Energy consumption adjusting method and device of server system, equipment, storage medium and program product

Abstract

The invention relates to an energy consumption adjusting method and device of a server system, equipment, a storage medium and a program product, and the method comprises the steps: obtaining the historical prediction state information of the server system and the actual state information of the current moment, and then carrying out the calculation of the historical prediction state information and the actual state information of the current moment; and inputting the state information into the state prediction model for prediction to obtain predicted state information at the current moment, and finally adjusting the energy consumption of the server system according to the predicted state information at the current moment and historical predicted state information to obtain an adjustment result. According to the method, the future state of the server system can be predicted more accurately by combining the historical prediction state information and the actual state information at the current moment, so that the energy consumption of the server system is accurately adjusted according to the future state, resource waste caused by inaccurate prediction is avoided, and the energy consumption of the server system is reduced. Energy and power cost can be saved to the greatest extent, and the call of green energy conservation and emission reduction is responded.

Description

Technical Field

[0001] The present application relates to the technical field of cloud computing servers, and in particular to a method, device, equipment, storage medium and program product for regulating energy consumption of a server system. Background Art

[0002] Cloud computing is a new trend in the field of information technology (IT). It not only involves computing, software, data access and storage services, but also occupies a core position in Infrastructure as a Service (IaaS). In the field of IaaS, energy saving is one of the most critical issues. Energy saving has become an important topic in various fields because it is not only related to economic benefits, but also closely linked to environmental protection and social sustainable development. In the information and communication technology (ICT) industry, especially in the field of cloud computing, the increase in energy consumption has become a significant problem. With the increasing distribution of cloud computing and the continuous growth of data storage and computing needs, the energy consumption of large data centers and infrastructure is also rising sharply. Under the normal operation of distributed cloud computing centers, it is crucial to effectively save energy costs and achieve green and low-carbon emission reduction in the field of cloud computing.

[0003] At present, the energy consumption adjustment method of cloud computing servers in the related art has the problem of inaccurate adjustment. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, equipment, storage medium and program product for adjusting the energy consumption of a server system that can adjust the accuracy in order to solve the above technical problems.

[0005] In a first aspect, the present application provides a method for adjusting energy consumption of a server system, the method comprising:

[0006] Obtain the historical predicted status information and the actual status information of the server system at the current moment;

[0007] Input the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction, and obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model;

[0008] According to the current prediction status information and the historical prediction status information, the energy consumption of the server system is adjusted to obtain an adjustment result.

[0009] In one embodiment, the historical predicted state information and the actual state information at the current moment are input into the state prediction model for prediction, and the predicted state information at the current moment is obtained, including:

[0010] Input the historical prediction state information into the theoretical state prediction model for prediction, and obtain the theoretical state information at the current moment;

[0011] The actual error noise information in the theoretical state information at the current moment and the actual state information at the current moment is input into the actual state prediction model for prediction to obtain the actual error covariance information at the current moment;

[0012] The theoretical state information at the current moment, the actual error covariance information at the current moment, and the actual state information at the current moment are input into the state optimization model for optimization to obtain the predicted state information at the current moment.

[0013] In one embodiment, the theoretical state information at the current moment includes the theoretical load information at the current moment and the theoretical error covariance information at the current moment, the theoretical state prediction model includes a first prediction sub-model and a second prediction sub-model, and the historical prediction state information is input into the theoretical state prediction model for prediction to obtain the theoretical state information at the current moment, including:

[0014] Inputting the historical prediction theoretical load information in the historical prediction state information into the first prediction sub-model for prediction to obtain the theoretical load information at the current moment;

[0015] The prediction error covariance information of the previous moment and the preset noise error in the historical prediction state information are input into the second prediction sub-model for prediction to obtain the theoretical error covariance information of the current moment.

