WO2023155904A1 - 一种网络设备运行状态的调整方法、装置及相关设备 - Google Patents

一种网络设备运行状态的调整方法、装置及相关设备 Download PDF

Info

Publication number
WO2023155904A1
WO2023155904A1 PCT/CN2023/077060 CN2023077060W WO2023155904A1 WO 2023155904 A1 WO2023155904 A1 WO 2023155904A1 CN 2023077060 W CN2023077060 W CN 2023077060W WO 2023155904 A1 WO2023155904 A1 WO 2023155904A1
Authority
WO
WIPO (PCT)
Prior art keywords
network device
energy
energy consumption
saving
traffic information
Prior art date
Application number
PCT/CN2023/077060
Other languages
English (en)
French (fr)
Inventor
吴俊�
张�杰
张亮
徐晓东
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2023155904A1 publication Critical patent/WO2023155904A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/02—Details
    • H04L12/12—Arrangements for remote connection or disconnection of substations or of equipment thereof
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0893—Assignment of logical groups to network elements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0894—Policy-based network configuration management
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/147—Network analysis or design for predicting network behaviour
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/02—Power saving arrangements
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • the present application relates to the technical field of communications, and in particular to a method, device and related equipment for adjusting the operating state of network equipment.
  • the present application provides a method and device for adjusting the operating state of network equipment and related equipment, so as to reduce the energy consumption of the network equipment by adjusting the operating parameters of the network equipment.
  • the present application provides a method for adjusting an operating state of a network device.
  • the first network device receives the first flow information sent by the second network device, where the first flow information indicates the value of the flow processed by the second network device within the first time period.
  • the first network device predicts the second traffic information corresponding to the second network device within the second time period according to the first traffic information.
  • the second time period is later than the first time period.
  • the first network device determines the energy consumption value corresponding to each energy-saving strategy in the multiple energy-saving strategies according to the second traffic information.
  • Each energy-saving policy includes configuration parameters corresponding to devices in the second network device.
  • the first network device determines the energy-saving strategy corresponding to the energy consumption value that meets the preset condition as the target energy-saving strategy, and sends the target energy-saving strategy to the second network device, so that the second network device follows the configuration parameters in the target energy-saving strategy run.
  • the first network device determines the target energy-saving strategy matching the flow information by predicting the second flow information that the second network device needs to process within the second time period, and sends the target energy-saving strategy to the second network device, so that the second network device
  • the second network device configures the parameters of the internal device according to the target energy saving strategy matched with the traffic to be processed. Therefore, this solution enables the second network device to consume low energy consumption during the low traffic time period, reducing the energy consumption generated by the network device.
  • the preset condition includes that the energy consumption value corresponding to the target energy-saving strategy is the minimum value among the multiple energy consumption values corresponding to the multiple energy-saving strategies, or the energy consumption value corresponding to the target energy-saving strategy is less than or equal to the energy consumption threshold.
  • the first network device may also determine the planned execution time corresponding to the target energy saving policy according to the second traffic information and the target energy saving policy, and send the planned execution time to the second network device, so that the second network device The second network device executes the target energy saving policy within the time period corresponding to the planned execution time.
  • the first network device may further determine a wakeup value according to the target energy saving policy, where the wakeup value indicates a condition for the second network device to terminate execution of the target energy saving policy.
  • the wake-up value may include a traffic threshold and/or a performance threshold. When the wake-up value includes a traffic threshold and the traffic processed by the second network device exceeds the traffic threshold, execution of the target energy-saving policy is terminated; when the wake-up value includes a performance threshold and the second network device When the performance value of the device for processing traffic exceeds the performance threshold, the execution of the target energy saving policy is terminated.
  • the traffic threshold may include a throughput threshold
  • the performance threshold may include a delay threshold, a jitter threshold, a packet loss rate threshold, and the like.
  • the first network device may use a traffic prediction model to predict traffic information corresponding to the first network device within the second time period. Specifically, the first network device inputs the first traffic information into the traffic prediction model, so as to obtain the second traffic information output by the traffic prediction model.
  • the traffic prediction model is trained and generated according to historical traffic information of the second network device.
  • the first network device may use an energy consumption prediction model to determine the energy consumption corresponding to each energy saving strategy. Specifically, for each energy saving strategy, the first network device inputs the second traffic information and configuration parameters corresponding to the energy saving strategy into the energy consumption prediction model, so as to obtain the energy consumption corresponding to the energy saving strategy output by the energy consumption prediction model.
  • the energy consumption prediction model is generated according to the training samples. Each training sample includes flow information, energy consumption value, and configuration parameters corresponding to the energy consumption value.
  • the energy consumption prediction model corresponds to the device type, and before the first network device uses the energy consumption prediction model to determine the energy consumption corresponding to the energy saving strategy, it can determine the energy consumption prediction corresponding to the device type according to the device type of the second network device Model.
  • the device types are the same, for example, the models of the network devices are the same.
  • the first network device may also receive the first The actual execution time of the target energy saving policy sent by the network device.
  • the first network device updates the target energy saving policy according to the actual execution time.
  • the execution duration corresponding to the actual execution time is shorter than the execution duration corresponding to the planned execution time.
  • the first network device may further receive third flow information sent by the second network device, where a statistical value corresponding to the third flow information exceeds the wakeup value.
  • the first network device optimizes the traffic prediction model according to the third traffic information. That is, when the second network device causes the statistical value to exceed the wake-up value due to the emergence of burst traffic information, it may send the burst traffic information to the first network device, so that the first network device optimizes the traffic prediction model according to the burst traffic information, Improve the accuracy of traffic forecasting models.
  • the first network device sends a local training sample to the third network device and receives the energy consumption prediction model sent by the third network device.
  • the local training samples include historical traffic information, historical energy consumption values, and configuration parameters corresponding to the historical energy consumption values of network devices managed by the first network device.
  • the energy consumption prediction model is trained and generated by the third network device using local training samples.
  • the first network device is a controller
  • the third network device is a cloud device.
  • the first network device uses local training samples to generate the first energy consumption prediction model, sends the model parameters of the first energy consumption prediction model to the third network device, and receives the data sent by the third network device.
  • Energy consumption prediction model is determined by the third network device according to model parameters of the first energy consumption prediction model respectively sent by the plurality of first network devices.
  • the local training samples include historical traffic information, historical energy consumption values, and configuration parameters corresponding to the historical energy consumption values of network devices managed by the first network device.
  • each first network device first uses local training samples to perform preliminary training to obtain model parameters of the first energy consumption prediction model, and sends the obtained model parameters to the third network device.
  • the third network device determines the final model parameters corresponding to the energy consumption prediction model by using the model parameters sent by the multiple first network devices, so as to not only improve the training efficiency, but also improve the accuracy of the energy consumption prediction model.
  • the present application provides a method for adjusting the operating state of a network device.
  • the second network device sends first traffic information to the first network device, where the first traffic information indicates the traffic processed by the second network device within the first time period value.
  • the second network device receives the target energy saving policy sent by the first network device, and applies configuration parameters corresponding to the target energy saving policy.
  • the target energy saving policy is determined by the first network device according to the second traffic information.
  • the second flow information is the flow corresponding to the second network device within a second time period predicted by the first network device according to the first flow information, and the second time period is later than the first time period.
  • the target energy saving strategy includes configuration parameters corresponding to devices in the second network device, and the energy consumption corresponding to the target energy saving strategy satisfies a preset condition.
  • the preset condition includes that the energy consumption value corresponding to the target energy-saving strategy is the minimum value among the multiple energy consumption values corresponding to the multiple energy-saving strategies, or the energy consumption value corresponding to the target energy-saving strategy is less than or equal to the energy consumption threshold.
  • the second network device may also receive the plan execution time sent by the first network device.
  • the planned execution time indicates the execution time of the target energy saving strategy and is determined by the first network device according to the second traffic information and the target energy saving strategy.
  • the second network device may also receive the wakeup value sent by the first network device.
  • the wakeup value indicates a condition for the second network device to terminate execution of the target energy saving policy.
  • the wakeup value is determined by the first network device according to the target energy saving policy.
  • the second network device during execution of the target energy saving policy by the second network device, when the statistical value of the second network device exceeds the wakeup value, the second network device terminates execution of the target energy saving policy.
  • the statistical value of the second network device exceeds the wake-up value, it indicates that the current traffic processed by the second network device does not match the predicted second traffic information, and a burst traffic occurs.
  • the second network device may terminate execution of the target energy saving policy in advance.
  • the second network device when the second network device executes the target energy saving policy, it may collect statistics on its corresponding transmission performance value. When the transmission performance value is greater than the transmission performance threshold, the second network device terminates execution of the target energy saving policy.
  • the second network device sends the actual execution time of the target energy saving policy to the first network device, so that the first network device can learn that the second network device terminates execution of the target energy saving policy in advance.
  • the execution duration corresponding to the actual execution time is shorter than the execution duration corresponding to the planned execution time.
  • the second network device may also send third traffic information to the first network device, so that the first network device uses the third traffic information to optimize the traffic prediction model and improve the prediction of the traffic prediction model accuracy.
  • the third flow information indicates the flow that causes the termination of the target energy-saving policy, that is, the burst flow.
  • the present application provides a network system.
  • the system includes a first network device and a second network device.
  • the first network device is configured to execute the method described in the first aspect or any possible implementation manner of the first aspect.
  • the second network device is configured to execute the method described in the second aspect or any possible implementation manner of the second aspect.
  • the network system further includes a third network device.
  • the third network device is configured to receive the local training samples sent by the first network device, train and generate an energy consumption prediction model according to the local training samples, and send the energy consumption prediction model to the first network device.
  • the local training samples include historical traffic information, historical energy consumption values, and configuration parameters corresponding to the historical energy consumption values of network devices managed by the first network device.
  • the present application provides a device for adjusting the running state of network equipment.
  • the device is applied to the first network equipment, and includes a receiving unit, a predicting unit, a determining unit and a sending unit.
  • the receiving unit is configured to receive the first flow information sent by the second network device.
  • the first traffic information indicates the value of the traffic processed by the second network device within the first time period.
  • the predicting unit is configured to predict the second traffic information corresponding to the second network device within the second time period according to the first traffic information, The second time period is later than the first time period.
  • the determining unit is configured to determine an energy consumption value corresponding to each energy-saving strategy in the plurality of energy-saving strategies according to the second flow information.
  • Each energy-saving strategy includes configuration parameters corresponding to components in the second network device running according to the energy-saving strategy.
  • the determining unit is further configured to determine an energy-saving strategy corresponding to an energy consumption value satisfying a preset condition as a target energy-saving strategy.
  • the sending unit is configured to send the target energy saving policy to the second network device, so that the second network device operates according to the configuration parameters in the target energy saving policy.
  • the determining unit is further configured to determine a planned execution time corresponding to the target energy saving strategy according to the second traffic information and the target energy saving strategy.
  • the sending unit is further configured to send the planned execution time to the second network device, so that the second network device executes the target energy saving policy within a time period corresponding to the planned execution time.
  • the determining unit is further configured to determine the wake-up value according to the target energy-saving policy.
  • the wakeup value indicates a condition for the second network device to terminate execution of the target energy saving policy.
  • the sending unit is also configured to send the wakeup value to the second network device.
  • the prediction unit is configured to input the first flow information into the flow prediction model, so as to obtain the second flow information output by the flow prediction model.
  • the traffic prediction model is trained and generated according to historical traffic information of the second network device.
  • the determining unit is configured to, for each energy-saving strategy, input the second flow information and configuration parameters corresponding to the energy-saving strategy into the energy consumption prediction model, so as to obtain the corresponding The energy consumption of the energy saving strategy.
  • the energy consumption prediction model is generated based on training samples. Each training sample includes flow information, energy consumption value and configuration parameters corresponding to the energy consumption value.
  • the determining unit is further configured to, before inputting the configuration parameters corresponding to the second traffic information and the energy saving strategy into the energy consumption prediction model, determine the energy consumption prediction according to the device type of the second network device Model.
  • the device type corresponds to the energy consumption prediction model.
  • the device further includes an updating unit.
  • the receiving unit is further configured to receive the actual execution time of the target energy saving policy sent by the second network device.
  • the updating unit is used for updating the target energy-saving policy according to the actual execution time.
  • the execution duration corresponding to the actual execution time is shorter than the execution duration corresponding to the planned execution time.
  • the device further includes an optimization unit.
  • the receiving unit is further configured to receive third flow information sent by the second network device.
  • the optimization unit is configured to optimize the traffic prediction model according to the third traffic information. The statistical value corresponding to the third flow information exceeds the wake-up value.
  • the sending unit is further configured to send the local training sample to the third network device.
  • the receiving unit is further configured to receive the energy consumption prediction model sent by the third network device.
  • the energy consumption prediction model is trained and generated by the third network device using local training samples.
  • the local training samples include historical traffic information, historical energy consumption values, and configuration parameters corresponding to the historical energy consumption values of network devices managed by the first network device.
  • the device further includes a generating unit.
  • the generation unit is used to generate a first energy consumption prediction model by using local training samples.
  • the sending unit is further configured to send the model parameters of the first energy consumption prediction model to the third network device.
  • the receiving unit is further configured to receive the energy consumption prediction model sent by the third network device.
  • the energy consumption prediction model is determined by the third network device according to model parameters of the first energy consumption prediction model respectively sent by multiple first network devices.
  • the local training samples include historical traffic information, historical energy consumption values, and configuration parameters corresponding to the historical energy consumption values of network devices managed by the first network device.
  • the preset condition includes that the energy consumption value corresponding to the target energy-saving strategy is the minimum value among the multiple energy consumption values corresponding to the multiple energy-saving strategies, or the energy consumption value corresponding to the target energy-saving strategy is less than or equal to Energy consumption threshold.
  • the present application provides a device for adjusting the running state of a network device.
  • the device is applied to the second network device.
  • the device includes a sending unit, a receiving unit, a sending unit and an application unit.
  • the sending unit is configured to send the first flow information to the first network device.
