Control device, control method and control program

Abstract

To reduce power consumption of a network device.SOLUTION: A storage part 11 stores past information 13 showing a first load of a network device 20 measured during a first period or first power mode of the network device 20 determined from the first load. A processing part 12 uses the past information 13 to determine whether an index value indicating the magnitude of a variation of the first load for the first period is a threshold or more. The processing part 12 operates the network device 20 in the first power mode during a second period after the first period in the case where the index value is not the threshold or more. The processing part 12 measures a second load of the network device 20 to operate the network device 20 with a second power mode determined from the second load for the second period in the case where the index value is the threshold or more.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a control device, a control method, and a control program. [Background technology]

[0002] Information processing systems often include network devices such as routers. Network devices often operate for long periods of time, sometimes 24 hours a day, 365 days a year, without interruption. This can lead to significant power consumption by network devices. Therefore, technologies for reducing the power consumption of network devices have been proposed.

[0003] For example, a communication device has been proposed that monitors CPU (Central Processing Unit) usage and data transfer volume and selects a CPU operating frequency appropriate for the current CPU usage and data transfer volume. Also, a communication system has been proposed that collects status data indicating past communication volumes, analyzes the status data, and determines an appropriate communication speed depending on the day of the week.

[0004] Furthermore, for example, an information processing device has been proposed that collects statistical information indicating the number of past transactions, analyzes the statistical information up to the previous day, and determines the time period during which an automated transaction device should be set to power-saving mode. The proposed information processing device monitors the number of transactions for the day, and if the number of transactions for the day exceeds a threshold, cancels the transition of the automated transaction device to power-saving mode. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2013-126055 [Patent Document 2] Japanese Patent Publication No. 2022-10968 [Patent Document 3] Japanese Patent Application Laid-Open No. 2012-203804 Summary of the Invention [Problem to be solved by the invention]

[0006] When attempting to select a power mode of a network device in real time based on the current load, the process of monitoring the load and selecting a power mode itself imposes a burden. Meanwhile, when determining a power mode schedule for a network device based on past load trends, there is a risk that the current load will deviate from the past load trends. Therefore, there is room for improvement in reducing the power consumption of network devices. Therefore, in one aspect, the present invention aims to reduce the power consumption of a network device. [Means for solving the problem]

[0007] In one aspect, a control device is provided that includes a storage unit and a processing unit. The storage unit stores past information indicating a first load of a network device measured during a first time period or a first power mode of the network device determined from the first load among multiple power modes with different power consumption. The processing unit uses the past information to determine whether an index value indicating the magnitude of fluctuation in the first load during the first time period is equal to or greater than a threshold. If the index value is not equal to or greater than the threshold, the processing unit operates the network device in the first power mode during a second time period that follows the first time period. If the index value is equal to or greater than the threshold, the processing unit measures a second load of the network device during the second time period and operates the network device in the second power mode determined from the second load among the multiple power modes.

[0008] In one embodiment, a control method executed by a computer is provided. In another embodiment, a control program executed by a computer is provided. [Effects of the Invention]

[0009] In one aspect, the power consumption of the network device is reduced. [Brief explanation of the drawings]

[0010] [Figure 1]FIG. 2 is a diagram for explaining a control device according to the first embodiment. [Figure 2] FIG. 10 illustrates an example of hardware of a network device according to a second embodiment. [Figure 3] FIG. 10 illustrates an example of hardware of a terminal device according to a second embodiment. [Figure 4] FIG. 10 is a diagram illustrating an example of a setting table. [Figure 5] FIG. 10 is a diagram illustrating an example of a measurement value table. [Figure 6] FIG. 10 is a diagram illustrating an example of a power mode table. [Figure 7] FIG. 10 is a diagram illustrating an example of updating the power mode table. [Figure 8] FIG. 10 is a diagram illustrating an example of a learning period and an operation period. [Figure 9] 10 is a flowchart illustrating an example of a procedure for power mode control. DETAILED DESCRIPTION OF THE INVENTION

[0011] The present embodiment will be described below with reference to the drawings. Note that a plurality of embodiments may be combined and implemented. [First embodiment] The control device 10 of the first embodiment controls the power mode of the network device 20. The network device 20 may be called a communication device or a packet processing device, and is, for example, a router or a switching hub (L3 switch or L2 switch). The network device 20 is expected to operate continuously for a long period of time. The control device 10 may be separated from the network device 20. The control device 10 may also be incorporated into the network device 20, or may be the same device as the network device 20. The control device 10 may also be called a computer or an information processing device.

[0012] 1 is a diagram illustrating a control device according to a first embodiment. The control device 10 includes a storage unit 11 and a processing unit 12. The storage unit 11 may be a volatile semiconductor memory such as a random access memory (RAM). Alternatively, the storage unit 11 may be a nonvolatile storage such as a hard disk drive (HDD) or a flash memory.

[0013] The processing unit 12 is, for example, a processor such as a CPU, a GPU (Graphics Processing Unit), or a DSP (Digital Signal Processor). However, the processing unit 12 may also include an electronic circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The processor executes a program stored in a memory such as a RAM (which may be the storage unit 11). A collection of processors may be called a multiprocessor or simply a "processor."

