However, the context actually acquired is too insufficient to be called context, such as the effective usage time of a PC, and does not take into account the device's unique characteristics or the device's state that reflects user behavior.

[0011] Both methods make it difficult to identify hotspots (e.g., processes or activities with particularly high environmental impacts) in the context of software development, making it difficult to achieve substantial reductions.

[0012] The present disclosure has been made in view of the above, and aims to enable more accurate and less time-consuming environmental impact assessment during software development.

[0013] An evaluation device according to one embodiment of the present disclosure acquires meta-information, power consumption, and internal state from each piece of equipment used in software development, timestamps the acquired power consumption and internal state and stores them as metrics, acquires information about the software development project, calculates the amount of power consumption based on the power consumption of the equipment during the development period of the software development project, and analyzes and visualizes the metrics and meta-information of the equipment during the development period of the software development project.

[0014] According to the present disclosure, more accurate and less time-consuming environmental impact assessment is possible during software development.

[0015] Fig. 1 is a diagram showing an example of the configuration of an evaluation system. Fig. 2 is a diagram showing an example of primary data. Fig. 3 is a flowchart showing an example of the flow of a process for evaluating environmental impact. Fig. 4 is a diagram showing an example of the hardware configuration of an evaluation device.

[0016] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present disclosure will be described below with reference to the accompanying drawings, which are intended to explain the environmental impact caused by the use of ICT devices during software development.

[0017] An example of the configuration of an evaluation system according to this embodiment will be described with reference to FIG. 1 . The evaluation system shown in the figure includes a metrics acquisition unit 11 and an evaluation device 10. ICT devices are devices used during software development, such as personal computers (PCs), displays, servers, network devices, UPSs, multifunction devices, and printers. The power consumption of the ICT devices is measured using a power meter. For example, the power consumption of the ICT devices is measured at regular intervals using a network-connectable power meter or an internal sensor of the ICT devices. Examples of the former include a smart plug, and examples of the latter include a baseboard management controller (BMC). While FIG. 1 illustrates one ICT device, one power meter, and one metrics acquisition unit 11, there may be multiple ICT devices, and the number of power meters and metrics acquisition units 11 may correspond to the number of ICT devices.

[0018] The metrics acquisition unit 11 is arranged on the client side, and the evaluation device 10 is arranged on the server side.

[0019] The evaluation device 10 comprises a metrics aggregation unit 12, a metrics storage unit 13, a metrics sequential analysis unit 14, a metrics processing unit 15, a development project information registration unit 16, a primary data storage unit 17, an environmental load calculation unit 18, and an environmental load analysis unit 19. Each unit will be described below.

[0020] The metrics acquisition unit 11 sets meta-information in the ICT device and acquires the power consumption and internal status of the ICT device. If the ICT device to be measured is a PC or server, the metrics acquisition unit 11 runs on a general OS of the ICT device. If the metrics acquisition unit 11 cannot be installed on ICT devices such as network devices or UPSs, a device on which the metrics acquisition unit 11 can run is separately installed, and the internal status and power consumption are acquired from the ICT device or power meter.

[0021] The metrics acquisition unit 11 manages meta information for each ICT device. The meta information includes, for example, an electric power company identifier, an ICT device identifier, an ICT device type identifier, and information indicating the specifications of the ICT device.

[0022] The power company identifier is an identifier used to identify the power company that supplies the electricity used by ICT equipment. The power company identifier is necessary to identify the CO2 emission coefficient of the ICT equipment.

[0023] The ICT device identifier is an identifier for identifying each ICT device.

[0024] The ICT device type identifier is an identifier that indicates the type of ICT device, such as a PC, a display, a server, storage, a network device, a UPS, a multifunction device, or a printer.

[0025] Information indicating the specifications of ICT equipment is information that indicates the specifications of the target ICT equipment, such as platform and hardware information, etc. For example, information indicating the specifications of ICT equipment includes the type of OS, the version of the OS, the CPU architecture, the number of physical cores of the CPU, the number of logical cores of the CPU, the minimum clock frequency, the maximum clock frequency, etc.