  • the first traffic information indicates the value of traffic processed by the second network device within the first time period.
  • the receiving unit is configured to receive the target energy saving policy sent by the first network device.
  • the target energy saving policy is determined by the first network device according to the second traffic information.
  • the second traffic information is the traffic corresponding to the second network device within the second time period predicted by the first network device according to the first traffic information.
  • the second time period is later than the first time period.
  • the target energy-saving policy includes configuration parameters corresponding to when devices in the second network device run according to the target energy-saving policy.
  • the energy consumption corresponding to the target energy-saving policy satisfies a preset condition.
  • the application unit is configured to apply the configuration parameters corresponding to the target energy saving policy.
  • the receiving unit is further configured to receive the plan execution time sent by the first network device.
  • the planned execution time indicates the execution time of the target energy saving policy.
  • the plan execution time is determined by the first network device according to the second traffic information and the target energy saving policy.
  • the receiving unit is further configured to receive the wakeup value sent by the first network device.
  • the wakeup value indicates a condition for the second network device to terminate execution of the target energy saving policy.
  • the wakeup value is determined by the first network device according to the target energy saving policy.
  • the device further includes a termination unit.
  • the terminating unit is configured to terminate execution of the target energy saving strategy when the second network device executes the target energy saving strategy and the statistical value of the second network device exceeds the wakeup value.
  • the device further includes a determination unit.
  • the determining unit is used for determining the transmission performance value of the second network device when executing the target energy saving strategy.
  • the terminating unit is configured to terminate execution of the target energy-saving strategy when the transmission performance value is greater than a transmission performance threshold.
  • the sending unit is further configured to send the actual execution time of the target energy saving policy to the first network device.
  • the execution duration corresponding to the actual execution time is shorter than the execution duration corresponding to the planned execution time.
  • the sending unit is further configured to send third traffic information to the first network device.
  • the third flow information indicates the flow that causes the termination of the target energy-saving policy.
  • the present application provides a network device.
  • the network device includes a processor and memory.
  • Memory is used to store instructions or computer programs.
  • the processor is configured to execute instructions or computer programs in the memory, so that the network device executes the network device operating state adjustment method described in the first aspect or any possible implementation manner of the first aspect, or executes the second aspect or the first aspect The method for adjusting the operating state of a network device described in any possible implementation manner of the two aspects.
  • the present application provides a computer-readable storage medium.
  • the storage medium includes instructions.
  • the computer is made to execute the method for adjusting the operating state of the network device described in the first aspect or any possible implementation of the first aspect, or execute the second aspect or any possible implementation of the second aspect The method for adjusting the operating state of the network equipment described in the manner.
  • the present application provides a computer program product.
  • the computer program product includes programs or codes.
  • the computer implements the method for adjusting the operating state of the network device as described in the first aspect or any possible implementation of the first aspect, or implements the second aspect or any of the second aspects The method for adjusting the running state of a network device described in a possible implementation manner.
  • the second network device can send the flow information processed by itself within the first time period, that is, the first flow information, to the first network device.
  • the first network device predicts the second flow information corresponding to the second network device within a second time period according to the first flow information, wherein the second time period is later than the first time period.
  • the first network device determines the energy consumption corresponding to each energy saving strategy in the multiple energy saving strategies according to the second flow information.
  • each energy saving strategy includes configuration parameters corresponding to the components in the second network device running according to the energy saving strategy.
  • the first network device After determining the energy consumption corresponding to each energy-saving strategy, the first network device determines the energy-saving strategy corresponding to the energy The configuration parameters of the target energy saving policy run. That is, the first network device determines a target energy-saving policy that matches the traffic information by predicting the second traffic information that the second network device needs to process within a second time period, and sends the target energy-saving policy to the second network device, Therefore, the second network device configures the parameters of the internal device according to the target energy saving policy matched with the traffic to be processed. Therefore, this solution enables the second network device to consume low energy consumption during the low traffic time period, thereby reducing energy consumption.
  • FIG. 1 is a flow chart of a method for adjusting the operating state of a network device provided in an embodiment of the present application
  • FIG. 2 is a schematic diagram of an application scenario provided by an embodiment of the present application.
  • FIG. 3 is a schematic structural diagram of an apparatus for adjusting the operating state of a network device provided in an embodiment of the present application
  • FIG. 4 is a schematic structural diagram of another apparatus for adjusting the operating state of network equipment provided by an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of a network device provided in an embodiment of the present application.
  • FIG. 6 is a schematic structural diagram of another network device provided by an embodiment of the present application.
  • the energy consumption of the network equipment is related to the configuration parameters of each device in the network equipment.
  • the configuration parameters of network devices are usually kept in a high configuration state, which causes the network devices to continuously generate high energy consumption.
  • the present application provides a method for adjusting the operating state of a network device, which is used to select an energy-saving strategy that matches the traffic to be processed by the network device from multiple energy-saving strategies, and send the matched energy-saving strategy to the network Devices, so that network devices operate according to the configuration parameters in the energy saving policy.
  • the network devices operate according to the configuration parameters that match the traffic to be processed, so that the network devices can not only handle high traffic normally, but also reduce energy consumption when the traffic decreases.
  • this figure is a flowchart of a method for adjusting the operating state of a network device provided in an embodiment of the present application. As shown in FIG. 1, the method includes:
  • S101 The second network device sends first traffic information to the first network device.
  • the second network device may collect flow information processed by itself within the first time period, that is, first flow information, and send the first flow information to the first network device.
  • the first flow information is used to indicate the second The value of traffic handled by the network device during the first time period.
  • the first traffic information may include: the sending rate and/or receiving rate of the second network device within the first time period, or the amount of data received and/or sent by the second network device within the first time period.
  • the sending rate may be rate statistical values such as an average sending rate and a maximum sending rate.
  • the amount of data may be the number of bits, bytes, packets, etc.
  • the first traffic information sent by the second network device may be device-level traffic information, board-level traffic information, or interface-level traffic information.
  • the first network device can adjust the parameters of the interface in a targeted manner based on the flow information of the interface, which can realize more accurate adjustment of the interface parameters .
  • the first network device receives first traffic information sent by the second network device, and predicts second traffic information corresponding to the second network device within a second time period according to the first traffic information.
  • the first network device After receiving the first flow information sent by the second network device, the first network device predicts the second flow information corresponding to the second network device within the second time period according to the first flow information.
  • the second time period is later than the first time period. That is, the first network device can predict the trend of future traffic according to the traffic information in the historical time period.
  • the first network device may input the first traffic information into the traffic prediction model, so as to obtain the second traffic information output by the traffic prediction model.
  • the traffic prediction model may be a preset model, for example, the first network device receives a traffic prediction model.
  • the received traffic prediction model may be trained by other network devices, or may be configured by an administrator.
  • the traffic prediction model may also be pre-trained and generated by the first network device according to historical traffic information of the second network device.
  • the first network device inputs the first traffic information into the traffic prediction model, so as to obtain the second traffic information through the traffic prediction model.
  • the historical traffic information refers to traffic values processed by the second network device in different time periods within the historical time period.
  • the traffic prediction model may be a regression prediction model, a neural network model, and the like.
  • the traffic prediction model is a neural network model
  • the first network device uses the acquired historical traffic information as training samples, and each training sample includes M traffic sequences and N traffic sequences.
  • Each flow sequence includes one or more flow values and a time corresponding to the one or more flow values.
  • the times corresponding to the N traffic sequences are later than the times corresponding to the M traffic sequences.
  • N flow sequences are labels.
  • the first network device can input a training sample into the neural network model, and the neural network model outputs inference results (predicted traffic sequences/values) for the M traffic sequences in the training sample .
  • the first network device can calculate the loss value between the inference result output by the neural network model and the actual result (label) of the group of training samples through a corresponding loss function. Then, the first network device can calculate the change gradient of the parameters in each network layer in the neural network model according to the calculated loss value. In this way, the first network device can calculate the adjustment value of the parameter during this round of iterative training based on the hyperparameters preset in the optimizer and the change gradient of the parameters in each network layer (also called parameter update ), the adjustment value can be, for example, the product of the change gradient and a hyperparameter (such as learning rate, etc.), so that the first network device can update the parameter value of the parameter based on the calculated adjustment value of each parameter.
  • the training is stopped to obtain the traffic prediction model.
  • the first network device determines the energy consumption corresponding to each energy-saving strategy in the multiple energy-saving strategies according to the second traffic information.
  • the first network device determines the energy corresponding to each of the multiple energy saving strategies according to the second traffic information. consumption.
  • Each energy saving strategy includes corresponding configuration parameters when devices in the second network device run according to the energy saving strategy.
  • the energy saving policy includes but not limited to various configuration parameters such as central processing unit (central processing unit, CPU) core on/off state, core frequency, port/switch board on/off state, port/switch board sleep state, etc.
  • the multiple energy-saving policies are pre-configured, and some or all of the configuration parameters included in each of the multiple energy-saving policies are different.
  • the plurality of energy-saving policies may be a group of energy-saving policies selected from a set of pre-configured energy-saving policies.
  • the first network device may determine the energy consumption corresponding to each energy-saving strategy in the following manner, specifically including: for each energy-saving strategy, the first network device inputs the second traffic information and the energy-saving strategy into the energy consumption prediction model, and obtains the energy consumption The energy consumption corresponding to the energy-saving strategy output by the consumption prediction model.
  • the energy consumption prediction model is generated according to training samples, and each training sample includes flow information, energy consumption value, and configuration parameters corresponding to the energy consumption value. That is, the first network device may determine the energy consumption corresponding to each energy-saving strategy by using the energy consumption prediction model generated through pre-training.
  • the training samples may come from the first network device or from other network devices.
  • the device type of the other network device is the same as that of the first network device, for example, the model of the other network device is the same as that of the first network device.
  • a training sample includes traffic information processed by the network device in a historical time period, energy consumption values corresponding to the historical time period, and configuration parameters corresponding to the historical time period.
  • the training of the energy consumption prediction model can be implemented in the following ways:
  • the first network device acquires local training samples, and uses the local training samples to train the energy consumption prediction model.
  • the local training samples refer to training samples provided by one or more network devices managed by the first network device.
  • the local training samples include training samples provided by one or more network devices managed by the first network device.
  • the multiple network devices have the same or similar device types, for example, the multiple network devices have the same model. Network devices of the same device type correspond to one energy consumption prediction model, and network devices of different device types correspond to different energy consumption prediction models.
  • a training sample includes traffic information processed by a network device in a historical time period, energy consumption values corresponding to the historical time period, and configuration parameters corresponding to the historical time period.
  • the first network device itself may complete the training of the energy consumption prediction model according to the acquired local training samples.
  • one or more network devices managed by the first network device may include the second network device, that is, the local training samples may include the historical traffic information, historical energy consumption value and historical energy consumption value corresponding to the second network device. configuration parameters.
  • the first network device acquires the local training samples, it sends the local training samples to the third network device, so that the third network device uses the received training samples to train the energy consumption prediction model.
  • the first network device receives the energy consumption prediction model sent by the third network device.
  • the local training samples include historical traffic information, historical energy consumption values, and configuration parameters corresponding to the historical energy consumption values of network devices managed by the first network device.
  • the network devices managed by the first network device may include the second network device, that is, the local training samples may include the historical flow information, the historical energy consumption value and the corresponding configuration parameters.
  • the third network device uses the local training samples reported by the first network device to train the energy consumption prediction model, and sends the trained energy consumption prediction model to the first network device.
  • the trained energy consumption prediction model can learn the relationship among flow information, configuration parameters, and energy consumption values, and then when the energy consumption prediction model is used to determine the energy consumption of each energy-saving strategy, the second flow information and The configuration parameters corresponding to the energy saving strategy are input into the energy consumption prediction model, and the energy consumption value corresponding to the energy saving strategy is determined through the energy consumption prediction model.
  • the third network device can receive the local training samples sent by the multiple first network devices respectively, so as to use a large number of local training samples to train and generate the energy consumption prediction model to improve energy consumption The accuracy of the prediction model's predictions.
  • the other is that the first network device uses local training samples to train and generate the first energy consumption prediction model, and sends the model parameters of the first energy consumption prediction model to the third network device, so that the third network device generates the first energy consumption prediction model according to the multiple first network devices.
  • the model parameters of the first energy consumption prediction model respectively sent by a network device determine the second energy consumption prediction model.
  • the first network device receives the second energy consumption prediction model sent by the third network device, and uses the second energy consumption prediction model to predict the energy consumption of each energy saving strategy.
  • the local training samples include historical traffic information, historical energy consumption values, and configuration parameters corresponding to the historical energy consumption values of network devices managed by the first network device.
  • each of the multiple first network devices managed by the third network device first uses local training samples to generate a first energy consumption prediction model, and uses the model parameters of the first energy consumption prediction model sent to the third network device.
  • the third network device determines the model parameters of the energy consumption prediction model according to the multiple model parameters, thereby generating the energy consumption prediction model. That is, in this implementation manner, the third network device and the multiple first network devices determine the energy consumption prediction model through federated learning.
  • network devices of the same device type correspond to one energy consumption prediction model
  • network devices of different device types correspond to different energy consumption prediction models.
  • the first network device or the third network device trains the energy consumption prediction model based on training samples from network devices of the same type.
  • a training sample includes traffic information processed by a network device of the device type in a historical time period, energy consumption values corresponding to the historical time period, and configuration parameters corresponding to the historical time period.
  • the first network device will determine the energy consumption prediction corresponding to the device type according to the device type of the second network device model, so that the energy consumption prediction model is used to predict the energy consumption corresponding to the energy saving strategy.
  • network devices of the same type are, for example, network devices of the same model.