[0014] The storage unit 11 stores past information 13. The past information 13 indicates at least one of the load of the network device 20 measured during a first period and the power mode of the network device 20 determined from the measured load. The first period is a relatively short period having a fixed time span, such as one hour, one day, or one week. The past information 13 may include information about each of a plurality of first periods.

[0015] The load on the network device 20 depends on the amount of data transferred by the network device 20 (e.g., the number of packets transferred per unit time). Examples of the load to be measured include CPU usage, memory usage, the size of a routing table, and traffic on each communication port. Multiple types of load may be measured for the network device 20. The measured load may vary during the first period.

[0016] The network device 20 has multiple power modes with different power consumptions. For example, the operating frequency of the CPU that processes packets differs between the multiple power modes. The higher the operating frequency, the more packets can be processed per unit time, but the higher the power consumption of the network device 20. The lower the operating frequency, the fewer packets can be processed per unit time, but the lower the power consumption of the network device 20.

[0017] For example, the network device 20 has power modes 21, 22, and 23. Power mode 21 is a high-performance mode with high data transfer capability and high power consumption. Power mode 22 is a normal mode with medium data transfer capability and medium power consumption. Power mode 23 is a power-saving mode with low data transfer capability and low power consumption.

[0018] The higher the measured load, the higher the power consumption mode (e.g., power mode 21) is selected. The lower the measured load, the lower the power consumption mode (e.g., power mode 23) is selected. When past information 13 indicates a power mode, that power mode may be the power mode actually used by network device 20 during the first period. Alternatively, that power mode may be an ideal power mode determined by retrospectively analyzing the load during the first period. The power mode may vary during the first period.

[0019] The processing unit 12 uses the past information 13 to calculate an index value indicating the magnitude of load fluctuation during the first period, and determines whether the index value is equal to or greater than a threshold value. If the past information 13 indicates the load, the index value may indicate the range of load fluctuation during the first period, such as the difference between the maximum and minimum load values ​​during the first period. Alternatively, if the past information 13 indicates the power mode, the index value may indicate the number of power mode changes during the first period. A large number of power mode changes reflects large load fluctuations. If there are multiple first periods, the processing unit 12 may determine whether the index value is equal to or greater than a threshold value for each of the multiple first periods.

[0020] The processing unit 12 controls the power mode of the network device 20 for a second period after the first period depending on whether the index value is equal to or greater than the threshold. The second period is preferably a period corresponding to the first period and a period during which the load on the network device 20 is likely to be similar to that of the first period. For example, the second period may be the day after the first period, the same day of the week following the first period, the same day of the month following the first period, or the same day of the year following the first period.

[0021] If the index value is not greater than or equal to the threshold, i.e., if the load fluctuations during the first period are not large, the processing unit 12 determines that the reliability of the load trend during the first period is high. In this case, the processing unit 12 operates the network device 20 during the second period in a power mode identified from the past information 13. For example, the processing unit 12 instructs the network device 20 on an operating mode. If the past information 13 indicates the power mode used during the first period, it can be said that the processing unit 12 reuses the power mode used during the first period during the second period. The processing unit 12 may omit measuring the load of the network device 20 during the second period.

[0022] The processing unit 12 may determine whether the index value is equal to or greater than the threshold value during the first period, during the second period, or between the first and second periods. The processing unit 12 may select an appropriate power mode according to the load during the first period, during the second period, or between the first and second periods.

[0023] On the other hand, if the index value is equal to or greater than the threshold, i.e., if the load fluctuation during the first period is large, the processing unit 12 determines that the reliability of the load trend during the first period is low. In this case, the processing unit 12 measures the load of the network device 20 during the second period. For example, the processing unit 12 instructs the network device 20 to measure the load. Then, the processing unit 12 operates the network device 20 in a power mode determined from the load measured during the second period. For example, the processing unit 12 instructs the network device 20 to select a power mode. It can also be said that the processing unit 12 does not reuse the power mode used during the first period during the second period.

[0024] As described above, the control device 10 of the first embodiment acquires past information 13 indicating a first load of the network device 20 measured during a first period, or a first power mode of the network device 20 determined from the first load among multiple power modes with different power consumptions. The control device 10 uses the past information 13 to determine whether an index value indicating the magnitude of fluctuation in the first load during the first period is equal to or greater than a threshold. If the index value is not equal to or greater than the threshold, the control device 10 operates the network device 20 in the first power mode during a second period following the first period. If the index value is equal to or greater than the threshold, the control device 10 measures a second load of the network device 20 during the second period, and operates the network device 20 in the second power mode determined from the second load among multiple power modes.

[0025] This eliminates the need to constantly monitor the load on the network device 20, reducing the burden of measuring the load and selecting an appropriate power mode. Furthermore, for periods of significant load fluctuation in the past, the load is measured again. This reduces reliance on unreliable load trends and reduces the risk of the network device 20 operating in an inappropriate power mode. As a result, the power consumption of the network device 20 is reduced.

[0026] [Second embodiment] The network device 100 of the second embodiment measures the load of the network device 100 and autonomously determines the power mode. The network device 100 is, for example, a router or an L3 switch. The network device 100 corresponds to a device that functions as both the control device 10 and the network device 20 of the first embodiment. However, the function of determining the power mode may be separated into an external device of the network device 100.