[0026] The user sets meta-information about each ICT device in the metrics acquisition unit 11. Information indicating the specifications of the ICT device may be acquired mechanically by the metrics acquisition unit 11 using a predetermined protocol. For example, information indicating the specifications of the ICT device can be acquired from information on the file system of a general OS or using a protocol such as SNMP. The items of information indicating the specifications of the ICT device that can be acquired vary depending on the ICT device.

[0027] The metrics acquisition unit 11 automatically acquires the power consumption of the ICT equipment from the power meter using a predetermined protocol. For example, if a network-connectable power meter is used as the power meter, the metrics acquisition unit 11 acquires the power consumption via Bluetooth Low Energy (BLE) or Wi-Fi (TCP / IP). If an internal sensor is used, the metrics acquisition unit 11 acquires the power consumption via IPMI / Redfish, SNMP, or the like.

[0028] The internal state of an ICT device is the state of each part of the ICT device that reflects user behavior, such as resource information and activity. For example, the internal state of an ICT device may include CPU usage, clock frequency, CPU temperature, CPU fan speed, memory usage, disk I / O, network I / O, resource usage by process, AFK / not AFK, and the application name of the active window. AFK stands for away from keyboard and means that the user is not actively performing an operation. Not AFK means that the user is actively performing an operation.

[0029] The metrics acquisition unit 11 mechanically acquires the internal state of ICT devices using a predetermined protocol. For example, the internal state of ICT devices can be acquired from information on the file system of a general OS or using a protocol such as SNMP. The items that can be acquired regarding the internal state of ICT devices vary depending on the ICT device.

[0030] The metrics acquisition unit 11 generates a timestamp for the time when the power consumption and internal state are acquired, and transmits the meta information of the ICT device, the acquired power consumption and internal state, and the timestamp to the metrics aggregator 12 using a predetermined protocol. Hereinafter, the power consumption and internal state with the attached timestamp will be referred to as metrics.

[0031] The metrics aggregator 12 processes the metrics received from each of the metrics acquirers 11 into a data format for the metrics storage unit 13 , and writes the processed metrics into the metrics storage unit 13 .

[0032] The metrics storage unit 13 stores the metrics written by the metrics aggregator 12 as time-series data. For example, the metrics storage unit 13 stores metrics in a key-value format, where a metric with meta-information and a timestamp added is the key, and the corresponding value is the value. To give a specific example, the data stored in the metrics storage unit 13 is in the format `wattage{device_id="1",device_type="pc",...} @1701442753718 50`. "wattage" is the metric name, the text in parentheses is the meta-information, and the number after @ is the timestamp. The part up to the timestamp is the key. "50" is the value, which in this case means the number of watts.

[0033] The metrics storage unit 13 outputs the metrics to be stored in response to a request from the metrics sequential analysis unit 14 or the metrics processing unit 15 .

[0034] The metrics sequential analysis unit 14 requests desired metrics from the metrics storage unit 13 and performs data summarization and visualization of the obtained metrics in real time. For example, the metrics sequential analysis unit 14 obtains the power consumption and CPU usage rate of each ICT device from the metrics storage unit 13, calculates the average value of the obtained power consumption, and visualizes changes in power consumption and changes in CPU usage rate.

[0035] The metrics processor 15 requests the metrics of the power consumption and internal state of each ICT device for a specific period (for example, one hour or one day) from the metrics storage unit 13 according to a predetermined method. The metrics processor 15 calculates the amount of power consumption based on the power consumption for the specific period, and performs data summarization (such as averaging) of the internal state for the specific period as needed. The metrics processor 15 writes the processed data to the primary data storage unit 17. For example, the processed data includes, for each ICT device, meta information, the specific period, the amount of power consumption for the specific period, and the internal state for the specific period.