  • the energy consumption prediction model may be a regression prediction model, a neural network model, or the like.
  • the network device obtains local training samples, and each local training sample includes traffic information processed by a network device in a historical time period, corresponding to the The energy consumption value of the historical time period and the configuration parameters corresponding to the historical time period, wherein the energy consumption value is used as a label.
  • the network device can input the traffic information and configuration parameters in a training sample into the neural network model, and the neural network model outputs an inference result (energy consumption value) for the training sample.
  • the network device can calculate the loss value between the inference result output by the neural network model and the actual result (label) of the training sample through the corresponding loss function. Then, the network device can calculate the change gradient of the parameters in each network layer in the neural network model according to the calculated loss value. In this way, the network device can calculate the adjustment value of the parameter in this round of iterative training process (also called the parameter update amount) based on the hyperparameters preset in the optimizer and the change gradient of the parameters in each network layer.
  • the adjustment value can be, for example, the product of the change gradient and a hyperparameter (such as learning rate, etc.), so that the network device can update the parameter value of each parameter based on the calculated adjustment value of each parameter.
  • the first network device can be a controller
  • the third network device can be a cloud device.
  • the controller and the cloud device can be integrated into the same hardware device, or they can be two independent hardware devices, or they can be two independent virtual equipment.
  • the first network device determines an energy saving strategy corresponding to the energy consumption meeting the preset condition as a target energy saving strategy.
  • the energy-saving strategy corresponding to the energy consumption meeting the preset condition is determined as the target energy-saving strategy.
  • the preset condition can be set according to the actual application situation, for example, the energy consumption value is smaller than the preset energy consumption threshold, or the energy consumption value is the smallest.
  • the determined target energy saving strategy may include one or more energy saving strategies.
  • the first network device may divide the second time period into multiple sub-time periods, and determine corresponding energy-saving policies for the multiple sub-time periods. In this case, the target energy-saving policy includes multiple energy-saving policies.
  • the second time period is from 1:00 am to 7:00 am, and a sub-time period is divided every 2 hours. Then, 1:00 am to 3:00 am corresponds to energy saving strategy 1, and 3 am to 5 am corresponds to energy saving strategy 2, and 5 am to Point 7 corresponds to energy-saving strategy 3, and the target energy-saving strategy includes energy-saving strategy 1, energy-saving strategy 2, and energy-saving strategy 3.
  • the first network device may determine an energy saving policy for multiple second time periods, that is, determine a target energy saving policy for each second time period. For example, the period from 1 am to 7 am is the first second time period, 7 am to 13 am is the second second time period, and 13 am to 19 am is the third second time period.
  • the first network device Energy-saving strategies are determined for the three second time periods respectively. At this time, the target energy-saving strategy may include the three energy-saving strategies.
  • S105 The first network device sends the target energy saving policy to the second network device.
  • the first network device may agree with the second network device on the execution time of the target energy saving policy, for example, the first network device and the second network device agree to apply the energy saving policy at the next hour when the energy saving policy is received. For example, if the second network device receives the energy saving policy at 13:30, the energy saving policy will be applied at 14:00.
  • the first network device may also determine the planned execution time corresponding to the target energy saving policy according to the second traffic information and the target energy saving policy. For example, when the target energy saving strategy includes only one energy saving strategy, the first network device determines the time period corresponding to the second traffic information as the execution time of the target energy saving strategy; when the target energy saving strategy includes multiple energy saving strategies, the first network device determines The time period corresponding to the traffic information corresponding to each energy saving strategy in the target energy saving strategy is determined as the execution time of the corresponding energy saving strategy. The first network device sends the planned execution time to the second network device, so that the second network device executes the target energy saving policy within a time period corresponding to the planned execution time.
  • the planned execution time may include the start time and end time of the second network device executing the target energy-saving policy, or the planned execution time may include information such as start time and execution duration, and the start time is used to instruct the second network device to start executing the target energy-saving policy
  • the end time is used to indicate the time when the second network device stops executing the target energy saving policy, and the start time is earlier than the end time.
  • the second network device receives the target energy saving policy, and applies configuration parameters corresponding to the target energy saving policy.
  • the first network device After the first network device determines the target energy saving policy, the first network device sends the target energy saving policy to the second network device, so that the second network device applies configuration parameters corresponding to the target energy saving policy.
  • the second network device may apply configuration parameters corresponding to the target energy saving policy according to a preset configured energy saving time.
  • the second network device may apply the configuration parameters corresponding to the target energy-saving policy according to the planned execution time sent by the first network device, and not apply the target energy-saving policy in a time period outside the planned execution time.
  • the first network device may send the target energy saving strategy and the planned execution time through the same message, or may send the target energy saving strategy and the planned execution time through different messages.
  • the first network device may also determine a wakeup value according to the target energy saving policy, where the wakeup value indicates a condition for the second network device to terminate execution of the target energy saving policy.
  • the wakeup value may be a traffic threshold, and when the traffic processed by the second network device exceeds the traffic threshold, execution of the target energy saving policy is terminated.
  • the wakeup value can be a performance threshold, such as a latency threshold value, jitter threshold, packet loss rate threshold, etc., when the performance value of the second network device when executing the target energy saving strategy exceeds the performance threshold, execution of the target energy saving strategy is terminated.
  • the traffic threshold may be a threshold corresponding to each interface in the second network device, or may be a threshold of the entire second network device.
  • the performance threshold may also be the threshold of each interface in the second network device, or may be the threshold of the entire second network device.
  • the first network device may calculate its corresponding wake-up value according to configuration parameters corresponding to the target energy-saving policy. For example, if the second network device includes 8 ports, and the target energy-saving policy indicates that port 1, port 3, and port 4 are closed, then the first network device calculates the throughput threshold according to the transmission bandwidth of the remaining ports. Alternatively, an association relationship between each energy-saving policy and its corresponding wake-up value is pre-configured, and after the target energy-saving policy is determined, a matching wake-up value is determined according to the above association relationship.
  • abnormal traffic may occur, so that the statistical value of the second network device exceeds the wake-up value.
  • the second network device will terminate the execution of the target in advance energy saving strategy.
  • the second network device may send the actual execution time of the target energy-saving policy to the first network device, and the execution duration corresponding to the actual execution time is shorter than the execution duration corresponding to the planned execution time.
  • the second network device terminates the execution of the target energy-saving strategy ahead of time, it can continue to run according to the configuration parameters of each device before executing the target energy-saving strategy, or continue to run according to the pre-stored default configuration parameters.
  • the first network device may update the target energy saving policy according to the actual execution time.
  • updating the target energy saving policy may include updating configuration parameters corresponding to the target energy saving policy. For example, in the energy-saving strategy before the update, the CPU frequency is reduced by 50%, and after the update, the CPU frequency is reduced by 45%.
  • the energy-saving policy of the system is to close a certain port + close a certain switching network board.
  • the second network device may also send third traffic information to the first network device, where the traffic value corresponding to the third traffic information exceeds the wake-up value.
  • the first network device optimizes the traffic prediction model according to the third traffic information, so that the prediction result of the traffic prediction model is more accurate.
  • the third flow information may include statistical values such as a time series indicating the size of the flow, a flow mean value, and a flow variance.
  • the first network device may not send the wake-up value to the second network device, but the second network device determines whether to terminate the execution of the target energy saving policy according to the current transmission performance value when executing the target energy saving policy.
  • Strategy Specifically, when the second network device executes the target energy saving strategy, the second network device determines the transmission performance value; when the transmission performance value is greater than the transmission performance threshold, the second network device terminates execution of the target energy saving strategy.
  • the transmission performance value may reflect the transmission quality of the second network device, and may include transmission delay, packet loss rate, and the like. For example, when the transmission delay of the second network device is greater than the delay threshold, the second network device stops executing the target energy saving policy.
  • the second network device stops executing the target energy saving policy.
  • the second network device stops executing the target energy saving policy.
  • the second network device sends the actual execution time of the target energy saving policy to the first network device, so that the first network device can know that the target energy saving policy is terminated in advance.
  • the second network device may also send third flow information to the first network device, where the third flow information indicates the flow that triggers termination of the target energy-saving policy, so that the first network device optimizes the flow prediction model according to the third flow information.
  • the second network device may report the traffic information processed by itself within the first time period to the first network device, That is, the first flow information.
  • the first network device predicts the second flow information corresponding to the second network device within a second time period according to the first flow information, wherein the second time period is later than the first time period.
  • the first network device determines the energy consumption corresponding to each energy saving strategy in the multiple energy saving strategies according to the second flow information.
  • each energy-saving strategy includes corresponding configuration parameters when devices in the second network device run according to the energy-saving strategy.
  • the first network device After determining the energy consumption corresponding to each energy-saving strategy, the first network device determines the energy-saving strategy corresponding to the energy The configuration parameters of the target energy saving policy run. That is, the first network device determines a target energy-saving strategy matching the flow information by predicting the second flow information processed by the second network device within the second time period, and sends the target energy-saving strategy to the second network device, Therefore, the second network device configures the parameters of the internal device according to the target energy saving policy matching the traffic to be processed. Therefore, the second network device does not need to maintain a high configuration all the time, but can flexibly apply a configuration matching the traffic to be processed, so as to reduce energy consumption generated during a low traffic time period.
  • Fig. 2 is a schematic diagram of an application scenario provided by an embodiment of the present application.
  • the application scenario includes cloud devices, analyzers, controllers and network devices.
  • Cloud devices can be deployed in public clouds, edge clouds or distributed clouds.
  • Network devices may include forwarding devices and terminal devices in the network.
  • the controller can collect data from the network device and send it to the analyzer, so that the analyzer can predict the flow information processed by the network device according to the collected data, and then determine the energy-saving strategy and the time to execute the energy-saving strategy.
  • the analyzer sends the determined energy-saving strategy and the corresponding execution time to the network device through the controller, and the network device applies the configuration parameters corresponding to the energy-saving strategy, thereby reducing the energy consumption of the network device.
  • the controller can collect training samples, and send the training samples to the cloud device through the analyzer, and the cloud device uses the training samples to train the energy consumption prediction model.
  • a training sample includes traffic information of a network device in a historical time period, energy consumption values corresponding to the historical time period, and configuration parameters corresponding to the historical time period.
  • the cloud device generates an energy consumption prediction model for the device type based on multiple training samples of network devices of the same device type, and sends the energy consumption prediction model to the analyzer.
  • the controller collects the first traffic information corresponding to the network device in the first time period, and sends the first traffic information to the analyzer, which uses the first traffic information and the traffic prediction model to obtain the network The second traffic information corresponding to the device in the second time period.
  • the analyzer inputs the second flow information and configuration parameters corresponding to an energy saving strategy into the energy consumption prediction model, determines the energy consumption corresponding to the energy saving strategy, and determines the energy saving strategy corresponding to the energy consumption meeting the preset condition as the target energy saving strategy.
  • the analyzer can also determine the planned execution time and wake-up value corresponding to the target energy-saving strategy, and send the target energy-saving strategy, planned execution time and wake-up value to the network device.
  • the policy execution module in the network device executes the target energy-saving policy according to the planned execution time, so as to reduce energy consumption.
  • the wake-up detection module in the network device is used to judge whether the traffic currently processed by the network device is greater than the wake-up value, and if so, an exception message can be sent to the policy execution module, so that the policy execution module terminates the execution of the target energy saving strategy.
  • the network device can also send the actual execution time and abnormal traffic information to the analyzer through the controller, so that the analyzer can update the energy saving strategy and the traffic prediction model based on the actual execution time and abnormal traffic information.
  • the analyzer and the controller may be the same physical device, or may be two independent physical devices.
  • the analyzer and cloud device can be the same physical device, or two separate physical devices.
  • the cloud device can collect training samples sent by multiple analyzers to use a large number of training samples to train the energy consumption prediction model, Improve the prediction accuracy of the energy consumption prediction model.
  • an embodiment of the present application further provides an apparatus for adjusting the running state of network equipment, which will be described below with reference to the accompanying drawings.
  • this figure is a schematic structural diagram of an apparatus 300 for adjusting the operating state of a network device provided in an embodiment of the present application.
  • the apparatus 300 may be used to implement the functions of the above-mentioned first network device.
  • the apparatus 300 includes a receiving unit 301 , a predicting unit 302 , a determining unit 303 and a sending unit 304 .
  • the receiving unit 301 is configured to receive the first flow information sent by the second network device.
  • the first traffic information indicates the value of traffic processed by the second network device within the first time period.
  • the predicting unit 302 is configured to predict second traffic information corresponding to the second network device within a second time period according to the first traffic information.
  • the second time period is later than the first time period.
  • the determining unit 303 is configured to determine an energy consumption value corresponding to each energy-saving strategy in the plurality of energy-saving strategies according to the second traffic information.
  • Each energy saving strategy includes configuration parameters corresponding to components in the second network device running according to the energy saving strategy.
  • the determining unit 303 is further configured to determine an energy saving strategy corresponding to an energy consumption value satisfying a preset condition as a target energy saving strategy.
  • the sending unit 304 is configured to send the target energy saving policy to the second network device, so that the second network device operates according to the configuration parameters in the target energy saving policy.
  • the determining unit 303 is further configured to determine the planned execution time corresponding to the target energy saving strategy according to the second traffic information and the target energy saving strategy.
  • the sending unit 304 is further configured to send the planned execution time to the second network device, so that the second network device executes the target energy saving policy within a time period corresponding to the planned execution time.
  • the determining unit 303 is further configured to determine a wake-up value according to the target energy-saving policy.
  • the wakeup value indicates a condition for the second network device to terminate execution of the target energy saving policy.
  • the sending unit 304 is further configured to send the wakeup value to the second network device.
  • the prediction unit 302 is configured to input the first flow information into the flow prediction model, so as to obtain the second flow information output by the flow prediction model.
  • the traffic prediction model is trained and generated according to historical traffic information of the second network device.
  • the determining unit 303 is configured to, for each energy-saving strategy, input the second flow information and configuration parameters corresponding to the energy-saving strategy into the energy consumption prediction model, so as to obtain the energy-saving strategy corresponding to the energy-saving strategy output by the energy consumption prediction model.
  • consumption The energy consumption prediction model is generated based on training samples. Each training sample includes flow information of a network device in a historical period, energy consumption value corresponding to the network device in the historical period, and configuration parameters corresponding to the network device in the historical period.