[0027] 2 is a diagram illustrating an example of hardware of a network device according to the second embodiment. The network device 100 includes LAN (Local Area Network) interfaces 101 and 102, a data processing unit 103, a measurement unit 111, a learning information storage unit 115, a power mode storage unit 116, and a power mode control unit 117.

[0028] The LAN interfaces 101 and 102 are wired communication interfaces connected to adjacent devices via LAN cables. While the network device 100 in FIG. 2 has two LAN interfaces, it may have three or more LAN interfaces. Furthermore, the network device 100 may have a communication interface conforming to a communication standard other than LAN, or may have a wireless communication interface.

[0029] The LAN interfaces 101 and 102 typically send and receive idle signals to and from adjacent devices even when there are no packets. By using the idle signals to synchronize with adjacent devices, the LAN interfaces 101 and 102 can quickly start high-speed communication when a packet arrives. However, the LAN interfaces 101 and 102 can also stop the idle signals or reduce the frequency of sending idle signals.

[0030] The data processing unit 103 receives packets from the LAN interfaces 101 and 102 and processes the received packets in accordance with a communication standard. The data processing unit 103 outputs the packets to the LAN interfaces 101 and 102 according to the processing results. The data processing unit 103 includes a CPU 104 and a memory 105.

[0031] The CPU 104 is a processor used for the above-mentioned packet processing. The CPU 104 executes a program for routing. The CPU 104 can operate at different clock frequencies. The memory 105 is a volatile memory such as a RAM. The memory 105 stores the program executed by the CPU 104. For example, the program is loaded into the memory 105 from a non-volatile memory (not shown). The memory 105 also includes a buffer for storing packets waiting to be processed. The memory 105 also stores a routing table including multiple records that associate packet destinations with LAN interfaces to which the packets are output.

[0032] The data processing unit 103 also measures the traffic of each of the LAN interfaces 101 and 102. The traffic is the amount of packets received or transmitted per unit time. The data processing unit 103 also detects adjacent devices connected to the LAN interfaces 101 and 102 and reports the adjacent devices according to the Link Layer Discovery Protocol (LLDP). The data processing unit 103 also detects when the utilization rate of the CPU 104 or the utilization rate of the memory 105 becomes significantly high (for example, when it exceeds 90%). When a high load is detected, the data processing unit 103 outputs an SNMP (Simple Network Management Protocol) trap to warn of the high load.

[0033] The measurement unit 111 collects various measurement values ​​indicating the load of the network device 100 from the data processing unit 103. However, whether or not to continuously collect the measurement values ​​is instructed by the power mode control unit 117. The measurement unit 111 has a CPU usage rate measurement unit 112, a memory usage rate measurement unit 113, and a table size measurement unit 114.

[0034] The CPU utilization rate measurement unit 112 measures the utilization rate of the CPU 104. The memory utilization rate measurement unit 113 measures the utilization rate of the memory 105. The table size measurement unit 114 measures the number of rows (number of records) in the routing table stored in the memory 105 and calculates the increase or decrease from the previous number of rows. The measurement unit 111 also acquires traffic information for each of the LAN interfaces 101 and 102 from the data processing unit 103. The measurement unit 111 also receives information about adjacent devices from the data processing unit 103 in accordance with LLDP. The measurement unit 111 also receives SNMP traps from the data processing unit 103.

[0035] The learning information storage unit 115 is a non-volatile memory that stores setting information and measurement values. For example, a flash memory is used as the learning information storage unit 115. The setting information indicates the measurement interval for each measurement item. The setting information also indicates the correspondence between the measurement value and the power mode. The setting information has a default value. The user can change the setting from this default value. The period for which the learning information storage unit 115 retains the measurement values ​​may depend on the algorithm used by the power mode control unit 117 to select the power mode.

[0036] The power mode storage unit 116 is a non-volatile memory that stores the power mode history. For example, a flash memory is used as the power mode storage unit 116. However, the learning information storage unit 115 and the power mode storage unit 116 may be different storage areas in the same non-volatile memory. The power mode storage unit 116 accumulates the power mode history for the length of the learning period. The learning period is, for example, one year.

[0037] The network device 100 has multiple power modes for each of the LAN interface 101, the LAN interface 102, and the data processing unit 103. The network device 100 can select the power modes of the LAN interface 101, the LAN interface 102, and the data processing unit 103 independently of one another. The power mode of the data processing unit 103 corresponds to the power mode of the entire network device 100.

[0038] There are three overall-level power modes: high-performance mode, normal mode, and power-saving mode. The clock frequency of the CPU 104 differs among the three overall-level power modes, resulting in different power consumption. In the high-performance mode, the clock frequency of the CPU 104 is high, power consumption is high, and the processing capability of the data processing unit 103 is high. In the normal mode, the clock frequency of the CPU 104 is medium, power consumption is medium, and the processing capability of the data processing unit 103 is medium. In the power-saving mode, the clock frequency of the CPU 104 is low, power consumption is low, and the processing capability of the data processing unit 103 is low.

[0039] There are two power modes at the LAN interface level: normal mode and power saving mode. The two power modes at the LAN interface level differ in the presence or absence of idle signals or the frequency of their transmission from the LAN interfaces 101 and 102. In normal mode, idle signals are constantly transmitted. This maintains synchronization with adjacent devices and allows high-speed communication to begin quickly, but consumes a lot of power. In power saving mode, idle signals are not transmitted or are transmitted less frequently. This causes a delay before synchronization with adjacent devices is achieved and high-speed communication begins, but consumes less power.