[0016] In one embodiment, the state optimization model includes a gain prediction sub-model, a first optimization sub-model and a second optimization sub-model, and the predicted state information at the current moment includes the predicted load information at the current moment and the predicted error covariance information at the current moment. The theoretical state information at the current moment, the actual error covariance information at the current moment and the actual state information at the current moment are input into the state optimization model for optimization to obtain the predicted state information at the current moment, including:

[0017] Inputting the theoretical error covariance information at the current moment in the theoretical state information at the current moment and the actual load information at the current moment into the gain prediction sub-model for prediction to obtain the target gain;

[0018] Input the target gain, the theoretical error covariance information at the current moment, and the actual state information at the current moment into the first optimization sub-model for prediction, and obtain the predicted load information at the current moment;

[0019] The target gain and the theoretical error covariance information at the current moment are input into the second optimization sub-model for prediction to obtain the prediction error covariance information at the current moment.

[0020] In one embodiment, the energy consumption of the server system is adjusted according to the current prediction state information and the historical prediction state information to obtain the adjustment result, including:

[0021] Obtain the total power of the cooling system in the server system at the last moment;

[0022] According to the current predicted state information, the historical predicted state information and the total power at the previous moment, the energy consumption of the server system is adjusted to obtain an adjustment result.

[0023] In one embodiment, the energy consumption of the server system is adjusted according to the current prediction state information, the historical prediction state information and the total power at the previous moment, and the adjustment result is obtained, including:

[0024] The current prediction status information, historical prediction status information and the total power at the previous moment are input into the energy consumption prediction model to obtain the target power;

[0025] Adjust the current power of the cooling system to the target power.

[0026] In a second aspect, the present application further provides an energy consumption adjustment device for a server system, the device comprising:

[0027] An acquisition module is used to acquire the historical predicted status information of the server system and the actual status information at the current moment;

[0028] The prediction module is used to input the historical prediction state information and the actual state information at the current moment into the state prediction model for prediction, and obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model;

[0029] The adjustment module is used to adjust the energy consumption of the server system according to the current prediction state information and the historical prediction state information to obtain the adjustment result.

[0030] In a third aspect, the present application further provides a computer device, the computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:

[0031] Obtain the historical predicted status information and the actual status information of the server system at the current moment;

[0032] Input the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction, and obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model;

[0033] According to the current prediction status information and the historical prediction status information, the energy consumption of the server system is adjusted to obtain an adjustment result.

[0034] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0035] Obtain the historical predicted status information and the actual status information of the server system at the current moment;

[0036] Input the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction, and obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model;

[0037] According to the current prediction status information and the historical prediction status information, the energy consumption of the server system is adjusted to obtain an adjustment result.

[0038] In a fifth aspect, the present application further provides a computer program product, the computer program product comprising a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0039] Obtain the historical predicted status information and the actual status information of the server system at the current moment;

[0040] Input the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction, and obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model;

[0041] According to the current prediction status information and the historical prediction status information, the energy consumption of the server system is adjusted to obtain an adjustment result.

[0042] The energy consumption regulation method, device, equipment, storage medium and program product of the above-mentioned server system obtains the historical predicted state information and the actual state information of the server system at the current moment, and then inputs the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction to obtain the predicted state information at the current moment, and finally adjusts the energy consumption of the server system according to the predicted state information at the current moment and the historical predicted state information to obtain the regulation result. Among them, the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model. The above-mentioned method can more accurately predict the future state of the server system by combining the historical predicted state information and the actual state information at the current moment, so as to accurately adjust the energy consumption of the server system according to the future state, thereby avoiding the waste of resources due to inaccurate prediction, and can save energy and electricity costs to the maximum extent and respond to the call for green energy conservation and emission reduction. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is an internal structure diagram of a computer device in one embodiment;

[0044] Figure 2 A schematic diagram of a flow chart of a method for adjusting energy consumption of a server system in one embodiment;

[0045] Figure 3 is a flow chart of a method for adjusting energy consumption of a server system in another embodiment;

[0046] Figure 4 is a flow chart of a method for adjusting energy consumption of a server system in another embodiment;

[0047] Figure 5 is a flow chart of a method for adjusting energy consumption of a server system in another embodiment;

[0048] Figure 6 is a flow chart of a method for adjusting energy consumption of a server system in another embodiment;

[0049] Figure 7 is a flow chart of a method for adjusting energy consumption of a server system in another embodiment;