  • the network device is a second network device, or a network device of the same device type as the second network device.
  • the determining unit 303 is further configured to determine the energy consumption prediction model according to the device type of the second network device before inputting the second traffic information and configuration parameters corresponding to the energy saving policy into the energy consumption prediction model.
  • the device type corresponds to the energy consumption prediction model.
  • the apparatus 300 further includes an updating unit.
  • the receiving unit 301 is further configured to receive the actual execution time of the target energy saving policy sent by the second network device.
  • the execution duration corresponding to the actual execution time is shorter than the execution duration corresponding to the planned execution time of the target energy saving measurement.
  • the updating unit is used for updating the target energy-saving strategy according to the actual execution time.
  • the apparatus 300 further includes an optimization unit.
  • the receiving unit 301 is further configured to receive third traffic information sent by the second network device.
  • the statistical value corresponding to the third traffic information exceeds the wake-up value.
  • the optimization unit is used for according to the third flow signal Information optimization traffic forecasting model.
  • the sending unit 304 is further configured to send the training sample to the third network device.
  • a training sample includes traffic information of a network device managed by the first network device within a historical time period, energy consumption value of the network device corresponding to the historical time period, and network Device configuration parameters.
  • the receiving unit 301 is further configured to receive the energy consumption prediction model sent by the third network device.
  • the energy consumption prediction model is trained and generated by the third network device using the above training samples.
  • the apparatus 300 further includes a generating unit.
  • the generation unit is used to generate a first energy consumption prediction model by using training samples.
  • the sending unit 304 is further configured to send the model parameters of the first energy consumption prediction model to the third network device.
  • the receiving unit 301 is further configured to receive the energy consumption prediction model sent by the third network device.
  • the energy consumption prediction model is determined by the third network device according to model parameters of the first energy consumption prediction model respectively sent by multiple first network devices.
  • the preset condition includes that the energy consumption value corresponding to the target energy-saving strategy is the minimum value among the multiple energy consumption values corresponding to the multiple energy-saving strategies, or that the energy consumption value corresponding to the target energy-saving strategy is less than or equal to the energy consumption threshold.
  • FIG. 4 this figure is a schematic structural diagram of an apparatus 400 for adjusting the operating state of a network device provided in an embodiment of the present application.
  • the apparatus 400 can be used to realize the function of the second network device.
  • the apparatus 400 includes a sending unit 401 , a receiving unit 402 and an application unit 403 .
  • a sending unit 401 configured to send first traffic information to a first network device.
  • the first traffic information indicates the value of traffic processed by the second network device within the first time period.
  • the receiving unit 402 is configured to receive the target energy saving policy sent by the first network device.
  • the target energy saving policy is determined by the first network device according to the second traffic information.
  • the second flow information is the flow corresponding to the second network device within the second time period predicted by the first network device according to the first flow information.
  • the second time period is later than the first time period.
  • the target energy-saving strategy includes configuration parameters corresponding to devices in the second network device running according to the target energy-saving strategy.
  • the energy consumption corresponding to the target energy-saving policy satisfies a preset condition.
  • the application unit 403 is configured to apply the configuration parameters corresponding to the target energy saving policy.
  • the receiving unit 402 is further configured to receive the plan execution time sent by the first network device.
  • the planned execution time indicates the execution time of the target energy saving policy.
  • the plan execution time is determined by the first network device according to the second traffic information and the target energy saving policy.
  • the receiving unit 402 is further configured to receive the wakeup value sent by the first network device.
  • the wakeup value indicates a condition for the second network device to terminate execution of the target energy saving policy.
  • the wakeup value is determined by the first network device according to the target energy saving policy.
  • the apparatus 400 further includes a termination unit.
  • the terminating unit is configured to terminate execution of the target energy saving strategy when the second network device executes the target energy saving strategy and the statistical value of the second network device exceeds the wakeup value.
  • the apparatus 400 further includes a determining unit.
  • the determining unit is configured to determine a transmission performance value when the second network device executes the target energy saving strategy.
  • the terminating unit is configured to terminate execution of the target energy-saving strategy when the transmission performance value is greater than the transmission performance threshold.
  • the sending unit 401 is further configured to send the actual execution time of the target energy saving policy to the first network device. Should The execution duration corresponding to the actual execution time is shorter than the execution duration corresponding to the planned execution time.
  • the sending unit 401 is further configured to send third traffic information to the first network device.
  • the third flow information indicates the flow that causes the termination of the target energy-saving policy.
  • FIG. 5 is a schematic structural diagram of a network device provided by an embodiment of the present application.
  • the network device can be, for example, the first network device, the second network device, or the third network device in the above method embodiment, or it can also be the device realization of the device 300 in the embodiment shown in FIG. 3 , or it can also be the device in FIG. 4
  • the network device 500 includes: a processor 510 , a communication interface 520 and a memory 530 .
  • the number of processors 510 in the network device 500 may be one or more, and one processor is taken as an example in FIG. 5 .
  • the processor 510, the communication interface 520, and the memory 530 may be connected through a bus system or other methods, wherein the connection through the bus system 540 is taken as an example in FIG. 5 .
  • the processor 510 may be a CPU, a network processor (network processor, NP), or a combination of a CPU and an NP.
  • the processor 510 may further include a hardware chip.
  • the aforementioned hardware chip may be an application-specific integrated circuit (application-specific integrated circuit, ASIC), a programmable logic device (programmable logic device, PLD) or a combination thereof.
  • the aforementioned PLD may be a complex programmable logic device (complex programmable logic device, CPLD), a field-programmable gate array (field-programmable gate array, FPGA), a general array logic (generic array logic, GAL) or any combination thereof.
  • the memory 530 may include a volatile memory (volatile memory), such as a random-access memory (random-access memory, RAM); the memory 530 may also include a non-volatile memory (non-volatile memory), such as a flash memory (flash memory). memory), a hard disk (hard disk drive, HDD) or a solid-state drive (solid-state drive, SSD); the memory 530 may also include a combination of the above types of memory.
  • volatile memory volatile memory
  • RAM random-access memory
  • non-volatile memory such as a flash memory (flash memory).
  • flash memory flash memory
  • HDD hard disk drive
  • solid-state drive solid-state drive
  • the memory 530 stores operating systems and programs, executable modules or data structures, or their subsets, or their extended sets, wherein the programs may include various operating instructions for implementing various operations.
  • the operating system may include various system programs for implementing various basic services and processing hardware-based tasks.
  • the processor 510 can read the program in the memory 530 to implement the method provided by the embodiment of the present application.
  • the memory 530 may be a storage device in the network device 500 , or may be a storage device independent of the network device 500 .
  • the bus system 540 may be a peripheral component interconnect standard (peripheral component interconnect, PCI) bus or an extended industry standard architecture (extended industry standard architecture, EISA) bus or the like.
  • PCI peripheral component interconnect
  • EISA extended industry standard architecture
  • the bus system 540 can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is used in FIG. 5 , but it does not mean that there is only one bus or one type of bus.
  • FIG. 6 is a schematic structural diagram of a network device 600 provided in an embodiment of the present application.
  • the network device can be, for example, the first network device, the second network device, or the third network device in the method embodiment, or it can also be the network device shown in FIG. 3
  • the network device 600 includes: a main control board 610 and an interface board 630 .
  • the main control board 610 is also called a main processing unit (main processing unit, MPU) or a route processing card (route processor card), the main control board 610 is used to control and manage each component in the network device 600, including routing calculation, device management, device maintenance, protocol processing, etc.
  • the main control board 610 includes: a CPU 611 and a memory 612 .
  • the interface board 630 is also called a line processing unit (line processing unit, LPU), a line card (line card), or a service board.
  • the interface board 630 is used to provide various service interfaces and implement forwarding of data packets.
  • Service interfaces include but are not limited to Ethernet interfaces, POS (Packet over SONET/SDH) interfaces, etc.
  • Ethernet interfaces are, for example, Ethernet ports, Gigabit Ethernet ports, Flexible Ethernet service interfaces (Flexible Ethernet Clients, FlexE Clients) and the like.
  • the interface board 630 includes: a central processing unit 631 , a network processor 632 , a forwarding entry storage 634 and a physical interface card (physical interface card, PIC) 633 .
  • the CPU 631 on the interface board 630 is used to control and manage the interface board 630 and communicate with the CPU 611 on the main control board 610 .
  • the network processor 632 is configured to implement message forwarding processing.
  • the form of the network processor 632 may be a forwarding chip.
  • the processing of the uplink message includes: processing of the inbound interface of the message, forwarding table search; the processing of the downlink message includes forwarding table search and so on.
  • the physical interface card 633 is used to implement the interconnection function of the physical layer.
  • the original traffic enters the interface board 630 through this, and the processed packets are sent out from the physical interface card 633 .
  • the physical interface card 633 includes at least one physical interface, which is also called a physical port.
  • the physical interface card 633 is also called a daughter card, which can be installed on the interface board 630, and is responsible for converting the photoelectric signal into a message, checking the validity of the message and forwarding it to the network processor 632 for processing.
  • the central processing unit 631 of the interface board 603 can also execute the functions of the network processor 632 , such as implementing software forwarding based on a general-purpose CPU, so that the physical interface card 633 does not need the network processor 632 .
  • the network device 600 includes multiple interface boards.
  • the network device 600 further includes an interface board 640
  • the interface board 640 includes: a central processing unit 641 , a network processor 642 , a forwarding entry storage 644 and a physical interface card 643 .
  • the network device 600 further includes a switching fabric unit 620 .
  • the SFU 620 may also be called a SFU unit (switch fabric unit, SFU).
  • SFU switch fabric unit
  • the switching fabric board 620 is used to complete data exchange between the interface boards.
  • the interface board 630 and the interface board 640 may communicate through the switching fabric board 620 .
  • the main control board 610 is coupled to the interface board 630 .
  • the main control board 610, the interface board 630, the interface board 640, and the switching fabric board 620 are connected to the system backplane through the system bus to realize intercommunication.
  • an inter-process communication protocol (inter-process communication, IPC) channel is established between the main control board 610 and the interface board 630, and the main control board 610 and the interface board 630 communicate through the IPC channel.
  • IPC inter-process communication
  • the network device 600 includes a control plane and a forwarding plane.
  • the control plane includes a main control board 610 and a central processing unit 631.
  • the forwarding plane includes various components for performing forwarding, such as forwarding entry storage 634, physical interface card 633, and network processing. device 632.
  • the control plane executes routers, generates forwarding tables, processes signaling and protocol packets, configures and maintains device status, and other functions.
  • the control plane sends the generated forwarding tables to the forwarding plane.
  • the network processor 632 The issued forwarding table looks up and forwards the packets received by the physical interface card 633 .
  • the forwarding table delivered by the control plane may be stored in the forwarding table item storage 634 .
  • the control plane and the forwarding plane can be completely separated and not on the same device.
  • the operations on the interface board 640 in the embodiment of the present application are consistent with the operations on the interface board 630.
  • the network device 600 in this embodiment may correspond to the network device in the foregoing method embodiments, and the main control board 610, interface board 630, and/or interface board 640 in the network device 600 may implement the foregoing method embodiments
  • the various steps in , for the sake of brevity, are not repeated here.
  • main control boards there may be one or more main control boards, and when there are multiple main control boards, it may include an active main control board and a standby main control board.
  • the network device can have at least one SFU, through which the data exchange between multiple interface boards can be realized, and large-capacity data exchange and processing capabilities can be provided. Therefore, the data access and processing capabilities of network devices with a distributed architecture are greater than those with a centralized architecture.
  • the form of the network device can also be that there is only one board, that is, there is no switching fabric board, and the functions of the interface board and the main control board are integrated on this board.
  • the central processing unit and the main control board on the interface board The central processing unit on the board can be combined into one central processing unit on the board to perform the superimposed functions of the two.
  • the data exchange and processing capabilities of this form of equipment are low (for example, low-end switches or routers and other network equipment). Which architecture to use depends on the specific networking deployment scenario.
  • the foregoing network device may be implemented as a virtualization device.
  • the virtualization device may be a virtual machine (virtual machine, VM) running a program for sending packets, and the virtual machine is deployed on a hardware device (for example, a physical server).
  • a virtual machine refers to a complete computer system that is simulated by software and has complete hardware system functions and runs in a completely isolated environment.
  • a virtual machine can be configured as a network device.
  • a network device may be implemented based on a common physical server combined with a network functions virtualization (network functions virtualization, NFV) technology.
  • the network device is a virtual host, a virtual router or a virtual switch.
  • a person skilled in the art can virtualize a network device with the above-mentioned functions on a general physical server by combining NFV technology by reading this application, so details will not be repeated here.
  • the embodiment of the present application also provides a chip, including a processor and an interface circuit.
  • the interface circuit is used to receive instructions and transmit them to the processor.
  • the processor for example, may be a specific implementation form of the adjustment apparatus 300 shown in FIG. 3 , and may be used to execute the above method for adjusting the operating state of the network device.
  • the processor is coupled with a memory, and the memory is used to store a program or an instruction, and when the program or instruction is executed by the processor, the system-on-a-chip implements the method in any one of the above method embodiments.
  • processors in the chip system there may be one or more processors in the chip system.
  • the processor can be realized by hardware or by software.
  • the processor may be a logic circuit, an integrated circuit, or the like.
  • the processor may be a general-purpose processor implemented by reading software codes stored in a memory.
  • the memory can be integrated with the processor, or can be set separately from the processor, which is not limited in this application.
  • the memory can be a non-transitory processor, such as a read-only memory ROM, which can be integrated with the processor on the same chip, or can be respectively arranged on different chips.
  • the setting method of the processor is not specifically limited.
  • the chip system can be, for example, FPGA, ASIC, system chip (system on chip, SoC), CPU, NP, digital signal processor (DSP), microcontroller (micro controller unit, MCU), PLD or other integrated chips.
  • SoC system on chip
  • DSP digital signal processor
  • MCU microcontroller
  • PLD PLD
  • the embodiments of the present application also provide a computer-readable storage medium, including instructions or computer programs, which, when run on a computer, cause the computer to execute the method for adjusting the operating state of the network device provided in the above embodiments.
  • the embodiment of the present application also provides a computer program product including an instruction or a computer program, which, when run on a computer, causes the computer to execute the method for adjusting the operating state of the network device provided in the above embodiments.
  • the disclosed system, device and method can be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of units is only a logical business division. In actual implementation, there may be other division methods.
  • multiple units or components can be combined or integrated. to another system, or some features may be ignored, or not implemented.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.
  • a unit described as a separate component may or may not be physically separated, and a component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each business unit in each embodiment of the present application may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit.
  • the above-mentioned integrated units can be implemented in the form of hardware or in the form of software business units.
  • the integrated unit is realized in the form of a software business unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.
  • the technical solution of the present application is essentially or part of the contribution to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods in various embodiments of the present application.
  • the aforementioned storage medium includes: various media capable of storing program codes such as U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk.
  • the services described in this application may be implemented by hardware, software, firmware or any combination thereof.
  • the services When implemented in software, the services may be stored on or transmitted over as one or more instructions or code on a computer-readable medium.
  • Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another.
  • a storage media may be any available media that can be accessed by a general purpose or special purpose computer.