[0040] The power mode control unit 117 autonomously determines the power mode of the network device 100 using different methods for the learning period and the operation period. Whether the network device 100 operates in the learning period and the length of the learning period are specified by the user.

[0041] During the learning period, the power mode control unit 117 causes the measurement unit 111 to measure the load at the measurement interval indicated by the setting information stored in the learning information storage unit 115. The power mode control unit 117 selects a power mode in real time based on the measurement value and the correspondence indicated by the setting information. The power mode control unit 117 determines an overall power mode based on CPU usage, memory usage, and table size. The power mode control unit 117 also determines a LAN interface power mode based on traffic and the presence or absence of adjacent devices. The power mode control unit 117 notifies the data processing unit 103 of the selected power mode and records the selected power mode in the power mode storage unit 116.

[0042] During the operation period, the power mode control unit 117 reuses the power mode recorded in the power mode storage unit 116. At this time, the power mode control unit 117 does not, as a rule, cause the measurement unit 111 to measure the load. The power mode control unit 117 uses the same power mode as the previous day corresponding to the current day as the power mode for that day. The power mode control unit 117 instructs the data processing unit 103 to use that power mode.

[0043] It is preferable that the corresponding past day is a day that is likely to have a similar load trend to the current day. If the learning period is one week, for example, the corresponding past day is the day of the previous week that falls on the same day of the week as the current day. If the learning period is one month, for example, the corresponding past day is the day of the previous month that falls on the same date as the current day, or the day of the previous month that falls on the same day of the week as the current day and is closest to the current day. If the learning period is one year, for example, the corresponding past day is the day of the previous year that falls on the same day of the week as the current day and is closest to the current day. The corresponding past day may also be specified by the user.

[0044] However, depending on the day, there may be a high risk that the load trend will not be similar to that of the corresponding past day. Therefore, even if the day is within the operation period, on such a high-risk day, the power mode control unit 117 causes the measurement unit 111 to measure the load again without reusing the past power mode. If the load fluctuation on the corresponding past day was large, the power mode control unit 117 determines that there is a high risk in reusing the past power mode. In this case, the power mode control unit 117 selects a power mode using the measured load on that day. If a day on which the load trend will be different from normal (for example, a special day on which communication volume is clearly different from the previous year, such as a service start date or service end date) is known in advance, such a day may be specified by the user.

[0045] The power mode control unit 117 determines whether to reuse a past power mode for each of the overall level and the LAN interface level, independently of each other. However, if the load fluctuation on the corresponding past day was large for at least one of the overall level and the LAN interface level, the power mode control unit 117 may avoid reusing all of the power modes for the overall level and the LAN interface level.

[0046] Furthermore, even on days when a previous power mode is to be reused, if an SNMP trap indicating a high load is received, the power mode control unit 117 forcibly changes the overall power mode to high performance mode. At this time, the power mode control unit 117 may also change the power mode of all LAN interfaces to normal mode.

[0047] To perform the above-described power mode control, the power mode control unit 117 has a CPU 118 and a memory 119. The CPU 118 is a processor used for power mode control. The CPU 118 executes a program for power mode control. The memory 119 is a volatile memory such as a RAM. The memory 119 stores the program executed by the CPU 118. For example, the program is loaded into the memory 119 from a non-volatile memory (e.g., the learning information storage unit 115, the power mode storage unit 116, or another non-volatile memory). The learning information storage unit 115, the power mode storage unit 116, or the memory 119 corresponds to the storage unit 11 of the first embodiment. The CPU 118 corresponds to the processing unit 12 of the first embodiment.

[0048] As will be described later, a terminal device 200 of a user who manages the network device 100 may access the network device 100 via a network. The terminal device 200 can read setting information, measurement values, and power mode history from the network device 100. The read information may be stored in non-volatile storage and may be displayed on a display device. The terminal device 200 can also rewrite setting information held by the network device 100. The terminal device 200 can also transmit a routing program and a power mode control program to the network device 100.

[0049] 3 is a diagram illustrating an example of hardware of a terminal device according to the second embodiment. The terminal device 200 includes a CPU 201, a RAM 202, an HDD 203, a GPU 204, an input interface 205, a medium reader 206, and a communication interface 207, all of which are connected to a bus.

[0050] The CPU 201 is a processor that executes programs. The CPU 201 loads the programs and data stored in the HDD 203 into the RAM 202 and executes the programs. The RAM 202 is a volatile semiconductor memory that temporarily stores the programs executed by the CPU 201 and data used in calculations by the CPU 201. The terminal device 200 may have a type of volatile memory other than RAM.

[0051] The HDD 203 is a non-volatile storage that stores software programs such as an operating system (OS), middleware, and application software, as well as data. The terminal device 200 may also have other types of non-volatile storage, such as a flash memory or an SSD (Solid State Drive).

[0052] The GPU 204 performs image processing in cooperation with the CPU 201 and displays images on a display device 211 connected to the terminal device 200. The display device 211 is, for example, a CRT (Cathode Ray Tube) display, a liquid crystal display, or an organic EL (Electro Luminescence) display. Other types of output devices may also be connected to the terminal device 200.