[0050] Figure 8 1 is a structural block diagram of an energy consumption regulating device of a server system in an embodiment. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0052] Cloud computing is a new trend in the field of information technology (IT). It not only involves computing, software, data access and storage services, but also occupies a core position in infrastructure as a service (IaaS). In the field of IaaS, energy saving is one of the most critical issues. Energy saving has become an important topic in various fields because it is not only related to economic benefits, but also closely linked to environmental protection and social sustainable development. In the information and communication technology (ICT) industry, especially in the field of cloud computing, the increase in energy consumption has become a significant problem. With the increasing distribution of cloud computing and the continuous growth of data storage and computing needs, the energy consumption of large data centers and infrastructure is also rising sharply. Under the normal operation of distributed cloud computing centers, it is crucial to effectively save energy costs and achieve green and low-carbon emission reduction in the field of cloud computing. At present, the energy consumption adjustment method of cloud computing servers in related technologies has the problem of inaccurate adjustment.

[0053] The present application provides an energy consumption regulation method for a server system, aiming to solve the above-mentioned technical problems. The following embodiments will specifically illustrate the energy consumption regulation method for a server system described in the present application.

[0054] The energy consumption adjustment method of the server system provided in the embodiment of the present application can be applied to Figure 1 In the computer device shown in the figure, the computer device can be a terminal or a server, and its internal structure diagram can be as shown in FIG. Figure 1As shown, the computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for regulating energy consumption of a server system is realized. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0055] Those skilled in the art will understand that Figure 1 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0056] In one embodiment, Figure 2 As shown, a method for adjusting energy consumption of a server system is provided, and the method is applied to Figure 1 The computer device in the example is used to illustrate, including the following steps:

[0057] S201, obtaining historical predicted status information and current actual status information of the server system.

[0058] The server system is a distributed cloud computing center server, which can be a single server or a server cluster composed of multiple servers. The historical prediction state information includes historical prediction load information and historical prediction error covariance information. Optionally, the historical prediction state information includes the prediction state information at the last moment and the prediction state information at the last moment. The actual state information at the current moment includes the actual load information at the current moment and the actual error noise information at the current moment.

[0059] In an embodiment of the present application, the computer can store the historical predicted state information of the server system predicted by the state prediction model in a database or a preset storage path. When the computer device needs to control the energy consumption of the server system at the current moment, the historical predicted state information of the server system can be obtained from the database or the preset storage path. The computer device can monitor the actual state of the server system in real time to obtain the actual state information at the current moment. Specifically, the preset state query function can be called periodically to monitor the actual state of the server system in real time. Optionally, a state information acquisition module can also be set on the server system, and the server system can be monitored in real time through the information acquisition module, and the results of the real-time monitoring are sent to the computer device, and the computer device can obtain the actual state information of the server system at the current moment. Optionally, the computer device can actively obtain the actual state information of the server system at the current moment from the information acquisition module at a regular time.

[0060] S202, inputting the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction, and obtaining the predicted state information at the current moment.

[0061] Among them, the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model. The theoretical state prediction model is used to predict the theoretical state of the server system at the current moment, the actual state prediction model is used to predict the actual state of the server system at the current moment, and the state optimization model is used to predict the current state of the server system based on the theoretical state information at the current moment and the actual state at the current moment. The predicted state information at the current moment includes the predicted load information at the current moment and the predicted error covariance information at the current moment.

[0062] In an embodiment of the present application, the computer device can train an initial state prediction model in advance according to the state sample data of the server system to obtain a state prediction model. Optionally, the computer device can obtain an existing prediction model and use it as the state prediction model in this embodiment. Optionally, the computer device can build a state prediction model in advance according to a preset algorithm. After the computer device obtains the historical prediction state information and the actual state information at the current moment based on the above steps, the historical prediction state information and the actual state information at the current moment can be input into the state prediction model, and the theoretical state prediction model in the state prediction model is used to predict the theoretical state of the server system at the current moment, and the actual state of the server system at the current moment is predicted by the actual state prediction model in the state prediction model, and the state optimization model in the state prediction model is used to predict the current state of the server system according to the theoretical state information at the current moment and the actual state at the current moment, and the predicted state information at the current moment is obtained. Optionally, after the computer device obtains the historical prediction status information and the actual status information at the current moment, it can perform preprocessing such as data cleaning, interpolation, and data enhancement on the historical prediction status information and the actual status information at the current moment to obtain the preprocessed historical prediction status information and the preprocessed actual status information at the current moment, and then input the preprocessed historical prediction status information and the preprocessed actual status information at the current moment into the status prediction model for prediction to obtain the prediction status information at the current moment.