Abstract

本申请公开了一种网络设备运行状态的调整方法、装置及相关设备，涉及通信技术领域。第二网络设备向第一网络设备发送自身在第一时间段内处理的第一流量信息。第一网络设备根据第一流量信息预测第二网络设备在第二时间段内对应的第二流量信息，第二时间段晚于第一时间段。第一网络设备根据第二流量信息确定每个节能策略对应的能耗，将满足预设条件的能耗对应的节能策略确定为目标节能策略。第一网络设备发送目标节能策略给第二网络设备，以使得第二网络设备按照目标节能策略的配置参数运行以调整第二网络设备的运行状态。第二网络设备按照与待处理流量匹配的节能策略对应的配置参数运行，使得第二网络设备的能耗满足预设条件。

Description

一种网络设备运行状态的调整方法、装置及相关设备
本申请要求于2022年02月21日提交的申请号为202210159692.9、申请名称为“一种网络设备运行状态的调整方法、装置及相关设备”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信技术领域，尤其涉及一种网络设备运行状态的调整方法、装置及相关设备。
背景技术
随着网络技术的不断发展，人们对网络服务的需求在急剧增长，从而使得提供网络服务的网络设备的数量和规模也在快速增加。网络规模的扩大导致网络能耗的增加，这不仅增加了网络运营的成本，还会产生大量的碳排放。
发明内容
本申请提供了一种网络设备运行状态的调整方法、装置及相关设备，以通过调整网络设备的运行参数，降低网络设备的能耗。
第一方面，本申请提供了一种网络设备运行状态的调整方法。第一网络设备接收第二网络设备发送的第一流量信息，该第一流量信息指示第二网络设备在第一时间段内所处理流量的值。第一网络设备根据第一流量信息预测第二网络设备在第二时间段内所对应的第二流量信息。第二时间段晚于第一时间段。第一网络设备根据第二流量信息确定多个节能策略中每个节能策略对应的能耗值。该每个节能策略包括第二网络设备中的器件所对应的配置参数。第一网络设备将满足预设条件的能耗值对应的节能策略确定为目标节能策略，并将该目标节能策略发送给第二网络设备，以使得第二网络设备按照目标节能策略中的配置参数运行。
第一网络设备通过预测第二网络设备需要在第二时间段内处理的第二流量信息，确定与该流量信息匹配的目标节能策略，并将该目标节能策略发送第二网络设备，从而使得第二网络设备按照与待处理的流量匹配的目标节能策略配置内部器件的参数。从而，该方案使得第二网络设备在低流量时间段内消耗低的能耗，减少该网络设备产生的能耗。
其中，预设条件包括目标节能策略对应的能耗值为多个节能策略对应的多个能耗值中的最小值，或者，目标节能策略对应的能耗值小于等于能耗阈值。
在一种可能的实现方式中，第一网络设备还可以根据第二流量信息以及目标节能策略确定目标节能策略对应的计划执行时间，并将该计划执行时间发送给第二网络设备，以使得第二网络设备在计划执行时间对应的时间段内执行目标节能策略。
在一种可能的实现方式中，第一网络设备还可根据目标节能策略确定唤醒值，该唤醒值指示第二网络设备终止执行目标节能策略的条件。其中，唤醒值可以包括流量阈值和/或性能阈值，当唤醒值包括流量阈值且第二网络设备处理的流量超过流量阈值时，则终止执行目标节能策略；当唤醒值包括性能阈值且第二网络设备处理流量的性能值超过性能阈值时，则终止执行目标节能策略。其中，流量阈值可以包括吞吐量阈值，性能阈值可以包括时延阈值、抖动阈值、丢包率阈值等。
在一种可能的实现方式中，第一网络设备可以利用流量预测模型预测第一网络设备在第二时间段内所对应的流量信息。具体为，第一网络设备将第一流量信息输入流量预测模型，以获得该流量预测模型输出的第二流量信息。该流量预测模型是根据第二网络设备的历史流量信息训练生成的。
在一种可能的实现方式中，第一网络设备可以利用能耗预测模型确定每个节能策略对应的能耗。具体为，针对每个节能策略，第一网络设备将第二流量信息以及节能策略对应的配置参数输入能耗预测模型，以获得能耗预测模型输出的对应该节能策略的能耗。其中，能耗预测模型是根据训练样本生成的。每个训练样本包括流量信息、能耗值以及能耗值对应的配置参数。
其中，能耗预测模型与设备类型相对应，第一网络设备在利用能耗预测模型确定节能策略对应的能耗之前，可以根据第二网络设备的设备类型确定与该设备类型对应的能耗预测模型。其中，设备类型相同例如为网络设备的型号相同。
在一种可能的实现方式中，在第二网络设备按照目标节能策略运行时，可能存在异常情况，导致第二网络设备提前终止执行目标节能策略，该情况下，第一网络设备还可以接收第二网络设备发送的目标节能策略的实际执行时间。第一网络设备根据实际执行时间更新目标节能策略。其中实际执行时间对应的执行时长小于计划执行时间对应的执行时长。
在一种可能的实现方式中，第一网络设备还可以接收第二网络设备发送的第三流量信息，该第三流量信息对应的统计值超过唤醒值。第一网络设备根据第三流量信息优化流量预测模型。即，当第二网络设备因突发流量信息的出现导致统计值超过唤醒值时，可以向第一网络设备发送突发流量信息，以使得第一网络设备根据突发流量信息优化流量预测模型，提高流量预测模型的准确性。
在一种可能的实现方式中，第一网络设备向第三网络设备发送本地训练样本并接收第三网络设备发送的能耗预测模型。该本地训练样本包括第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及历史能耗值对应的配置参数。该能耗预测模型由第三网络设备利用本地训练样本训练生成的。例如，第一网络设备为控制器，第三网络设备为云端设备。
在一种可能的实现方式中，第一网络设备利用本地训练样本生成第一能耗预测模型，将第一能耗预测模型的模型参数发送给第三网络设备，并接收第三网络设备发送的能耗预测模型。该能耗预测模型由第三网络设备根据多个第一网络设备分别发送的第一能耗预测模型的模型参数确定。其中，本地训练样本包括第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及历史能耗值对应的配置参数。在该实现方式中，每个第一网络设备首先利用本地训练样本进行初步的训练获得第一能耗预测模型的模型参数，并将获得的模型参数发送给第三网络设备。第三网络设备利用多个第一网络设备发送的模型参数确定能耗预测模型对应的最终模型参数，从而不仅提高训练效率，还可以提高能耗预测模型的准确性。
第二方面，本申请提供了一种网络设备运行状态的调整方法。第二网络设备向第一网络设备发送第一流量信息，该第一流量信息指示第二网络设备在第一时间段内所处理流量 的值。第二网络设备接收第一网络设备发送的目标节能策略，并应用该目标节能策略所对应的配置参数。其中，目标节能策略是由第一网络设备根据第二流量信息确定的。第二流量信息是由第一网络设备根据第一流量信息预测的第二网络设备在第二时间段内所对应的流量，第二时间段晚于所述第一时间段。目标节能策略包括第二网络设备中的器件所对应的配置参数，且目标节能策略所对应的能耗满足预设条件。其中，预设条件包括目标节能策略对应的能耗值为多个节能策略对应的多个能耗值中的最小值，或者，目标节能策略对应的能耗值小于等于能耗阈值。
在一种可能的实现方式中，第二网络设备还可以接收第一网络设备发送的计划执行时间。该计划执行时间指示所述目标节能策略的执行时间且由第一网络设备根据第二流量信息以及目标节能策略确定的。
在一种可能的实现方式中，第二网络设备还可以接收第一网络设备发送的唤醒值。该唤醒值指示第二网络设备终止执行目标节能策略的条件。其中，所述唤醒值是由第一网络设备根据目标节能策略确定的。
在一种可能的实现方式中，第二网络设备在执行目标节能策略期间，当第二网络设备的统计值超过唤醒值时，第二网络设备终止执行目标节能策略。在该实现方式中，当第二网络设备的统计值超过唤醒值时，表明当前第二网络设备所处理的流量与预测的第二流量信息不匹配，出现突发流量，为更好地处理突发流量，第二网络设备可以提前终止执行目标节能策略。
在一种可能的实现方式中，在第二网络设备执行目标节能策略时，可以统计自身对应的传输性能值。在传输性能值大于传输性能阈值时，第二网络设备终止执行目标节能策略。
在一种可能的实现方式中，第二网络设备向第一网络设备发送目标节能策略的实际执行时间，从而使得第一网络设备可以获知第二网络设提前终止执行目标节能策略。该实际执行时间对应的执行时长小于计划执行时间对应的执行时长。
在一种可能的实现方式中，第二网络设备还可以向第一网络设备发送第三流量信息，从而使得第一网络设备利用第三流量信息对流量预测模型进行优化，提高流量预测模型的预测准确性。该第三流量信息指示引发目标节能策略终止的流量，即突发流量。
第三方面，本申请提供了一种网络系统。该系统包括第一网络设备和第二网络设备。第一网络设备用于执行第一方面或第一方面中任一可能的实现方式中所述的方法。第二网络设备用于执行第二方面或第二方面中任一可能的实现方式中所述的方法。
在一种可能的实现方式中，该网络系统还包括第三网络设备。第三网络设备，用于接收第一网络设备发送的本地训练样本，根据本地训练样本训练生成能耗预测模型，并将能耗预测模型发送给所述第一网络设备。该本地训练样本包括第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及所述历史能耗值所对应的配置参数。
第四方面，本申请提供了一种网络设备运行状态的调整装置。该装置应用于第一网络设备，包括接收单元、预测单元、确定单元和发送单元。接收单元用于接收第二网络设备发送的第一流量信息。该第一流量信息指示第二网络设备在第一时间段内所处理流量的值。预测单元用于根据第一流量信息预测第二网络设备在第二时间段内所对应的第二流量信息， 第二时间段晚于第一时间段。确定单元用于根据第二流量信息确定多个节能策略中每个节能策略对应的能耗值。每个节能策略包括第二网络设备中的器件根据该节能策略运行时所对应的配置参数。确定单元还用于将满足预设条件的能耗值对应的节能策略确定为目标节能策略。发送单元用于将目标节能策略发送给第二网络设备，以使得第二网络设备按照该目标节能策略中的配置参数运行。
在一种可能的实现方式中，确定单元还用于根据第二流量信息以及目标节能策略确定该目标节能策略对应的计划执行时间。发送单元还用于将该计划执行时间发送给第二网络设备，以使得第二网络设备在该计划执行时间对应的时间段内执行目标节能策略。
在一种可能的实现方式中，确定单元还用于根据目标节能策略确定唤醒值。该唤醒值指示第二网络设备终止执行该目标节能策略的条件。发送单元还用于将该唤醒值发送给第二网络设备。
在一种可能的实现方式中，预测单元用于将第一流量信息输入流量预测模型，以获得该流量预测模型输出的第二流量信息。流量预测模型是根据第二网络设备的历史流量信息训练生成的。
在一种可能的实现方式中，确定单元用于，对于每个节能策略，将第二流量信息以及该节能策略对应的配置参数输入能耗预测模型，以获得该能耗预测模型输出的对应该节能策略的能耗。该能耗预测模型是根据训练样本生成的。每个训练样本包括流量信息、能耗值以及该能耗值对应的配置参数。
在一种可能的实现方式中，确定单元还用于，在将所述第二流量信息以及节能策略对应的配置参数输入能耗预测模型之前，根据第二网络设备的设备类型确定该能耗预测模型。所述设备类型与所述能耗预测模型对应。
在一种可能的实现方式中，该装置还包括更新单元。接收单元还用于接收第二网络设备发送的目标节能策略的实际执行时间。更新单元用于根据该实际执行时间更新目标节能策略。该实际执行时间对应的执行时长小于所述计划执行时间对应的执行时长。
在一种可能的实现方式中，该装置还包括优化单元。接收单元还用于接收第二网络设备发送的第三流量信息。优化单元用于根据第三流量信息优化所述流量预测模型。该第三流量信息对应的统计值超过唤醒值。
在一种可能的实现方式中，发送单元还用于向第三网络设备发送本地训练样本。接收单元还用于接收第三网络设备发送的能耗预测模型。该能耗预测模型由第三网络设备利用本地训练样本训练生成的。本地训练样本包括第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及所述历史能耗值所对应的配置参数。
在一种可能的实现方式中，该装置还包括生成单元。该生成单元用于利用本地训练样本生成第一能耗预测模型。发送单元还用于将第一能耗预测模型的模型参数发送给第三网络设备。接收单元还用于接收第三网络设备发送的能耗预测模型。该能耗预测模型由第三网络设备根据多个第一网络设备分别发送的第一能耗预测模型的模型参数确定的。本地训练样本包括第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及所述历史能耗值所对应的配置参数。
在一种可能的实现方式中，预设条件包括目标节能策略对应的能耗值为多个节能策略对应的多个能耗值中的最小值，或者，目标节能策略对应的能耗值小于等于能耗阈值。
第五方面，本申请提供了一种网络设备运行状态的调整装置。该装置应用于第二网络设备。该装置包括发送单元、接收单元、发送单元和应用单元。发送单元用于向第一网络设备发送第一流量信息。第一流量信息指示第二网络设备在第一时间段内所处理流量的值。接收单元用于接收第一网络设备发送的目标节能策略。目标节能策略是由第一网络设备根据第二流量信息确定的。第二流量信息是由第一网络设备根据第一流量信息预测的第二网络设备在第二时间段内所对应的流量。所述第二时间段晚于所述第一时间段。目标节能策略包括第二网络设备中的器件根据所述目标节能策略运行时所对应的配置参数。目标节能策略所对应的能耗满足预设条件。应用单元用于应用所述目标节能策略所对应的配置参数。
在一种可能的实现方式中，接收单元还用于接收所述第一网络设备发送的计划执行时间。该计划执行时间指示目标节能策略的执行时间。该计划执行时间是由所述第一网络设备根据所述第二流量信息以及所述目标节能策略确定的。
在一种可能的实现方式中，接收单元还用于接收第一网络设备发送的唤醒值。该唤醒值指示第二网络设备终止执行目标节能策略的条件。该唤醒值是由第一网络设备根据目标节能策略确定的。
在一种可能的实现方式中，该装置还包括终止单元。终止单元用于在第二网络设备执行目标节能策略且第二网络设备的统计值超过唤醒值时，终止执行该目标节能策略。
在一种可能的实现方式中，该装置还包括确定单元。该确定单元用于确定第二网络设备在执行目标节能策略时的传输性能值。终止单元用于在该传输性能值大于传输性能阈值时，终止执行该目标节能策略。
在一种可能的实现方式中，发送单元还用于向第一网络设备发送目标节能策略的实际执行时间。该实际执行时间对应的执行时长小于所述计划执行时间对应的执行时长。
在一种可能的实现方式中，发送单元还用于向第一网络设备发送第三流量信息。该第三流量信息指示引发目标节能策略终止的流量。
第六方面，本申请提供了一种网络设备。该网络设备包括处理器和存储器。存储器用于存储指令或计算机程序。处理器用于执行该存储器中的指令或计算机程序，以使得该网络设备执行第一方面或第一方面中任一可能的实现方式所述的网络设备运行状态调整方法，或者执行第二方面或第二方面中任一可能的实现方式所述的网络设备运行状态调整方法。
第七方面，本申请提供了一种计算机可读存储介质。该存储介质包括指令。当指令在计算机上运行时，使得计算机执行第一方面或第一方面中任一可能的实现方式所述的网络设备运行状态调整方法，或者执行第二方面或第二方面中任一可能的实现方式所述的网络设备运行状态调整方法。
第八方面，本申请提供了一种计算机程序产品。该计算机程序产品包括程序或代码。程序或代码在计算机上运行时，使得该计算机实现如第一方面或第一方面中任一可能的实现方式所述的网络设备运行状态调整方法，或者实现第二方面或第二方面中任一可能的实现方式所述的网络设备运行状态调整方法。
通过本申请提供的技术方案，第二网络设备可以向第一网络设备发送自身在第一时间段内所处理的流量信息，即第一流量信息。第一网络设备在接收到第一流量信息后，根据第一流量信息预测第二网络设备在第二时间段内对应的第二流量信息，其中，第二时间段晚于第一时间段。在预测出第二流量信息后，第一网络设备根据第二流量信息确定多个节能策略中每个节能策略对应的能耗。其中，每个节能策略包括第二网络设备中的器件根据节能策略运行时所对应的配置参数。在确定出每个节能策略对应的能耗后，第一网络设备将满足预设条件的能耗对应的节能策略确定为目标节能策略，并发送给第二网络设备，以使得第二网络设备按照目标节能策略的配置参数运行。也就是，第一网络设备通过预测第二网络设备需要在第二时间段内处理的第二流量信息，确定与该流量信息匹配的目标节能策略，并将该目标节能策略发送第二网络设备，从而使得第二网络设备按照与待处理的流量匹配的目标节能策略配置内部器件的参数。从而，该方案使得第二网络设备在低流量时间段内消耗低的能耗，减少能耗。
附图说明
图1为本申请实施例提供的一种网络设备运行状态的调整方法流程图；
图2为本申请实施例提供的一种应用场景示意图；
图3为本申请实施例提供的一种网络设备运行状态的调整装置的结构示意图；
图4为本申请实施例提供的另一种网络设备运行状态的调整装置的结构示意图；
图5为本申请实施例提供的一种网络设备的结构示意图；
图6为本申请实施例提供的另一种网络设备的结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本申请中的方案，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅是本申请一部分实施例，而不是全部的实施例。
随着网络建设规模的不断扩大，网络能耗增加所导致的运营成本增加已成为运营商面临的棘手问题。网络设备的能耗与网络设备中各器件的配置参数相关。然而，网络设备的配置参数通常保持在高配状态，这导致网络设备持续产生高能耗。
基于此，本申请提供了一种网络设备运行状态的调整方法，用于实现从多个节能策略中选出与网络设备待处理的流量匹配的节能策略，并将该匹配的节能策略发送给网络设备，以使得网络设备按照该节能策略中的配置参数运行。网络设备按照与待处理流量匹配的配置参数运行，使得网络设备既可以正常处理高流量，又可以在流量降低时，减少产生的能耗。
为便于理解本申请实施例提供的技术方案，下面将结合附图进行说明。
参见图1，该图为本申请实施例提供的一种网络设备运行状态的调整方法流程图，如图1所示，该方法包括：
S101：第二网络设备向第一网络设备发送第一流量信息。
本实施例中，第二网络设备可以收集自身在第一时间段内所处理的流量信息，即第一流量信息，并将该第一流量信息发送给第一网络设备。其中，第一流量信息用于指示第二 网络设备在第一时间段内所处理流量的值。具体地，第一流量信息可以包括：第二网络设备在第一时间段内的发送速率和/或接收速率，或者第二网络设备在第一时间段内接收和/或发送的数据量等。其中，发送速率可以是平均发送速率、最大发送速率等速率统计值。数据量可以是比特数、字节数、报文数等。