[0036] The development project information registration unit 16 accepts registration of information related to a development project from a user and writes the development project information to the primary data storage unit based on a predetermined protocol. The development project information includes, for example, the software type, development period, the period of each process, and development method. The development project information may also include information on ICT equipment used during development and daily work reports of developers.

[0037] The primary data storage unit 17 stores the primary data written by the metrics processing unit 15 and the development project information registration unit 16. RDBMS or NoSQL can be used to store the primary data. The primary data storage unit 17 also stores the CO2 emission coefficient for each electric power company, which is used to calculate CO2 emissions.

[0038] The primary data storage unit 17 outputs the stored primary data in response to requests from the environmental load calculation unit 18 and the environmental load analysis unit 19. For example, when the primary data storage unit 17 receives a request specifying a period and ICT equipment, it outputs meta information for the specified ICT equipment, as well as information about the development project for the period, power consumption, and internal state. Figure 2 shows an example of the primary data output by the primary data storage unit 17. The primary data in the figure includes period information 110, development project information 120, ICT equipment meta information 130, ICT equipment power consumption 140, and ICT equipment internal state 150. The "electric company" in the meta information 130 is the identifier of the electric power company that supplies the electricity used by the ICT equipment. The ICT equipment meta information, ICT equipment internal state, and development project information are collectively referred to as context.

[0039] The environmental load calculation unit 18 requests all power consumption data for the development period from the primary data storage unit 17 and calculates the total power consumption and total CO2 emissions. The total power consumption can be calculated by adding up the power consumption of each ICT device obtained from the primary data storage unit 17. The total CO2 emissions can be calculated by multiplying the power consumption of each ICT device by the CO2 emission coefficient of the power company used by the ICT device to calculate the CO2 emissions of each ICT device, and then adding up the CO2 emissions of each ICT device. For example, when a user specifies a development project to be calculated, the environmental load calculation unit 18 requests primary data of the ICT devices used in the specified development project during the development period from the primary data storage unit 17 and calculates the total power consumption and total CO2 emissions from the obtained primary data.

[0040] The environmental load analysis unit 19 retrieves desired data from the primary data storage unit 17 based on a predetermined protocol and performs data summarization (also called analysis) and visualization. Data summarization and visualization include, for example, calculating statistical values ​​such as the mean, median, mode, variance, or standard deviation, visualizing the data distribution, indicating data ranges and outliers, and showing the relationship between power consumption and internal states or between internal states. Because the primary data includes not only power consumption but also context, multifaceted analysis of power consumption and CO2 emissions is possible. In other words, it becomes possible to understand the causal relationship between device-specific characteristics and user behavior and power consumption and CO2 emissions.

[0041] For example, visualizing the power consumption and CO2 emissions at each development stage and analyzing hotspots in the development process and tasks (meetings, documentation, coding, testing, etc.) can be useful in planning measures to reduce CO2 emissions. Visualizing the differences in power consumption and CO2 emissions between high-spec and low-spec ICT equipment can be useful in selecting the optimal ICT equipment. Visualizing the relationship between ICT equipment status (e.g., CPU load rate, memory usage rate, CPU temperature) and power consumption can help reduce power consumption by balancing the load, improving CPU cooling functions, or using higher-spec ICT equipment. Regarding user behavior, visualizing the relationship between AFK and power consumption can be used to analyze the impact of user operations on power consumption and recommend appropriate sleep mode settings.

[0042] The environmental load calculation unit 18 may calculate the total power consumption and total CO2 emissions for each process during the development period (e.g., requirements definition, basic design, detailed design, manufacturing / unit testing, etc.). The environmental load analysis unit 19 may summarize and visualize data for each process during the development period.

[0043] An example of the flow of a process for evaluating environmental load will be described with reference to the flowchart of FIG.

[0044] In step S11 , the metrics acquisition unit 11 of each ICT device acquires power consumption and internal state, and transmits them to the metrics aggregator 12 .