[0063] S203, adjusting the energy consumption of the server system according to the current prediction state information and the historical prediction state information to obtain an adjustment result.

[0064] The adjustment result includes the target power.

[0065] In an embodiment of the present application, after the computer device obtains the predicted state information at the current moment based on the above steps, it can determine the target energy consumption value according to the predicted state information at the current moment and the historical predicted state information, and then adjust the energy consumption of the server system according to the target energy consumption value, so that the energy consumption of the current server system reaches the target energy consumption value, and obtain the adjusted server system. Optionally, the computer device can determine the target power adjustment amount of the server system according to the predicted state information at the current moment and the historical predicted state information, and then adjust the power of the server system according to the target power adjustment amount value, and obtain the adjusted server system. Optionally, the computer device can input the predicted state information at the current moment and the historical predicted state information into a preset adjustment model to perform energy consumption prediction, obtain the minimum energy consumption that satisfies the normal operation of the server system, and then determine the target power adjustment amount of the server system according to the minimum energy consumption, and then adjust the power of the server system according to the target power adjustment amount value, and obtain the adjusted server system.

[0066] The energy consumption regulation method of the server system provided in the embodiment of the present application obtains the historical predicted state information of the server system and the actual state information at the current moment, and then inputs the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction, obtains the predicted state information at the current moment, and finally adjusts the energy consumption of the server system according to the predicted state information at the current moment and the historical predicted state information to obtain the regulation result. Among them, the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model. The above method can more accurately predict the future state of the server system by combining the historical predicted state information and the actual state information at the current moment, so as to accurately adjust the energy consumption of the server system according to the future state, thereby avoiding the waste of resources due to inaccurate prediction, and can save energy and electricity costs to the maximum extent and respond to the call for green energy conservation and emission reduction.

[0067] In one embodiment, a specific implementation method for predicting the prediction status information at the current moment is also provided, such as Figure 3 As shown, the above step S202 of "inputting the historical predicted state information and the actual state information at the current moment into the state prediction model for prediction to obtain the predicted state information at the current moment" includes:

[0068] S301, inputting historical prediction state information into a theoretical state prediction model for prediction to obtain theoretical state information at the current moment.

[0069] The theoretical state prediction model may be a neural network model, a machine algorithm model, or a mathematical model. The theoretical state information at the current moment includes the theoretical load information at the current moment and the theoretical error covariance information at the current moment.

[0070] In an embodiment of the present application, the computer device may train an initial theoretical state prediction model in advance based on the state sample data of the server system to obtain a theoretical state prediction model. Optionally, the computer device may obtain an existing prediction model and use it as the theoretical state prediction model in this embodiment. Optionally, the computer device may construct a theoretical state prediction model in advance based on a preset mathematical algorithm. After obtaining the historical prediction state information, the computer device may input the historical prediction state information into the theoretical state prediction model, and predict the theoretical state of the server system at the current moment through the theoretical state prediction model to obtain the theoretical state information at the current moment.

[0071] Optional, such as Figure 4 As shown, the above step S301 of "inputting the historical prediction state information into the theoretical state prediction model for prediction to obtain the theoretical state information at the current moment" includes:

[0072] S3011, inputting the historical predicted theoretical load information in the historical predicted state information into the first prediction sub-model for prediction to obtain the theoretical load information at the current moment.

[0073] The theoretical state prediction model includes a first prediction sub-model and a second prediction sub-model. The first prediction sub-model is a theoretical load information prediction model.

[0074] In the embodiment of the present application, the load of the server system is closely related to the CPU usage rate, memory usage rate and disk IO, among which the main loads are still characterized by the CPU usage rate (load) and the memory usage rate (load). Therefore, this algorithm will use the CPU load and the memory load to jointly characterize the load of the server system. Usually, due to the uncertain changes in business loads and process-specific scheduled tasks, the CPU usage rate will experience instantaneous surges and sudden decreases. Obviously, it is not reasonable to use a rough CPU instantaneous change rate to predict the CPU load. If the CPU instantaneous change rate is analogized to power, it is obvious that the accumulation of CPU load over a period of time, that is, analogy to the concept of work, is more in line with it. The memory load changes are usually slow and insignificant, and there will not often be sudden increases and decreases. Therefore, it is relatively reasonable to use the instantaneous change rate of the memory.