在具体实现时，第二网络设备所发送的第一流量信息可以是设备级的流量信息，也可以是单板级的流量信息，还可以是接口级的流量信息。例如，当第一流量信息包括第二网络设备的某一接口的流量信息时，第一网络设备可以基于该接口的流量信息针对性的调整该接口的参数，可以实现对接口参数更准确的调整。
S102：第一网络设备接收第二网络设备发送的第一流量信息，并根据第一流量信息预测第二网络设备在第二时间段内对应的第二流量信息。
第一网络设备在接收到第二网络设备发送的第一流量信息后，将根据该第一流量信息预测第二网络设备在第二时间段内对应的第二流量信息。其中，第二时间段晚于第一时间段。即，第一网络设备可以根据历史时间段的流量信息预测未来流量的趋势。具体地，第一网络设备可以将第一流量信息输入流量预测模型，以获得该流量预测模型输出的第二流量信息。该流量预测模型可以是预设的模型，例如，第一网络设备接收一个流量预测模型。该接收的流量预测模型可以是其他网络设备训练的，也可以是管理员配置的。该流量预测模型还可以是第一网络设备根据第二网络设备的历史流量信息预先训练生成的。当需要进行流量预测时，第一网络设备将第一流量信息输入流量预测模型，以通过该流量预测模型获得第二流量信息。其中，历史流量信息是指第二网络设备在历史时间段内的不同时间段所处理的流量值。
其中，流量预测模型可以为回归预测模型、神经网络模型等。当流量预测模型为神经网络模型时，第一网络设备将获取的历史流量信息作为训练样本，每个训练样本包括M个流量序列和N个流量序列。每个流量序列包括一个或多个流量值以及该一个或多个流量值对应的时间。其中，N个流量序列对应的时间晚于M个流量序列对应的时间。N个流量序列为标签。在每轮的迭代训练过程中，第一网络设备可以将一个训练样本输入至神经网络模型中，由神经网络模型输出针对该训练样本中M个流量序列的推理结果(预测的流量序列/值)。然后，第一网络设备可以通过相应的损失函数，计算神经网络模型输出的推理结果与该组训练样本的实际结果(标签)之间的损失值。然后，第一网络设备可以根据计算出的损失值，计算出神经网络模型中各个网络层中参数的变化梯度。这样，第一网络设备可以基于优化器中预先设置的超参数以及各个网络层中参数的变化梯度，计算出该参数在这一轮迭代训练过程中的调整值(也可称之为参数更新量)，该调整值例如可以是变化梯度与超参数(如学习率等)的乘积等，从而第一网络设备可以基于计算出的各个参数的调整值更新该参数的参数值。经过上述多次训练后，当损失值小于预设阈值时，则停止训练，获得流量预测模型。
S103：第一网络设备根据第二流量信息确定多个节能策略中每个节能策略对应的能耗。
当第一网络设备预测出第二网络设备在第二时间段内可能处理的流量为第二流量信息时，第一网络设备根据第二流量信息确定多个节能策略中每个节能策略对应的能耗。其中， 每个节能策略包括第二网络设备中的器件根据该节能策略运行时对应的配置参数。例如，节能策略包括但不限于中央处理器(central processing unit,CPU)核的开关状态、核的频率、端口/交换板的开关状态、端口/交换板的休眠状态等各种配置参数。
其中，多个节能策略为预先配置的，该多个节能策略中每个节能策略包括的配置参数中部分或全部配置参数不同。具体地，该多个节能策略可以为从预先配置的节能策略集合中选出的一组节能策略。
其中，第一网络设备可以通过以下方式确定每个节能策略对应的能耗，具体包括：针对每个节能策略，第一网络设备将第二流量信息以及节能策略输入能耗预测模型，获得该能耗预测模型输出的对应节能策略的能耗。其中，能耗预测模型是根据训练样本生成的，每个训练样本包括流量信息、能耗值以及能耗值对应的配置参数。即，第一网络设备可以利用预先训练生成的能耗预测模型确定每个节能策略对应的能耗大小。其中，训练样本可以来自于第一网络设备，也可以来自于其他网络设备。其他网络设备与第一网络设备的设备类型相同，例如，其他网络设备和第一网络设备的型号相同。一个训练样本包括网络设备在一个历史时间段处理的流量信息、对应于该历史时间段的能耗值以及对应于该历史时间段的配置参数。
其中，关于能耗预测模型的训练可以采用以下方式实现：
一种是，第一网络设备获取本地训练样本，并利用本地训练样本训练能耗预测模型。其中，本地训练样本是指第一网络设备所管理的一个或多个网络设备提供的训练样本。具体地，本地训练样本包括第一网络设备所管理的一个或多个网络设备提供的训练样本。该多个网络设备具备相同或相似的设备类型，例如，该多个网络设备的型号相同。同一设备类型的网络设备对应于一个能耗预测模型，不同设备类型的网络设备对应于不同的能耗预测模型。一个训练样本包括一个网络设备在一个历史时间段处理的流量信息、对应于该历史时间段的能耗值以及对应于该历史时间段的配置参数。即，在该实现方式中，第一网络设备自身可以根据获取的本地训练样本完成能耗预测模型的训练。其中，第一网络设备所管理的一个或多个网络设备中可以包括第二网络设备，也就是，本地训练样本可以包括第二网络设备的历史流量信息、历史能耗值以及历史能耗值对应的配置参数。
一种是，第一网络设备在获取到本地训练样本后，向第三网络设备发送本地训练样本，以使得第三网络设备利用接收到的训练样本训练能耗预测模型。第一网络设备接收第三网络设备发送的能耗预测模型。其中，本地训练样本包括第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及历史能耗值对应的配置参数。需要说明的是，第一网络设备所管理的网络设备中可以包括第二网络设备，也就是，本地训练样本可以包括第二网络设备的历史流量信息、历史能耗值以及历史能耗值对应的配置参数。即，在该实现方式中，由第三网络设备利用第一网络设备上报的本地训练样本训练能耗预测模型，并将训练完成的能耗预测模型发送给第一网络设备。其中，训练完成的能耗预测模型可以学习到流量信息、配置参数以及能耗值之间的关联关系，进而在利用能耗预测模型确定每个节能策略的能耗时，将第二流量信息以及节能策略对应的配置参数输入能耗预测模型中，通过能耗预测模型确定该节能策略对应的能耗值。
当第三网络设备对应多个第一网络设备时，第三网络设备可以接收多个第一网络设备分别发送的本地训练样本，从而利用大量的本地训练样本训练生成能耗预测模型，提高能耗预测模型预测的准确度。
另一种是，第一网络设备利用本地训练样本训练生成第一能耗预测模型，并将第一能耗预测模型的模型参数发送给第三网络设备，以使得第三网络设备根据多个第一网络设备分别发送的第一能耗预测模型的模型参数确定第二能耗预测模型。第一网络设备接收第三网络设备发送的第二能耗预测模型，并将该第二能耗预测模型用于预测每个节能策略的能耗。其中，本地训练样本包括第一网络设备管理的网络设备的历史流量信息、历史能耗值以及历史能耗值对应的配置参数。在该训练方式中，第三网络设备管理的多个第一网络设备中每个第一网络设备首先利用本地训练样本生成第一能耗预测模型，并将该第一能耗预测模型的模型参数发送给第三网络设备。第三网络设备在接收到多个第一网络设备各自发送的模型参数后，根据该多个模型参数确定能耗预测模型的模型参数，从而生成能耗预测模型。即，在该实现方式中，第三网络设备和多个第一网络设备通过联邦学习的方式确定能耗预测模型。
其中，同一设备类型的网络设备对应于一个能耗预测模型，不同设备类型的网络设备对应于不同的能耗预测模型。针对一个能耗预测模型，第一网络设备或第三网络设备基于来自于相同类型的网络设备的训练样本训练该能耗预测模型。一个训练样本包括具备该设备类型的一个网络设备在一个历史时间段处理的流量信息、对应于该历史时间段的能耗值以及对应于该历史时间段的配置参数。基于此，第一网络设备在将第二流量信息以及节能策略对应的配置参数输入能耗预测模型之前，第一网络设备将根据第二网络设备的设备类型确定与该设备类型对应的能耗预测模型，从而利用该能耗预测模型预测节能策略对应的能耗。其中，同类型的网络设备例如是具有相同型号的网络设备。
其中，能耗预测模型可以为回归预测模型、神经网络模型等。当能耗预测模型为神经网络时，网络设备(第一网络设备或第三网络设备)获取本地训练样本，每个本地训练样本包括一个网络设备在一个历史时间段处理的流量信息、对应于该历史时间段的能耗值以及对应于该历史时间段的配置参数，其中能耗值作为标签。在每轮的迭代训练过程中，网络设备可以将一个训练样本中的流量信息和配置参数输入至神经网络模型中，由神经网络模型输出针对该训练样本的推理结果(能耗值)。然后，网络设备可以通过相应的损失函数，计算神经网络模型输出的推理结果与该训练样本的实际结果(标签)之间的损失值。然后，网络设备可以根据计算出的损失值，计算出神经网络模型中各个网络层中参数的变化梯度。这样，网络设备可以基于优化器中预先设置的超参数以及各个网络层中参数的变化梯度，计算出该参数在这一轮迭代训练过程中的调整值(也可称之为参数更新量)，该调整值例如可以是变化梯度与超参数(如学习率等)的乘积等，从而网络设备可以基于计算出的各个参数的调整值更新该参数的参数值。经过上述多次训练后，当损失值小于预设阈值时，则停止训练，获得能耗预测模型。
其中，第一网络设备可以为控制器、第三网络设备为云端设备，控制器和云端设备可以集成在同一个硬件设备，也可以为两个独立的硬件设备，还可以为两个独立的虚拟设备。
S104：第一网络设备将满足预设条件的能耗对应的节能策略确定为目标节能策略。
在第一网络设备确定出多个节能策略中每个节能策略对应的能耗后，将满足预设条件的能耗对应的节能策略确定为目标节能策略。其中，预设条件可以根据实际应用情况进行设定，例如能耗值小于预设能耗阈值，或者能耗值最小等。其中，确定出的目标节能策略可以包括一个或多个节能策略。第一网络设备可以将第二时间段划分为多个子时间段，分别为多个子时间段确定对应的节能策略，此时目标节能策略包括多个节能策略。例如，第二时间段为凌晨1点-7点，每间隔2个小时划分一个子时间段，则凌晨1点-3点对应节能策略1、3点-5点对应节能策略2、5点-7点对应节能策略3，则目标节能策略包括节能策略1、节能策略2和节能策略3。或者，第一网络设备可以为多个第二时间段确定节能策略，即，针对每个第二时间段确定一个目标节能策略。例如，凌晨1点-7点期间为第1个第二时间段，7点-13点为第2个第二时间段，13点-19点为第3个第二时间段，第一网络设备分别为该3个第二时间段确定节能策略，此时，目标节能策略可以包括该3个节能策略。
S105：第一网络设备将目标节能策略发送给第二网络设备。
第一网络设备可以和第二网络设备约定目标节能策略的执行时间，例如，第一网络设备和第二网络设备约定在接收到节能策略的下个整点时间时应用节能策略。例如，第二网络设备在13:30接收到节能策略，则将在14:00应用该节能策略。
可选地，第一网络设备还可以根据第二流量信息以及目标节能策略确定目标节能策略对应的计划执行时间。例如，当目标节能策略仅包括一个节能策略时，第一网络设备将第二流量信息对应的时段确定为目标节能策略的执行时间；当目标节能策略包括多个节能策略时，第一网络设备将该目标节能策略中的每个节能策略对应的流量信息对应的时段确定为对应的节能策略的执行时间。第一网络设备将计划执行时间发送给第二网络设备，以使得第二网络设备在计划执行时间对应的时间段内执行目标节能策略。其中，计划执行时间可以包括第二网络设备执行目标节能策略的开始时间和终止时间，或者计划执行时间可以包括开始时间以及执行时长等信息，开始时间用于指示第二网络设备开始执行目标节能策略的时间，终止时间用于指示第二网络设备停止执行目标节能策略的时间，开始时间早于终止时间。
S106：第二网络设备接收目标节能策略，并应用目标节能策略对应的配置参数。
在第一网络设备确定出目标节能策略后，第一网络设备将目标节能策略发送给第二网络设备，以使得第二网络设备应用该目标节能策略对应的配置参数。
其中，第二网络设备在接收到目标节能策略时，可以按照预设配置的节能时间应用目标节能策略对应的配置参数。或者，第二网络设备可以按照第一网络设备发送的计划执行时间应用目标节能策略对应的配置参数，在计划执行时间外的时间段不应用目标节能策略。其中，第一网络设备可以通过同一消息发送目标节能策略和计划执行时间，也可以通过不同消息发送目标节能策略、计划执行时间。
此外，第一网络设备还可以根据目标节能策略确定唤醒值，该唤醒值指示第二网络设备终止执行目标节能策略的条件。该唤醒值可以为流量阈值，当第二网络设备处理的流量超过流量阈值时，则终止执行目标节能策略。或者，唤醒值可以为性能阈值，例如时延阈 值、抖动阈值、丢包率阈值等，当第二网络设备执行目标节能策略时的性能值超过性能阈值时，终止执行目标节能策略。其中，流量阈值可以为第二网络设备中各个接口对应的阈值，也可以为整个第二网络设备的阈值。性能阈值也可以为第二网络设备中各个接口的阈值，也可以为整个第二网络设备的阈值。
在具体实现中，当第一网络设备确定出目标节能策略后，可以根据目标节能策略对应的配置参数计算出其对应的唤醒值。例如，第二网络设备包括8个端口，目标节能策略中指示关闭端口1、端口3和端口4，则第一网络设备根据剩余端口的传输带宽计算吞吐量阈值。或者，预先配置每个节能策略与其对应的唤醒值之间的关联关系，在确定目标节能策略后，根据上述关联关系确定匹配的唤醒值。
在一些应用场景下，在第二网络设备执行目标节能策略时，可能会发生流量异常的情况，使得第二网络设备的统计值超过唤醒值，该情况下，第二网络设备将提前终止执行目标节能策略。同时，第二网络设备可以向第一网络设备发送该目标节能策略的实际执行时间，该实际执行时间对应的执行时长小于计划执行时间对应的执行时长。当第二网络设备提前终止执行目标节能策略后，可以按照执行目标节能策略之前各器件的配置参数继续运行，或者按照预先存储的默认配置参数继续运行。
第一网络设备在接收到第二网络设备发送的实际执行时间后，可以根据实际执行时间更新目标节能策略。其中，更新目标节能策略可以包括更新目标节能策略对应的配置参数。例如，更新前的节能策略中CPU降频50％，更新后的CPU降频为45％；更新前的节能策略为关闭某一端口+关闭某个交换网板+关闭CPU某个核，更新后的节能策略为关闭某一端口+关闭某个交换网版。
另外，第二网络设备还可以向第一网络设备发送第三流量信息，该第三流量信息对应的流量值超过唤醒值。第一网络设备根据第三流量信息优化流量预测模型，从而使得流量预测模型的预测结果更为准确。其中，第三流量信息可以包括指示流量大小的时间序列、流量均值、流量方差等统计值。
在一种应用场景下，第一网络设备也可以不向第二网络设备发送唤醒值，而是由第二网络设备在执行目标节能策略时，根据当前的传输性能值确定是否需要终止执行目标节能策略。具体地，在第二网络设备执行目标节能策略时，第二网络设备确定传输性能值；在传输性能值大于传输性能阈值时，第二网络设备终止执行目标节能策略。其中，传输性能值可以反映第二网络设备的传输质量，可以包括传输时延、丢包率等。例如，当第二网络设备的传输时延大于时延阈值时，第二网络设备终止执行目标节能策略。又例如，当第二网络设备的丢包率大于丢包率阈值时，第二网络设备终止执行目标节能策略。再例如，当第二网络设备的丢包率大于预设丢包率阈值且传输时延也大于预设时延阈值时，第二网络设备终止执行目标节能策略。在终止执行目标节能策略后，第二网络设备向第一网络设备发送目标节能策略的实际执行时间，以使得第一网络设备可以获知目标节能策略提前终止。第二网络设备还可以向第一网络设备发送第三流量信息，该第三流量信息指示引发目标节能策略终止的流量，以使得第一网络设备根据第三流量信息优化流量预测模型。
可见，第二网络设备可以向第一网络设备上报自身在第一时间段内所处理的流量信息， 即第一流量信息。第一网络设备在接收到第一流量信息后，根据第一流量信息预测第二网络设备在第二时间段内对应的第二流量信息，其中，第二时间段晚于第一时间段。在预测出第二流量信息后，第一网络设备根据第二流量信息确定多个节能策略中每个节能策略对应的能耗。其中，每个节能策略包括第二网络设备中的器件根据节能策略运行时对应的配置参数。在确定出每个节能策略对应的能耗后，第一网络设备将满足预设条件的能耗对应的节能策略确定为目标节能策略，并发送给第二网络设备，以使得第二网络设备按照目标节能策略的配置参数运行。也就是，第一网络设备通过预测第二网络设备在第二时间段内处理的第二流量信息，确定与该流量信息匹配的目标节能策略，并将该目标节能策略发送给第二网络设备，从而使得第二网络设备按照与待处理流量匹配的目标节能策略配置内部器件的参数。从而，该第二网络设备无需始终保持高配置，而可以灵活的应用与待处理流量匹配的配置，降低在低流量时间段内产生的能耗。
图2是本申请实施例提供的一种应用场景示意图。参见图2所示的应用场景示意图，该应用场景包括云端设备、分析器、控制器和网络设备。云端设备可以部署于公有云、边缘云或者分布式云。网络设备可以包括网络中的转发设备、终端设备。控制器可以从网络设备中采集数据并发送给分析器，以使得分析器根据采集的数据对网络设备处理的流量信息进行预测，进而确定出节能策略和执行该节能策略的时间。分析器通过控制器将确定的节能策略以及对应的执行时间发送给网络设备，网络设备应用节能策略对应的配置参数，从而降低网络设备的能耗。
在图2所示的应用场景中，控制器可以采集训练样本，并通过分析器将训练样本发送给云端设备，由云端设备利用训练样本训练能耗预测模型。一个训练样本包括一个网络设备在一个历史时间段的流量信息、对应于该历史时间段的能耗值以及对应于该历史时间段的配置参数。云端设备基于相同设备类型的网络设备的多个训练样本训练生成针对该设备类型的能耗预测模型，将该能耗预测模型发送给分析器。在实际的应用过程中，控制器采集网络设备在第一时间段对应的第一流量信息，并将该第一流量信息发送给分析器，该分析器利用第一流量信息以及流量预测模型获得网络设备在第二时间段对应的第二流量信息。分析器将第二流量信息和一个节能策略对应的配置参数输入能耗预测模型中，确定该节能策略对应的能耗，并将满足预设条件的能耗对应的节能策略确定为目标节能策略。同时，分析器还可以确定该目标节能策略对应的计划执行时间以及唤醒值，并将目标节能策略、计划执行时间以及唤醒值发送给网络设备。网络设备中的策略执行模块根据计划执行时间执行目标节能策略，以降低能耗。在目标节能策略执行过程中，网络设备中的唤醒检测模块用于判断网络设备当前处理的流量是否大于唤醒值，如果是，则可以向策略执行模块发送异常消息，以使得策略执行模块终止执行目标节能策略。网络设备还可以将实际执行时间和异常流量信息通过控制器发送给分析器，以使得分析器可以根据实际执行时间和异常流量信息更新节能策略以及流量预测模型等。