[0053] The input interface 205 receives an input signal from an input device 212 connected to the terminal device 200. The input device 212 is, for example, a mouse, a touch panel, or a keyboard. A plurality of input devices may be connected to the terminal device 200.

[0054] The medium reader 206 is a reading device that reads programs and data recorded on the recording medium 213. The recording medium 213 is, for example, a magnetic disk, an optical disk, or a semiconductor memory. Magnetic disks include flexible disks (FDs) and HDDs. Optical disks include compact discs (CDs) and digital versatile discs (DVDs). The medium reader 206 copies the programs and data read from the recording medium 213 to other recording media such as the RAM 202 or the HDD 203. The read programs may be executed by the CPU 201 and may be transmitted to the network device 100.

[0055] The recording medium 213 may be a portable recording medium. The recording medium 213 may be used to distribute programs and data. The recording medium 213 and other non-volatile storage may also be referred to as a computer-readable recording medium.

[0056] The communication interface 207 communicates with the network device 100 via the network 214. The communication interface 207 may be a wired communication interface connected to a wired communication device such as a switch or a router, or may be a wireless communication interface connected to a wireless communication device such as a base station or an access point.

[0057] 4 is a diagram showing an example of a setting table. The learning information storage unit 115 stores setting tables 121 and 122. The setting table 121 indicates the settings for power mode control at an overall level. The setting table 121 associates measurement items, measurement intervals, numerical conditions, and power modes. The measurement items indicate the load related to the data processing unit 103, and include CPU usage, memory usage, and table size.

[0058] The default value for the CPU usage measurement interval is 5 minutes. The default value for the memory usage measurement interval is 5 minutes. The default value for the table size measurement interval is 30 minutes. The user can change these measurement intervals. The table size measurement value indicates the increase or decrease in the number of rows in the routing table since the previous measurement time.

[0059] The numerical condition for CPU usage is divided into three ranges. The higher the CPU usage, the higher the processing power and power consumption of the selected power mode. For example, if CPU usage exceeds 70% three times in a row, high performance mode is selected. If CPU usage is between 30% and 70%, or exceeds 70% two or fewer times in a row, normal mode is selected. If CPU usage is below 30%, power saving mode is selected.

[0060] The memory usage numerical condition divides memory usage into three ranges. The higher the memory usage, the higher the processing power and power consumption of the selected power mode. For example, if memory usage exceeds 70% three times in a row, high performance mode is selected. If memory usage is between 30% and 70%, or exceeds 70% two or fewer times in a row, normal mode is selected. If memory usage is less than 30%, power saving mode is selected.

[0061] The numerical conditions for table size divide the number of rows that have increased or decreased into three ranges. The greater the number of rows that have increased or decreased, the higher the processing power and power consumption of the selected power mode. For example, if the number of rows that have increased or decreased is 10 or more, high performance mode is selected. If the number of rows that have increased or decreased is between 5 and 9, normal mode is selected. If the number of rows that have increased or decreased is 4 or less, or if there is no increase / decrease (no change), power saving mode is selected. The user can change these numerical conditions.

[0062] When the measurement values ​​of different measurement items correspond to different power modes, the power mode control unit 117 selects the power mode with the highest processing performance from among the different power modes. For example, when the CPU usage rate corresponds to the normal mode and the memory usage rate and table size correspond to the power saving mode, the power mode control unit 117 selects the normal mode.

[0063] The setting table 122 indicates the settings for power mode control at the LAN interface level. The setting table 122 associates measurement items, measurement intervals, numerical conditions, and power modes. The measurement items indicate the load related to the LAN interfaces 101 and 102, and include traffic and the presence or absence of adjacent devices.

[0064] The default value for the traffic measurement interval is 5 minutes. The default value for the measurement interval for the presence or absence of adjacent devices is 30 minutes. The user can change these measurement intervals. The traffic numerical condition divides the traffic into two sections. The more traffic there is, the higher the power mode with higher processing capacity and power consumption is selected. For example, if the traffic is 20% or more of the bandwidth, normal mode is selected. If the traffic is less than 20% of the bandwidth, power saving mode is selected. The user can change these numerical conditions.

[0065] For neighboring devices, if there are neighboring devices, normal mode or power saving mode is selected according to traffic. Neighboring devices are other network devices or terminal devices. If there are no neighboring devices, the LAN interface is shut down. In the shutdown state, control signals to detect neighboring devices are transmitted, but packet signals are not transmitted. The shutdown state is released when neighboring devices are detected.

[0066] FIG. 5 is a diagram showing an example of a measurement value table. The learning information storage unit 115 stores measurement value tables 123 and 124. The measurement value table 123 stores measurement values ​​related to overall-level power mode control. The measurement value table 123 associates time, CPU usage rate, memory usage rate, and table size. At each time, depending on the measurement interval, some of the CPU usage rate, memory usage rate, and table size may not be measured. For example, the measurement frequency of table size is one-sixth of the CPU usage rate and memory usage rate.

[0067] The measurement value table 124 stores measurement values ​​related to power mode control at the LAN interface level. The measurement value table 124 associates time, interface, traffic, and the presence or absence of adjacent devices. The interfaces included in the measurement value table 124 identify the LAN interfaces 101 and 102. At each time, depending on the measurement interval, either traffic or the presence or absence of adjacent devices may not be measured. For example, the measurement frequency for the presence or absence of adjacent devices is one-sixth of that for traffic.