[0045] In step S12 , the metrics aggregator 12 processes the received metrics into the data format of the metrics storage unit 13 and stores them in the metrics storage unit 13 .

[0046] In step S13, the metrics processing unit 15 acquires metrics for a specific period from the metrics storage unit 13, and processes the metrics, such as calculating the amount of power consumption and summarizing the internal state.

[0047] In step S14, the environmental load calculation unit 18 calculates the total power consumption and total CO2 emissions during the development period.

[0048] In step S15, the environmental load analysis unit 19 summarizes and visualizes data in the context of the development period.

[0049] As described above, in the evaluation device 10 of this embodiment, the metrics acquisition unit 11 acquires meta-information, power consumption, and internal state data from each piece of equipment used in software development. The metrics storage unit 13 timestamps and stores the acquired power consumption and internal state data as metrics. The development project information registration unit 16 acquires information about the software development project. The environmental load calculation unit 18 calculates the amount of power consumption based on the power consumption of the equipment during the development period. The environmental load analysis unit 19 analyzes and visualizes the metrics and meta-information for the equipment during the development period. The evaluation device 10 of this embodiment measures the power consumption of ICT equipment used during software development, and also acquires information about the specifications and internal state of the ICT equipment and the development project, thereby enabling causal relationships with power consumption to be traced. This enables accurate calculations based on actual measurements and autonomous calculations related to the "use of ICT equipment" during development on a development project-by-development (i.e., software-by-software) basis, enabling more accurate and effortless evaluation of the environmental load during software development.

[0050] According to this embodiment, in addition to measuring the power consumption of ICT devices used during software development, information on the use and internal state of the ICT devices and development projects is also acquired, making it possible to trace causality with power consumption. This allows context to be associated with the power consumption of ICT devices, making it possible to capture when, in what circumstances, and to what extent power is consumed, thereby identifying hotspots and ultimately facilitating effective CO2 reduction.

[0051] The evaluation device 10 described above can be, for example, a general-purpose computer system including a central processing unit (CPU) 901, a memory 902, a storage 903, a communication device 904, an input device 905, and an output device 906, as shown in Fig. 4. In this computer system, the evaluation device 10 is realized by the CPU 901 executing a predetermined program loaded onto the memory 902. This program can be recorded on a computer-readable non-transitory recording medium such as a magnetic disk, an optical disk, or a semiconductor memory, or can be distributed via a network.

[0052] REFERENCE SIGNS LIST 10 Evaluation device 11 Metrics acquisition unit 12 Metrics aggregation unit 13 Metrics storage unit 14 Metrics sequential analysis unit 15 Metrics processing unit 16 Development project information registration unit 17 Primary data storage unit 18 Environmental load calculation unit 19 Environmental load analysis unit

Claims

1. An evaluation device that acquires meta-information, power consumption, and internal states from each piece of equipment used in software development, assigns timestamps to the acquired power consumption and internal states and stores them as metrics, acquires information about the software development project, calculates power consumption based on the power consumption of the equipment during the development period of the software, and analyzes and visualizes the metrics and meta-information of the equipment during the development period of the software.

2. An evaluation device according to claim 1, wherein the meta-information includes an identifier of an electric power company that supplies the power used by the device, and the CO2 emissions are calculated from the amount of power consumed and the CO2 emission coefficient of the electric power company.

3. The evaluation device according to claim 1, wherein the evaluation device calculates the amount of power consumption in each stage of the software development, and analyzes and visualizes metrics for each stage.

4. An evaluation method in which a computer acquires meta-information, power consumption, and internal state from each piece of equipment used in software development, timestamps the acquired power consumption and internal state and stores them as metrics, acquires information about the software development project, calculates the amount of power consumption based on the power consumption of the equipment during the development period of the software, and analyzes and visualizes the metrics and meta-information of the equipment during the development period of the software.

Images