其中，分析器和控制器可以是同一物理设备，也可以为两个独立的物理设备。或者，分析器和云端设备可以是同一个物理设备，或者是两个独立的物理设备。在实际应用时，云端设备可以收集多个分析器发送的训练样本，以利用大量的训练样本训练能耗预测模型， 提高能耗预测模型预测准确度。
基于上述方法实施例，本申请实施例还提供了一种网络设备运行状态的调整装置，下面将结合附图进行说明。
参见图3，该图为本申请实施例提供的一种网络设备运行状态的调整装置300的结构示意图。该装置300可以用于实现上述第一网络设备的功能。如图3所示，该装置300包括接收单元301、预测单元302、确定单元303和发送单元304。
接收单元301，用于接收第二网络设备发送的第一流量信息。第一流量信息指示第二网络设备在第一时间段内所处理流量的值。
预测单元302，用于根据第一流量信息预测第二网络设备在第二时间段内所对应的第二流量信息。第二时间段晚于第一时间段。
确定单元303，用于根据第二流量信息确定多个节能策略中每个节能策略对应的能耗值。每个节能策略包括第二网络设备中的器件根据该节能策略运行时所对应的配置参数。
确定单元303，还用于将满足预设条件的能耗值对应的节能策略确定为目标节能策略。
发送单元304，用于将目标节能策略发送给第二网络设备，以使得第二网络设备按照目标节能策略中的配置参数运行。
可选地，确定单元303还用于根据第二流量信息以及目标节能策略确定该目标节能策略对应的计划执行时间。发送单元304还用于将该计划执行时间发送给第二网络设备，以使得第二网络设备在该计划执行时间对应的时间段内执行该目标节能策略。
可选地，确定单元303还用于根据该目标节能策略确定唤醒值。该唤醒值指示第二网络设备终止执行该目标节能策略的条件。发送单元304还用于将该唤醒值发送给第二网络设备。
可选地，预测单元302用于将第一流量信息输入流量预测模型，以获得该流量预测模型输出的第二流量信息。该流量预测模型是根据第二网络设备的历史流量信息训练生成的。
可选地，确定单元303用于，对于每个节能策略，将第二流量信息以及该节能策略对应的配置参数输入能耗预测模型，以获得该能耗预测模型输出的对应该节能策略的能耗。该能耗预测模型是根据训练样本生成的。每个训练样本包括一个网络设备在一个历史时段内的流量信息、对应于该历史时间段内的该网络设备的能耗值以及对应于该历史时间段内的该网络设备的配置参数。该网络设备为第二网络设备，或者为与第二网络设备的设备类型相同的网络设备。
可选地，确定单元303还用于在将第二流量信息以及节能策略对应的配置参数输入能耗预测模型之前，根据第二网络设备的设备类型确定该能耗预测模型。该设备类型与该能耗预测模型对应。
可选地，该装置300还包括更新单元。接收单元301还用于接收第二网络设备发送的目标节能策略的实际执行时间。该实际执行时间对应的执行时长小于该目标节能测量的计划执行时间对应的执行时长。该更新单元用于根据该实际执行时间更新该目标节能策略。
可选地，该装置300还包括优化单元。接收单元301还用于接收第二网络设备发送的第三流量信息。第三流量信息对应的统计值超过唤醒值。优化单元，用于根据第三流量信 息优化流量预测模型。
可选地，发送单元304还用于向第三网络设备发送训练样本。一个训练样本包括第一网络设备所管理的一个网络设备在一个历史时间段内的流量信息、对应于该历史时间段内的该网络设备的能耗值以及对应于该历史时间段内的该网络设备的的配置参数。接收单元301还用于接收第三网络设备发送的能耗预测模型。该能耗预测模型是由第三网络设备利用上述训练样本训练生成的。
可选地，该装置300还包括生成单元。该生成单元用于利用训练样本生成第一能耗预测模型。发送单元304还用于将该第一能耗预测模型的模型参数发送给第三网络设备。接收单元301还用于接收第三网络设备发送的能耗预测模型。该能耗预测模型是由第三网络设备根据多个第一网络设备分别发送的第一能耗预测模型的模型参数确定的。
可选的，预设条件包括目标节能策略对应的能耗值为多个节能策略对应的多个能耗值中的最小值，或者，目标节能策略对应的能耗值小于等于能耗阈值。
需要说明的是，本实施例中各个单元的实现可以参见上述方法实施例中的相关描述，本实施例在此不再赘述。
参见图4，该图为本申请实施例提供的一种网络设备运行状态的调整装置400的结构示意图。该装置400可以用于实现第二网络设备的功能。该装置400包括发送单元401、接收单元402和应用单元403。
发送单元401，用于向第一网络设备发送第一流量信息。第一流量信息指示第二网络设备在第一时间段内所处理流量的值。
接收单元402，用于接收第一网络设备发送的目标节能策略。该目标节能策略是由第一网络设备根据第二流量信息确定的。该第二流量信息是由第一网络设备根据第一流量信息预测的第二网络设备在第二时间段内所对应的流量。第二时间段晚于第一时间段。该目标节能策略包括第二网络设备中的器件根据该目标节能策略运行时所对应的配置参数。该目标节能策略所对应的能耗满足预设条件。
应用单元403，用于应用目标节能策略所对应的配置参数。
可选地，接收单元402还用于接收第一网络设备发送的计划执行时间。计划执行时间指示目标节能策略的执行时间。该计划执行时间是由第一网络设备根据第二流量信息以及目标节能策略确定的。
可选地，接收单元402还用于接收第一网络设备发送的唤醒值。该唤醒值指示第二网络设备终止执行该目标节能策略的条件。该唤醒值是由第一网络设备根据目标节能策略确定的。
可选地，该装置400还包括终止单元。该终止单元用于在第二网络设备执行目标节能策略且第二网络设备的统计值超过唤醒值时，终止执行该目标节能策略。
可选地，该装置400还包括确定单元。该确定单元用于在第二网络设备执行目标节能策略时，确定传输性能值。终止单元用于在传输性能值大于传输性能阈值时，终止执行所述目标节能策略。
可选地，发送单元401还用于向第一网络设备发送目标节能策略的实际执行时间。该 实际执行时间对应的执行时长小于计划执行时间对应的执行时长。
可选地，发送单元401还用于向第一网络设备发送第三流量信息。第三流量信息指示引发所述目标节能策略终止的流量。
需要说明的是，本实施例中各个单元的实现可以参见上述方法实施例中的相关描述，本实施例在此不再赘述。
图5为本申请实施例提供的一种网络设备的结构示意图。该网络设备例如可以是上述方法实施例中的第一网络设备、第二网络设备或第三网络设备，或者也可以是图3所示实施例中装置300的设备实现，或者也可以是图4所示实施例中装置400的设备实现。
该网络设备500包括：处理器510、通信接口520和存储器530。网络设备500中的处理器510的数量可以一个或多个，图5中以一个处理器为例。本申请实施例中，处理器510、通信接口520和存储器530可通过总线系统或其它方式连接，其中，图5中以通过总线系统540连接为例。
处理器510可以是CPU、网络处理器(network processor,NP)、或者CPU和NP的组合。处理器510还可以进一步包括硬件芯片。上述硬件芯片可以是应用专用集成电路(application-specific integrated circuit，ASIC)，可编程逻辑器件(programmable logic device，PLD)或其组合。上述PLD可以是复杂可编程逻辑器件(complex programmable logic device，CPLD)，现场可编程逻辑门阵列(field-programmable gate array，FPGA)，通用阵列逻辑(generic array logic，GAL)或其任意组合。
存储器530可以包括易失性存储器(volatile memory)，例如随机存取存储器(random-access memory，RAM)；存储器530也可以包括非易失性存储器(non-volatile memory)，例如快闪存储器(flash memory)，硬盘(hard disk drive，HDD)或固态硬盘(solid-state drive，SSD)；存储器530还可以包括上述种类的存储器的组合。
可选地，存储器530存储有操作系统和程序、可执行模块或者数据结构，或者它们的子集，或者它们的扩展集，其中，程序可包括各种操作指令，用于实现各种操作。操作系统可包括各种系统程序，用于实现各种基础业务以及处理基于硬件的任务。处理器510可以读取存储器530中的程序，实现本申请实施例提供的方法。
其中，存储器530可以为网络设备500中的存储器件，也可以为独立于网络设备500的存储装置。
总线系统540可以是外设部件互连标准(peripheral component interconnect，PCI)总线或扩展工业标准结构(extended industry standard architecture，EISA)总线等。总线系统540可以分为地址总线、数据总线、控制总线等。为便于表示，图5中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
图6是本申请实施例提供的一种网络设备600的结构示意图，该网络设备例如可以是方法实施例中的第一网络设备、第二网络设备或第三网络设备，或者也可以是图3所述实施例中装置300的设备实现，或者也可以为图4所述实施例中装置400的设备实现。
网络设备600包括：主控板610和接口板630。
主控板610也称为主处理单元(main processing unit，MPU)或路由处理卡(route  processor card)，主控板610用于对网络设备600中各个组件进行控制和管理，包括路由计算、设备管理、设备维护、协议处理等。主控板610包括：中央处理器611和存储器612。
接口板630也称为线路处理单元(line processing unit，LPU)、线卡(line card)或业务板。接口板630用于提供各种业务接口并实现数据包的转发。业务接口包括而不限于以太网接口、POS(Packet over SONET/SDH)接口等，以太网接口例如是以太口、千兆以太口、灵活以太网业务接口(Flexible Ethernet Clients，FlexE Clients)等。接口板630包括：中央处理器631、网络处理器632、转发表项存储器634和物理接口卡(physical interface card，PIC)633。
接口板630上的中央处理器631用于对接口板630进行控制管理并与主控板610上的中央处理器611进行通信。
网络处理器632用于实现报文的转发处理。网络处理器632的形态可以是转发芯片。具体而言，上行报文的处理包括：报文入接口的处理，转发表查找；下行报文的处理包括转发表查找等等。
物理接口卡633用于实现物理层的对接功能，原始的流量由此进入接口板630，以及处理后的报文从该物理接口卡633发出。物理接口卡633包括至少一个物理接口，物理接口也称物理口。物理接口卡633也称为子卡，可安装在接口板630上，负责将光电信号转换为报文并对报文进行合法性检查后转发给网络处理器632处理。在一些实施例中，接口板603的中央处理器631也可执行网络处理器632的功能，比如基于通用CPU实现软件转发，从而物理接口卡633中不需要网络处理器632。
可选地，网络设备600包括多个接口板，例如网络设备600还包括接口板640，接口板640包括：中央处理器641、网络处理器642、转发表项存储器644和物理接口卡643。
可选地，网络设备600还包括交换网板620。交换网板620也可以称为交换网板单元(switch fabric unit，SFU)。在网络设备有多个接口板630的情况下，交换网板620用于完成各接口板之间的数据交换。例如，接口板630和接口板640之间可以通过交换网板620通信。
主控板610和接口板630耦合。例如。主控板610、接口板630和接口板640，以及交换网板620之间通过系统总线与系统背板相连实现互通。在一种可能的实现方式中，主控板610和接口板630之间建立进程间通信协议(inter-process communication，IPC)通道，主控板610和接口板630之间通过IPC通道进行通信。
在逻辑上，网络设备600包括控制面和转发面，控制面包括主控板610和中央处理器631，转发面包括执行转发的各个组件，比如转发表项存储器634、物理接口卡633和网络处理器632。控制面执行路由器、生成转发表、处理信令和协议报文、配置与维护设备的状态等功能，控制面将生成的转发表下发给转发面，在转发面，网络处理器632基于控制面下发的转发表对物理接口卡633收到的报文查表转发。控制面下发的转发表可以保存在转发表项存储器634中。在一些实施例中，控制面和转发面可以完全分离，不在同一设备上。
应理解，本申请实施例中接口板640上的操作与接口板630的操作一致，为了简洁， 不再赘述。应理解，本实施例的网络设备600可对应于上述各个方法实施例中的网络设备，该网络设备600中的主控板610、接口板630和/或接口板640可以实现上述各个方法实施例中的各种步骤，为了简洁，在此不再赘述。
应理解，主控板可能有一块或多块，有多块的时候可以包括主用主控板和备用主控板。接口板可能有一块或多块，网络设备的数据处理能力越强，提供的接口板越多。接口板上的物理接口卡也可以有一块或多块。交换网板可能没有，也可能有一块或多块，有多块的时候可以共同实现负荷分担冗余备份。在集中式转发架构下，网络设备可以不需要交换网板，接口板承担整个系统的业务数据的处理功能。在分布式转发架构下，网络设备可以有至少一块交换网板，通过交换网板实现多块接口板之间的数据交换，提供大容量的数据交换和处理能力。所以，分布式架构的网络设备的数据接入和处理能力要大于集中式架构的设备。可选地，网络设备的形态也可以是只有一块板卡，即没有交换网板，接口板和主控板的功能集成在该一块板卡上，此时接口板上的中央处理器和主控板上的中央处理器在该一块板卡上可以合并为一个中央处理器，执行两者叠加后的功能，这种形态设备的数据交换和处理能力较低(例如，低端交换机或路由器等网络设备)。具体采用哪种架构，取决于具体的组网部署场景。
在一些可能的实施例中，上述网络设备可以实现为虚拟化设备。例如，虚拟化设备可以是运行有用于发送报文功能的程序的虚拟机(virtual machine，VM)，虚拟机部署在硬件设备上(例如，物理服务器)。虚拟机指通过软件模拟的具有完整硬件系统功能的、运行在一个完全隔离环境中的完整计算机系统。可以将虚拟机配置为网络设备。例如，可以基于通用的物理服务器结合网络功能虚拟化(network functions virtualization，NFV)技术来实现网络设备。网络设备为虚拟主机、虚拟路由器或虚拟交换机。本领域技术人员通过阅读本申请即可结合NFV技术在通用物理服务器上虚拟出具有上述功能的网络设备，此处不再赘述。
应理解，上述各种产品形态的网络设备，具有上述方法实施例中网络设备的任意功能，此处不再赘述。
本申请实施例还提供了一种芯片，包括处理器和接口电路。接口电路用于接收指令并传输至处理器。处理器，例如可以是图3示出的调整装置300的一种具体实现形式，可以用于执行上述网络设备运行状态的调整方法。其中，所述处理器与存储器耦合，所述存储器用于存储程序或指令，当所述程序或指令被所述处理器执行时，使得该芯片系统实现上述任一方法实施例中的方法。
可选地，该芯片系统中的处理器可以为一个或多个。该处理器可以通过硬件实现也可以通过软件实现。当通过硬件实现时，该处理器可以是逻辑电路、集成电路等。当通过软件实现时，该处理器可以是一个通用处理器，通过读取存储器中存储的软件代码来实现。
可选地，该芯片系统中的存储器也可以为一个或多个。该存储器可以与处理器集成在一起，也可以和处理器分离设置，本申请并不限定。示例性的，存储器可以是非瞬时性处理器，例如只读存储器ROM，其可以与处理器集成在同一块芯片上，也可以分别设置在不同的芯片上，本申请对存储器的类型，以及存储器与处理器的设置方式不作具体限定。
示例性的，该芯片系统可以例如是FPGA、ASIC、系统芯片(system on chip，SoC)，、 CPU、NP、数字信号处理电路(digital signal processor，DSP)、微控制器(micro controller unit，MCU)、PLD或其他集成芯片。
本申请实施例还提供了一种计算机可读存储介质，包括指令或计算机程序，当其在计算机上运行时，使得计算机执行以上实施例提供的网络设备运行状态的调整方法。
本申请实施例还提供了一种包含指令或计算机程序的计算机程序产品，当其在计算机上运行时，使得计算机执行以上实施例提供的网络设备运行状态的调整方法。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”、“第四”等(如果存在)是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的实施例能够以除了在这里图示或描述的内容以外的顺序实施。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统，装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统，装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，单元的划分，仅仅为一种逻辑业务划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各业务单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件业务单元的形式实现。
集成的单元如果以软件业务单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、ROM、RAM、磁碟或者光盘等各种可以存储程序代码的介质。
本领域技术人员应该可以意识到，在上述一个或多个示例中，本申请所描述的业务可以用硬件、软件、固件或它们的任意组合来实现。当使用软件实现时，可以将这些业务存储在计算机可读介质中或者作为计算机可读介质上的一个或多个指令或代码进行传输。计 算机可读介质包括计算机存储介质和通信介质，其中通信介质包括便于从一个地方向另一个地方传送计算机程序的任何介质。存储介质可以是通用或专用计算机能够存取的任何可用介质。
以上的具体实施方式，对本申请的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上仅为本申请的具体实施方式而已。
以上，以上实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。