[0068] For example, the measurement value table 123 indicates that at 0:00 on January 1, 2023, the CPU utilization rate of the data processing unit 103 is 20%, the memory utilization rate is 25%, and the table size has not changed since the previous measurement. Also, for example, the measurement value table 124 indicates that at 0:00 on January 1, 2023, the traffic of the LAN interface 101 is 15%, and that there is an adjacent device connected to the LAN interface 101.

[0069] The extent to which measurement value tables 123, 124 store old measurement values ​​may depend on the method used by power mode control unit 117 to determine the magnitude of load fluctuations. When the magnitude of load fluctuations for each day is determined from measurement values ​​after a learning period has elapsed, measurement value tables 123, 124 may store measurement values ​​for the learning period. When the magnitude of load fluctuations for each day is determined from measurement values ​​at the end of each day, measurement value tables 123, 124 may store only measurement values ​​for one day. When the magnitude of load fluctuations for each day is determined from the power mode, measurement value tables 123, 124 may store only measurement values ​​for an even shorter period.

[0070] When a load is measured exceptionally during an operation period, the power mode control unit 117 may save the measurement value of that day in the measurement value tables 123, 124. At this time, if the measurement value of a past day corresponding to that day is stored in the measurement value tables 123, 124, the power mode control unit 117 may delete the measurement value of that past day from the measurement value tables 123, 124. For example, when a load is measured exceptionally on January 1st during an operation period, the power mode control unit 117 may overwrite the measurement value of January 1st one year ago with the measurement value of this year.

[0071] 6 is a diagram showing an example of a power mode table. The power mode storage unit 116 stores power mode tables 125 and 126. The power mode table 125 stores the results of overall-level power mode selection. The power mode table 125 associates a date, time, power mode, and fluctuation flag. When an overall-level power mode is newly set or changed, the time and power mode are recorded in the power mode table 125.

[0072] The power mode table 126 stores the results of power mode selection at the LAN interface level. The power mode table 126 associates the date, interface, time, power mode, and fluctuation flag. When the power mode of a LAN interface is newly set or changed, the time and power mode are recorded in the power mode table 126.

[0073] For example, the power mode table 125 indicates that the power mode of the data processing unit 103 changed as follows on January 1, 2023: power saving mode from midnight to 10:00, high performance mode from 10:00 to 10:15, normal mode from 10:15 to 12:00, power saving mode from 12:00 to 19:00, high performance mode from 19:00 to 19:20, and normal mode from 19:20 onwards.Furthermore, for example, the power mode table 126 indicates that the power mode of the LAN interface 101 changed as follows on January 1, 2023: power saving mode from midnight to 10:00, normal mode from 10:00 to 12:00, power saving mode from 12:00 to 19:00, and normal mode from 19:00 onwards.

[0074] Power mode tables 125, 126 store the power mode selection results for the length of the learning period. If the learning period is one year, power mode tables 125, 126 store the power mode selection results for one year. During the operation period, as a general rule, the power modes indicated in power mode tables 125, 126 are reused.

[0075] For example, if the power mode of January 1, 2023 is reused on January 1, 2024, the data processing unit 103 operates according to the following schedule: power saving mode from midnight to 10:00, high performance mode from 10:00 to 10:15, normal mode from 10:15 to 12:00, power saving mode from 12:00 to 19:00, high performance mode from 19:00 to 19:20, and normal mode from 19:20 onwards. Also, the LAN interface 101 operates according to the following schedule: power saving mode from midnight to 10:00, normal mode from 10:00 to 12:00, power saving mode from 12:00 to 19:00, and normal mode from 19:00 onwards. In this case, load measurement is omitted on January 1, 2024.

[0076] If the load is measured exceptionally during the operation period, the power mode control unit 117 may record the power mode selected on that day in the power mode tables 125, 126. At this time, the power mode control unit 117 may delete the power mode of the previous day corresponding to that day from the power mode tables 125, 126. For example, if the load is measured exceptionally on January 1st during the operation period, the power mode control unit 117 may overwrite the power mode of January 1st from one year ago with the power mode of this year. Furthermore, because the power mode of a day with large load fluctuations is not reused, the power mode tables 125, 126 do not need to store the power mode of that day.

[0077] The fluctuation flag indicates whether or not the load fluctuation is large on that day. When the fluctuation flag is ON, it indicates that the load fluctuation on that day is large. When the fluctuation flag is OFF, it indicates that the load fluctuation on that day is not large. The fluctuation flag is determined after the end of one day. The power mode control unit 117 may determine the fluctuation flag for each day at the end of each day during the learning period. Alternatively, the power mode control unit 117 may determine the fluctuation flag for all days collectively after the end of the learning period. Alternatively, the power mode control unit 117 may determine the fluctuation flag for the corresponding past day for each day during the operation period.

[0078] There are several possible methods for determining the fluctuation flag. For example, the power mode control unit 117 scans the power mode selection results for a day and calculates the number of power mode changes. If the number of power mode changes is equal to or greater than a threshold (e.g., three times), the power mode control unit 117 determines that the day is one with a large load fluctuation. On the other hand, if the number of power mode changes is less than the threshold, the power mode control unit 117 determines that the day is one with a small load fluctuation.