[0075] The computer device may pre-build the first prediction sub-model, and the specific construction method is as follows:

[0076] Assume that the CPU is The load at the moment is , the rate of change is ; Memory exists The load at the moment is , the rate of change is ;in and Can be used and To approximate, we can use the last moment ( The forecast load information at the previous time (t 2 moment) to calculate the server system's load at the current moment ( Theoretical load information at the time :

[0077]

[0078]

[0079] The CPU and memory load prediction states of the above two formulas can be written in the form of a matrix, namely:

[0080]

[0081]

[0082] in, Indicates the theoretical load information at the current moment. Indicates the theoretical CPU load information at the current moment. Indicates the theoretical memory load information at the current moment. (or ) indicates the theoretical CPU load information at the last moment, (or ) indicates the theoretical memory load information at the last moment, Indicates the theoretical CPU load information at the last moment. Indicates the theoretical memory load information at the last moment.

[0083] After the computer device obtains the historical prediction state information, the historical prediction theoretical load information in the historical prediction state information can be input into the first prediction sub-model for prediction to obtain the theoretical load information at the current moment.

[0084] S3012, input the prediction error covariance information of the previous moment and the preset noise error in the historical prediction state information into the second prediction sub-model for prediction, and obtain the theoretical error covariance information of the current moment.

[0085] Among them, the second prediction sub-model is a theoretical error covariance information prediction model.

[0086] In the embodiment of the present application, the computer device may pre-build the second prediction sub-model, and the specific construction method is as follows:

[0087] because and It is used and The approximate calculation inevitably introduces errors, and the actual CPU load change is not uniform, so the current moment ( The actual load status information at the time It is also necessary to introduce a preset noise error ,in ,Right now The mean is 0 and the covariance is , then the actual load status information at the current moment for:

[0088]

[0089] The theoretical load information of the server system at the current moment can be calculated by the above formula. and the actual load status information at the current moment There are differences, and the difference between the two can be expressed by the theoretical error covariance information (i.e., the second prediction sub-model):

[0090]

[0091] After the computer device obtains the historical prediction state information, the prediction error covariance information of the previous moment and the preset noise error in the historical prediction state information can be input into the second prediction sub-model for prediction to obtain the theoretical error covariance information of the current moment.

[0092] S302, inputting the actual error noise information in the theoretical state information at the current moment and the actual state information at the current moment into the actual state prediction model for prediction, and obtaining the actual error covariance information at the current moment.

[0093] Among them, the actual state prediction model is the actual error covariance information prediction model.

[0094] In the embodiment of the present application, the computer device can pre-build an actual state prediction model, and the specific construction method is as follows:

[0095] Computer devices can use sensors to measure variables such as time slices, temperature, etc. by measuring the conversion matrix (For example, we see CPU and memory load values, but the sensor measures temperature or time slices. There is a linear conversion relationship between them. Here we use the conversion matrix Characterization), the actual load status information at the current moment Convert to sensor measurement value At the same time, due to the existence of systematic errors in sensor measurement, it is assumed that the sensor error noise (i.e., the actual error noise information) is , then the actual load status information at the current moment Sensor measurement value It can be expressed as:

[0096]

[0097] The actual load status information at the current moment Sensor measurement values ​​and theoretical load information at the current moment The error covariance between (i.e., the actual state prediction model) can be expressed as:

[0098] After the computer device obtains the theoretical state information at the current moment and the actual state information at the current moment, the actual error noise information in the theoretical state information at the current moment and the actual state information at the current moment can be input into the actual state prediction model for prediction to obtain the actual error covariance information at the current moment.

[0099] S303, inputting the theoretical state information at the current moment, the actual error covariance information at the current moment and the actual state information at the current moment into the state optimization model for optimization to obtain the predicted state information at the current moment.