Claims (25)

  1. 一种网络设备运行状态的调整方法，其特征在于，所述方法包括：
    第一网络设备接收第二网络设备发送的第一流量信息，所述第一流量信息指示所述第二网络设备在第一时间段内所处理流量的值；
    所述第一网络设备根据所述第一流量信息预测所述第二网络设备在第二时间段内所对应的第二流量信息，所述第二时间段晚于所述第一时间段；
    所述第一网络设备根据所述第二流量信息确定多个节能策略中每个节能策略对应的能耗值，所述每个节能策略包括所述第二网络设备中的器件根据所述节能策略运行时所对应的配置参数；
    所述第一网络设备将满足预设条件的能耗值对应的节能策略确定为目标节能策略，将所述目标节能策略发送给所述第二网络设备，以使得所述第二网络设备按照所述目标节能策略中的配置参数运行。
  2. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    所述第一网络设备根据所述第二流量信息以及所述目标节能策略确定所述目标节能策略对应的计划执行时间；
    所述第一网络设备将所述计划执行时间发送给所述第二网络设备，以使得所述第二网络设备在所述计划执行时间对应的时间段内执行所述目标节能策略。
  3. 根据权利要求1或2所述的方法，其特征在于，所述方法还包括：
    所述第一网络设备根据所述目标节能策略确定唤醒值，所述唤醒值指示所述第二网络设备终止执行所述目标节能策略的条件；
    所述第一网络设备将所述唤醒值发送给所述第二网络设备。
  4. 根据权利要求1-3任一项所述的方法，其特征在于，所述第一网络设备根据所述第一流量信息预测所述第二网络设备在第二时间段内所对应的第二流量信息，包括：
    所述第一网络设备将所述第一流量信息输入流量预测模型，以获得所述流量预测模型输出的所述第二流量信息，所述流量预测模型是根据所述第二网络设备的历史流量信息训练生成的。
  5. 根据权利要求1-4任一项所述的方法，其特征在于，所述第一网络设备根据所述第二流量信息确定多个节能策略中每个节能策略对应的能耗，包括：
    对于每个节能策略，所述第一网络设备将所述第二流量信息以及所述节能策略输入能耗预测模型，以获得所述能耗预测模型输出的对应所述节能策略的能耗，所述能耗预测模型是根据训练样本生成的，每个训练样本包括流量信息、能耗值以及所述能耗值对应的配置参数。
  6. 根据权利要求5所述的方法，其特征在于，在所述第一网络设备将所述第二流量信息以及所述节能策略输入能耗预测模型之前，所述方法还包括：
    所述第一网络设备根据所述第二网络设备的设备类型确定所述能耗预测模型，所述设备类型与所述能耗预测模型对应。
  7. 根据权利要求2所述的方法，其特征在于，所述方法还包括：
    所述第一网络设备接收所述第二网络设备发送的所述目标节能策略的实际执行时间，所述实际执行时间对应的执行时长小于所述计划执行时间对应的执行时长；
    所述第一网络设备根据所述实际执行时间更新所述目标节能策略。
  8. 根据权利要求4所述的方法，其特征在于，所述方法还包括：
    所述第一网络设备接收所述第二网络设备发送的第三流量信息，所述第三流量信息对应的统计值超过唤醒值；
    所述第一网络设备根据所述第三流量信息优化所述流量预测模型。
  9. 根据权利要求5所述的方法，其特征在于，所述方法还包括：
    所述第一网络设备向第三网络设备发送本地训练样本，所述本地训练样本包括所述第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及所述历史能耗值所对应的配置参数；
    所述第一网络设备接收所述第三网络设备发送的所述能耗预测模型，所述能耗预测模型是由所述第三网络设备利用所述本地训练样本训练生成的。
  10. 根据权利要求5所述的方法，其特征在于，所述方法还包括：
    所述第一网络设备利用本地训练样本生成第一能耗预测模型，并将所述第一能耗预测模型的模型参数发送给所述第三网络设备，所述本地训练样本包括所述第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及所述历史能耗值所对应的配置参数；
    所述第一网络设备接收所述第三网络设备发送的所述能耗预测模型，所述能耗预测模型是由所述第三网络设备根据多个第一网络设备分别发送的第一能耗预测模型的模型参数确定的。
  11. 根据权利要求1-10任一项所述的方法，其特征在于，所述预设条件包括所述目标节能策略对应的能耗值为所述多个节能策略对应的多个能耗值中的最小值，或者，所述目标节能策略对应的能耗值小于等于能耗阈值。
  12. 一种网络设备运行状态的调整方法，其特征在于，所述方法包括：
    第二网络设备向第一网络设备发送第一流量信息，所述第一流量信息指示所述第二网络设备在第一时间段内所处理流量的值；
    所述第二网络设备接收所述第一网络设备发送的目标节能策略，所述目标节能策略是由所述第一网络设备根据第二流量信息确定的，所述第二流量信息是由所述第一网络设备根据所述第一流量信息预测的所述第二网络设备在第二时间段内所对应的流量，所述第二时间段晚于所述第一时间段，所述目标节能策略包括所述第二网络设备中的器件根据所述目标节能策略运行时所对应的配置参数，所述目标节能策略所对应的能耗满足预设条件；
    所述第二网络设备应用所述目标节能策略所对应的配置参数。
  13. 根据权利要求12所述的方法，其特征在于，所述方法还包括：
    所述第二网络设备接收所述第一网络设备发送的计划执行时间，所述计划执行时间指示所述目标节能策略的执行时间，所述计划执行时间是由所述第一网络设备根据所述第二流量信息以及所述目标节能策略确定的。
  14. 根据权利要求12或13所述的方法，其特征在于，所述方法还包括：
    所述第二网络设备接收所述第一网络设备发送的唤醒值，所述唤醒值指示所述第二网络设备终止执行所述目标节能策略的条件，所述唤醒值是由所述第一网络设备根据所述目标节能策略确定的。
  15. 根据权利要求14所述的方法，其特征在于，所述方法还包括：
    在所述第二网络设备执行所述目标节能策略且所述第二网络设备的统计值超过所述唤醒值时，所述第二网络设备终止执行所述目标节能策略。
  16. 根据权利要求12或13所述的方法，其特征在于，所述方法还包括：
    在所述第二网络设备执行所述目标节能策略时，所述第二网络设备确定传输性能值；
    在所述传输性能值大于传输性能阈值时，所述第二网络设备终止执行所述目标节能策略。
  17. 根据权利要求15或16所述的方法，其特征在于，所述方法还包括：
    所述第二网络设备向所述第一网络设备发送所述目标节能策略的实际执行时间，所述实际执行时间对应的执行时长小于所述计划执行时间对应的执行时长。
  18. 根据权利要求15-17任一项所述的方法，其特征在于，所述方法还包括：
    所述第二网络设备向所述第一网络设备发送第三流量信息，所述第三流量信息指示引发所述目标节能策略终止的流量。
  19. 一种网络设备运行状态的调整装置，其特征在于，所述装置应用于第一网络设备，包括：
    接收单元，用于接收第二网络设备发送的第一流量信息，所述第一流量信息指示所述第二网络设备在第一时间段内所处理流量的值；
    预测单元，用于根据所述第一流量信息预测所述第二网络设备在第二时间段内所对应的第二流量信息，所述第二时间段晚于所述第一时间段；
    确定单元，用于根据所述第二流量信息确定多个节能策略中每个节能策略对应的能耗值，所述每个节能策略包括所述第二网络设备中的器件根据所述节能策略运行时所对应的配置参数；
    所述确定单元，还用于将满足预设条件的能耗值对应的节能策略确定为目标节能策略；
    发送单元，用于将所述目标节能策略发送给所述第二网络设备，以使得所述第二网络设备按照所述目标节能策略中的配置参数运行。
  20. 一种网络设备运行状态的调整装置，其特征在于，所述装置应用于第二网络设备，包括：
    发送单元，用于向第一网络设备发送第一流量信息，所述第一流量信息指示所述第二网络设备在第一时间段内所处理流量的值；
    接收单元，用于接收所述第一网络设备发送的目标节能策略，所述目标节能策略是由所述第一网络设备根据第二流量信息确定的，所述第二流量信息是由所述第一网络设备根据所述第一流量信息预测的所述第二网络设备在第二时间段内所对应的流量，所述第二时间段晚于所述第一时间段，所述目标节能策略包括所述第二网络设备中的器件根据所述目标节能策略运行时所对应的配置参数，所述目标节能策略所对应的能耗满足预设条件；
    应用单元，用于应用所述目标节能策略所对应的配置参数。
  21. 一种网络系统，其特征在于，所述系统包括：第一网络设备和第二网络设备；
    所述第一网络设备，用于执行权利要求1-11任一项所述的方法；
    所述第二网络设备，用于执行权利要求12-18任一项所述的方法。
  22. 根据权利要求21所述的系统，其特征在于，所述系统还包括：第三网络设备；
    所述第三网络设备，用于接收所述第一网络设备发送的本地训练样本，所述本地训练样本包括所述第一网络设备所管理的网络设备的历史流量信息、历史能耗值以及所述历史能耗值所对应的配置参数；
    所述第三网络设备，还用于根据所述本地训练样本训练生成能耗预测模型，并将所述能耗预测模型发送给所述第一网络设备。
  23. 一种网络设备，其特征在于，所述网络设备包括：处理器和存储器；
    所述存储器，用于存储指令或计算机程序；
    所述处理器，用于执行所述存储器中的所述指令或计算机程序，以使得所述网络设备执行权利要求1-11任意一项所述的网络设备运行状态调整方法，或者执行权利要求12-18任一项所述的网络设备运行状态调整方法。
  24. 一种计算机可读存储介质，其特征在于，包括指令，当其在计算机上运行时，使得计算机执行以上权利要求1-11任意一项所述的网络设备运行状态调整方法，或者执行权利要求12-18任一项所述的网络设备运行状态调整方法。
  25. 一种计算机程序产品，其特征在于，所述计算机程序产品包括程序或代码，所述程序或代码在计算机上运行时，实现如权利要求1至11任一项所述的网络设备运行状态调整方法，或者实现如权利要求12-18任一项所述的网络设备运行状态调整方法。
PCT/CN2023/077060 2022-02-21 2023-02-20 一种网络设备运行状态的调整方法、装置及相关设备 WO2023155904A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210159692.9A CN116668209A (zh) 2022-02-21 2022-02-21 一种网络设备运行状态的调整方法、装置及相关设备
CN202210159692.9 2022-02-21

Publications (1)

Publication Number Publication Date
WO2023155904A1 true WO2023155904A1 (zh) 2023-08-24

Family

ID=87577609

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/077060 WO2023155904A1 (zh) 2022-02-21 2023-02-20 一种网络设备运行状态的调整方法、装置及相关设备

Country Status (2)

Country Link
CN (1) CN116668209A (zh)
WO (1) WO2023155904A1 (zh)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101110701A (zh) * 2007-08-02 2008-01-23 杭州华三通信技术有限公司 一种分布式系统的节能方法、系统和设备
CN101598967A (zh) * 2009-06-23 2009-12-09 中兴通讯股份有限公司 多核防火墙单板及其运行方法
CN101873680A (zh) * 2010-06-25 2010-10-27 华为技术有限公司 一种动态能耗控制方法及系统、相关设备
CN103428282A (zh) * 2013-08-06 2013-12-04 浪潮(北京)电子信息产业有限公司 一种云计算数据中心的在线节能控制方法及装置
WO2014032477A1 (zh) * 2012-08-29 2014-03-06 惠州Tcl移动通信有限公司 一种多核中央处理器的调控方法及系统
WO2017148253A1 (zh) * 2016-02-29 2017-09-08 中兴通讯股份有限公司 一种节能管理实现方法、装置及网络设备

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101110701A (zh) * 2007-08-02 2008-01-23 杭州华三通信技术有限公司 一种分布式系统的节能方法、系统和设备
CN101598967A (zh) * 2009-06-23 2009-12-09 中兴通讯股份有限公司 多核防火墙单板及其运行方法
CN101873680A (zh) * 2010-06-25 2010-10-27 华为技术有限公司 一种动态能耗控制方法及系统、相关设备
WO2014032477A1 (zh) * 2012-08-29 2014-03-06 惠州Tcl移动通信有限公司 一种多核中央处理器的调控方法及系统
CN103428282A (zh) * 2013-08-06 2013-12-04 浪潮(北京)电子信息产业有限公司 一种云计算数据中心的在线节能控制方法及装置
WO2017148253A1 (zh) * 2016-02-29 2017-09-08 中兴通讯股份有限公司 一种节能管理实现方法、装置及网络设备

Also Published As

Publication number Publication date
CN116668209A (zh) 2023-08-29

Similar Documents

Publication Publication Date Title
Zhang et al. Load balancing in data center networks: A survey
CN104081718B (zh) 用于远程系统管理的网络控制器
CN102055667B (zh) 用于实现网络规则的方法和设备
US11574241B2 (en) Adaptive threshold selection for SD-WAN tunnel failure prediction
US11888744B2 (en) Spin-leaf network congestion control method, node, system, and storage medium
CN112422443B (zh) 一种拥塞算法的自适应控制方法、存储介质、设备及系统
US11128561B1 (en) Auto load balancing
US20120191998A1 (en) Dynamic power management in a communications device
US11671368B2 (en) Congestion control method, apparatus, and system, and computer storage medium
CN114500218B (zh) 一种控制网络设备的方法及装置
CN102783092A (zh) 通信系统，网络管理方法和交换设备
US11645166B2 (en) Method, device and computer program product for backuping data
US20160191412A1 (en) Technologies for directed power and performance management
US20140047260A1 (en) Network management system, network management computer and network management method
CN111211988B (zh) 面向分布式机器学习的数据传输方法及系统
CN112565102A (zh) 一种负载均衡方法、装置、设备及介质
CN110290020B (zh) 一种以太网测试仪表高精度流量百分比产生方法和装置
US20220200870A1 (en) Inter-terminal connection state prediction method and apparatus and analysis device
Meng et al. Greening the internet using multi-frequency scaling scheme
De Pellegrini et al. Blind, adaptive and robust flow segmentation in datacenters
US20240022501A1 (en) Data Packet Sending Method and Device
WO2023155904A1 (zh) 一种网络设备运行状态的调整方法、装置及相关设备
Tong et al. A novel and efficient link discovery mechanism in SDN
CN113678404A (zh) 用于在对边缘网关节点的维护期间减轻中断的方法
Bolla et al. Energy-aware load balancing for parallel packet processing engines

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23755904

Country of ref document: EP

Kind code of ref document: A1