[0079] Furthermore, for example, the power mode control unit 117 scans the power mode selection results for one day and detects a certain pattern of power mode changes. The certain patterns include a pattern in which the power mode decreases in the order of high performance mode, normal mode, and power saving mode, and a pattern in which the power mode increases in the order of power saving mode, normal mode, and high performance mode. If any of the above patterns is detected, the power mode control unit 117 determines that the day is one with a large load fluctuation. On the other hand, if none of the above patterns is detected, the power mode control unit 117 determines that the day is one with a small load fluctuation.

[0080] Furthermore, for example, the power mode control unit 117 scans the measurement values ​​for one day and calculates the fluctuation range of the measurement values ​​(the difference between the maximum and minimum values). For example, the CPU usage rate or memory usage rate is used as the measurement value. If the fluctuation range is equal to or greater than a threshold value (for example, 30%), the power mode control unit 117 determines that the day is one with a large load fluctuation. On the other hand, if the fluctuation range is less than the threshold value, the power mode control unit 117 determines that the day is one with a small load fluctuation.

[0081] Also, for example, the power mode control unit 117 scans the table size for one day. If the table size fluctuates by five or more rows in one day, the power mode control unit 117 determines that the day is one with a large load fluctuation. On the other hand, if the table size does not fluctuate by five or more rows in one day, the power mode control unit 117 determines that the day is one with a small load fluctuation.

[0082] As a result, the fluctuation flag for each day is determined to be ON or OFF. For example, the number of power mode changes on January 1, 2023 is five, which is greater than the threshold value of three. Therefore, the power mode control unit 117 determines that the fluctuation flag for January 1, 2023 is ON. On the other hand, the power mode control unit 117 determines that the fluctuation flag for January 2, 2023 is OFF.

[0083] 7 is a diagram showing an example of updating the power mode table. If the fluctuation flag for January 1, 2023 is ON, the power mode control unit 117 causes the measurement unit 111 to measure the load on January 1, 2024 in accordance with the setting tables 121 and 122. The power mode control unit 117 selects a power mode for January 1, 2024 using the measurement value.

[0084] For example, the power mode of the data processing unit 103 is selected as follows: power saving mode from midnight to 10:00, high performance mode from 10:00 to 10:15, normal mode from 10:15 to 12:00, power saving mode from 12:00 to 18:00, high performance mode from 18:00 to 19:20, and normal mode from 19:20 onwards.

[0085] The power mode control unit 117 updates the power mode table 125 to power mode table 125a. In power mode table 125a, the power mode selection result for January 1, 2023 is overwritten with the power mode selection result for January 1, 2024. The power mode control unit 117 will determine the power mode of the data processing unit 103 for January 1, 2025, one year from now, by referring to the information for January 1, 2024.

[0086] Also, for example, the power mode of the LAN interface 101 is selected as follows: power saving mode from midnight to 10:00, normal mode from 10:00 to 12:00, power saving mode from 12:00 to 18:00, and normal mode from 18:00 onwards.

[0087] The power mode control unit 117 updates the power mode table 126 to power mode table 126a. In power mode table 126a, the power mode selection result for January 1, 2023 is overwritten with the power mode selection result for January 1, 2024. The power mode control unit 117 will determine the power mode of the LAN interface 101 for January 1, 2025, one year from now, by referring to the information for January 1, 2024.

[0088] FIG. 8 is a diagram showing examples of a learning period and an operation period. Here, the learning period 127 is one year starting on January 1, 2023. The learning period 127 is specified by the user. The operation period 128 is the period immediately following the learning period 127 and starts on January 1, 2024. The operation period 128 continues unless the user instructs re-learning.

[0089] The power mode control unit 117 references the power mode records for January 2023 to determine the power mode for January 2024. For example, the power mode control unit 117 generally changes the power mode for each day in January 2024 to match the records for that day in January 2023. Similarly, the power mode control unit 117 references the power mode records for February 2023 to determine the power mode for February 2024. Furthermore, the power mode control unit 117 references the power mode records for December 2023 to determine the power mode for December 2024.

[0090] 9 is a flowchart showing an example of a procedure for power mode control. Here, a description will be given of the processing procedure of the network device 100. The network device 100 repeatedly executes the following steps S10 to S18 at measurement intervals, for example.

[0091] In step S10, power mode control unit 117 determines whether the current period is the learning period. If the current period is the learning period, the process proceeds to step S15. If the current period is not the learning period but the operation period, the process proceeds to step S11.

[0092] In step S11, the power mode control unit 117 determines whether the fluctuation flag for the past corresponding day is ON. The corresponding day is determined based on the length of the learning period, such as the same day of the week one week ago, the same day one month ago, or the same day one year ago. If the fluctuation flag is ON, this indicates that the load fluctuation on the corresponding day was large. The power mode control unit 117 may also determine the fluctuation flag from the number of power mode changes on the corresponding day. If the fluctuation flag is ON, processing proceeds to step S15. If the fluctuation flag is OFF, processing proceeds to step S12.

[0093] In step S12, power mode control unit 117 determines whether an SNMP trap indicating a high load has been received within the most recent fixed time period. This SNMP trap indicates that the current CPU usage rate or memory usage rate of data processing unit 103 is significantly high. If an SNMP trap indicating a high load has been received, processing proceeds to step S13. If an SNMP trap indicating a high load has not been received, processing proceeds to step S14.