[0100] The state optimization model may be a neural network model, a machine algorithm model, or a mathematical model. The state optimization model includes a gain prediction sub-model, a first optimization sub-model, and a second optimization sub-model. The predicted state information at the current moment includes the predicted load information at the current moment and the predicted error covariance information at the current moment.

[0101] In the embodiment of the present application, the computer device can train the initial state optimization model in advance according to the state sample data of the server system to obtain the state optimization model. Optionally, the computer device can obtain an existing prediction model and use it as the state optimization model in the present embodiment. Optionally, the computer device can construct a state optimization model in advance according to a preset mathematical algorithm. After obtaining the theoretical state information at the current moment, the actual error covariance information at the current moment, and the actual state information at the current moment, the computer device can input the theoretical state information at the current moment, the actual error covariance information at the current moment, and the actual state information at the current moment into the state optimization model for optimization to obtain the predicted state information at the current moment.

[0102] Optional, such as Figure 5 As shown, the above step S303 of "inputting the theoretical state information at the current moment, the actual error covariance information at the current moment and the actual state information at the current moment into the state optimization model for optimization to obtain the predicted state information at the current moment" includes:

[0103] S3031, inputting the theoretical error covariance information at the current moment in the theoretical state information at the current moment and the actual load information at the current moment into the gain prediction sub-model for prediction to obtain the target gain.

[0104] Among them, the target gain is the optimal Kalman gain.

[0105] In the embodiment of the present application, the computer device may pre-build a gain prediction sub-model, and the specific construction method is as follows:

[0106] Forecast load information at the current moment When the optimal value is reached, it should be consistent with the actual load status information at the current moment The difference is the smallest, I is the unit matrix, and the expected mean can be expressed as:

[0107]

[0108] According to the knowledge of the posterior estimation, the above formula is equivalent to minimizing the posterior estimation covariance matrix .right beg The partial derivative of and equal to 0 can be obtained:

[0109]

[0110] Then we can get the optimal Kalman gain (target gain):

[0111]

[0112] After obtaining the theoretical error covariance information and the actual load information at the current moment, the computer device can input the theoretical error covariance information and the actual load information at the current moment into the gain prediction sub-model for prediction to obtain the target gain.

[0113] S3032, input the target gain, the theoretical error covariance information at the current moment, and the actual state information at the current moment into the first optimization sub-model for prediction, to obtain the predicted load information at the current moment.

[0114] Among them, the first optimization sub-model is a load information prediction model.

[0115] In the embodiment of the present application, the computer device can pre-build a load prediction information prediction model, and the specific construction method is as follows:

[0116] According to Kalman filtering, combined with the theoretical state information at the current moment and the actual load status information at the current moment Sensor measurement value The forecast load information at the current moment can be determined for:

[0117] Right now, for and The weighted is the Kalman gain, i.e. the weighting coefficient, when For the optimal gain, is the optimal predicted load information at the current moment. The above formula can be understood more intuitively as follows: first, , which is inaccurate, and There are also errors. and The errors are weighted to be closer to the actual load status information.

[0118] After the computer device obtains the target gain based on the above steps, it can input the target gain, the theoretical error covariance information at the current moment, and the actual state information at the current moment into the first optimization sub-model for prediction to obtain the predicted load information at the current moment.

[0119] S3033, input the target gain and the theoretical error covariance information at the current moment into the second optimization sub-model for prediction, and obtain the prediction error covariance information at the current moment.

[0120] Among them, the second optimization sub-model is the prediction error covariance information prediction model.

[0121] In the embodiment of the present application, the computer device can pre-build a prediction error covariance information prediction model, and the specific construction method is as follows:

[0122] First, the actual load status information and the current load forecast information The covariance error can be expressed as:

[0123]

[0124] According to the target gain, it can be determined that the optimal covariance error (i.e., the second optimization sub-model) can be expressed as:

[0125]

[0126] After the computer device obtains the target gain and the theoretical error covariance information at the current moment based on the above steps, the target gain and the theoretical error covariance information at the current moment can be input into the second optimization sub-model for prediction to obtain the predicted error covariance information at the current moment.

[0127] The method described in the embodiment of the present application predicts the system load at the next moment based on the CPU load and memory load in combination with the Kalman filter algorithm, and adaptively adjusts the power of the cooling and heat dissipation system of the hyper-converged all-in-one machine through load estimation, thereby saving energy and electricity costs to the maximum extent and responding to the call for green energy conservation and emission reduction while ensuring the normal operation of the hyper-converged all-in-one cloud computing server.