[0094] In step S13, the power mode control unit 117 selects the high performance mode as the power mode of the data processing unit 103. Additionally, the power mode control unit 117 may select the normal mode as the power mode of the LAN interfaces 101 and 102. If this changes the power mode of at least one of the LAN interfaces 101 and 102 and the data processing unit 103, the power mode control unit 117 notifies the data processing unit 103 of the change in power mode. Then, the process proceeds to step S17.

[0095] In step S14, the power mode control unit 117 reuses the power mode that was selected at the same time on the corresponding date. The power mode control unit 117 may reuse only the LAN interfaces 101, 102 and the data processing unit 103 whose fluctuation flags are OFF. If this changes the power mode of at least one of the LAN interfaces 101, 102 and the data processing unit 103, the power mode control unit 117 notifies the data processing unit 103 of the change in power mode. Then, processing proceeds to step S17.

[0096] In step S15, the measurement unit 111 measures the load on the network device 100. During the operation period, the measurement unit 111 may measure only the load associated with the LAN interfaces 101, 102 and the data processing unit 103 whose fluctuation flags are ON.

[0097] In step S16, the power mode control unit 117 determines the power mode from the measurement values ​​in step S15. During the operation period, the power mode control unit 117 may re-determine the power mode only for those LAN interfaces 101, 102 and data processing unit 103 whose fluctuation flags are ON. If this results in a change in the power mode of at least one of the LAN interfaces 101, 102 and data processing unit 103, the power mode control unit 117 notifies the data processing unit 103 of the change in power mode.

[0098] In step S17, the power mode control unit 117 records the power mode determined in step S13, step S14, or step S16. In step S18, the power mode control unit 117 determines whether the load fluctuations within a day are large. For example, the power mode control unit 117 determines whether the number of power mode changes has reached a threshold value for each of the LAN interfaces 101 and 102 and the data processing unit 103. The power mode control unit 117 records a fluctuation flag indicating whether the load fluctuations are large.

[0099] As described above, the network device 100 of the second embodiment autonomously changes its power mode depending on the load during the learning period. This reduces the power consumption of the network device 100. Furthermore, the network device 100 sets power modes separately for the entire network and the LAN interface. This makes it easier to reduce power consumption compared to changing the power modes of all components at once.

[0100] After the learning period has elapsed, the network device 100 determines the power mode schedule by referencing past power mode performance. This eliminates the need to constantly monitor the load after the learning period has elapsed, reducing the burden of power mode control and reducing power consumption.

[0101] Furthermore, even after the learning period has elapsed, the network device 100 measures the load again and reselects a power mode on days when the load fluctuated significantly in the past. This evaluates the risk that the load trend will deviate from past performance and reduces the risk of setting a power mode that is incompatible with the current load. As a result, power consumption is reduced.

[0102] Furthermore, even on days when a previous power mode is reused, if a significant increase in load is detected by an existing communication control protocol, the network device 100 is forced to operate in high-performance mode, thereby preventing a decrease in communication performance. [Explanation of symbols]

[0103] 10 Control device 11 Storage section 12 Processing section 13 Past information 20 Network Equipment 21, 22, 23 Power Mode

Claims

1. a storage unit that stores past information indicating a first load of a network device measured during a first period or a first power mode of the network device determined from the first load among a plurality of power modes with different power consumption; a processing unit that uses the past information to determine whether an index value indicating a magnitude of fluctuation in the first load during the first time period is equal to or greater than a threshold, and if the index value is not equal to or greater than the threshold, operates the network device in the first power mode during a second time period after the first time period, and if the index value is equal to or greater than the threshold, measures a second load on the network device during the second time period and operates the network device in a second power mode among the plurality of power modes that is determined based on the second load; A control device having:

2. the historical information indicating the first power mode used during the first time period; the index value indicates the number of times the first power mode changes during the first period. The control device according to claim 1 .

3. the processing unit identifies the first power mode from the past information when the index value is not equal to or greater than the threshold value. The control device according to claim 1 .

4. The first period is plural, the second period is plural, the processing unit determines whether the index value is equal to or greater than the threshold value for each of the plurality of first periods, and uses, for each of the plurality of second periods, a determination result for a corresponding first period among the plurality of first periods; The control device according to claim 1 .

5. determining whether an index value indicating a magnitude of fluctuation in the first load during the first time period is equal to or greater than a threshold value, using past information indicating a first load of the network device measured during the first time period, or a first power mode of the network device determined from the first load among a plurality of power modes with different power consumption; if the index value is not greater than or equal to the threshold, operating the network device in the first power mode for a second time period that is after the first time period; if the index value is equal to or greater than the threshold, measuring a second load on the network device during the second time period, and operating the network device in a second power mode among the plurality of power modes that is determined based on the second load; A control method for computer-implemented processing.

6. determining whether an index value indicating a magnitude of fluctuation in the first load during the first time period is equal to or greater than a threshold value, using past information indicating a first load of the network device measured during the first time period, or a first power mode of the network device determined from the first load among a plurality of power modes with different power consumption; if the index value is not greater than or equal to the threshold, operating the network device in the first power mode for a second time period that is after the first time period; if the index value is equal to or greater than the threshold, measuring a second load on the network device during the second time period, and operating the network device in a second power mode among the plurality of power modes that is determined based on the second load; A control program that causes a computer to execute a process.

Images