[0128] In one embodiment, a specific implementation method for adjusting the energy consumption of the server system is also provided, such as Figure 6As shown, the above step S203 of "adjusting the energy consumption of the server system according to the current prediction state information and the historical prediction state information to obtain the adjustment result" includes:

[0129] S401, obtaining the total power of a cooling system in a server system at the last moment.

[0130] In the embodiment of the present application, a corresponding sensor may be pre-installed on the cooling and heat dissipation system in the server system, and then the computer device may obtain the total power of the cooling and heat dissipation system in the server system at the last moment through the sensor.

[0131] S402, adjusting the energy consumption of the server system according to the current prediction state information, the historical prediction state information and the total power at the previous moment to obtain an adjustment result.

[0132] In an embodiment of the present application, after obtaining the predicted status information at the current moment, the historical predicted status information and the total power at the previous moment, the computer can determine the adjustment coefficient based on the predicted status information at the current moment and the historical predicted status information, and then determine the target adjustment loss based on the adjustment coefficient and the total power at the previous moment, and finally adjust the energy consumption of the server system according to the target adjustment loss to obtain an adjustment result.

[0133] Optional, such as Figure 7 As shown, the above step S602 of "adjusting the energy consumption of the server system according to the current prediction state information, the historical prediction state information and the total power at the previous moment to obtain the adjustment result" includes:

[0134] S4021, input the current prediction status information, historical prediction status information and the total power at the previous moment into the energy consumption prediction model to obtain the target power.

[0135] In the embodiment of the present application, the computer device can pre-build an energy consumption prediction model, and the specific construction method is as follows:

[0136] Based on the current forecast load information in the current forecast status information and the predicted load information of the previous moment in the predicted status information of the previous moment , to control the power of the cooling system. For example, if the current forecast load information Greater than the forecast load information of the previous moment , the power of the heat dissipation and cooling system can be relatively increased to ensure the normal operation of the cloud computing server; if the current moment's predicted load information Less than the forecast load information of the previous moment , the power of the heat dissipation and cooling system can be relatively reduced to reduce the consumption of power resources. The total power of the cooling system at the moment (i.e. the total power at the previous moment) is ,So The total power of the cooling system at this moment can be:

[0137]

[0138] The computer equipment obtains the current predicted state information, historical predicted state information and the total power of the previous moment Afterwards, the current forecast load information can be , the predicted load information of the last moment and the total power at the previous moment , input into the energy consumption prediction model to obtain the target power .

[0139] S4022, adjusting the current power of the cooling system to the target power.

[0140] In the embodiment of the present application, after the computer device obtains the target power based on the above steps, the current power of the cooling and heat dissipation system can be adjusted to the target power.

[0141] The method described in the embodiment of the present application controls the total power of the cooling and heat dissipation system by optimally predicting the server system load, which not only meets the requirements of stable operation of the server system, but also avoids additional energy consumption, and is more in line with the requirements of green and low-carbon emission reduction. The method described in the embodiment of the present application controls the total power of the cooling and heat dissipation system by optimally predicting the server system load, which not only meets the requirements of stable operation of the server system, but also avoids additional energy consumption, and is more in line with the requirements of green and low-carbon emission reduction. At the same time, the above method can calculate the optimal load state prediction estimate of the system at the current moment according to the Kalman filter algorithm based on the load of the hyper-converged all-in-one server resource pool, and adaptively adjust the power of the cooling and heat dissipation system of the cloud computing server at the current moment in combination with the optimal load state prediction estimate, while ensuring the normal operation of the hyper-converged all-in-one cloud computing server, and save energy and electricity costs to the maximum extent and respond to the call for green energy conservation and emission reduction.

[0142] Based on all the above embodiments, a method for adjusting energy consumption of a server system is also provided, the method comprising:

[0143] S501, obtaining historical predicted state information and current actual state information of the server system.

[0144] S502: Input the historical predicted theoretical load information in the historical predicted state information into the first prediction sub-model for prediction to obtain the theoretical load information at the current moment.

[0145] S503: Input the prediction error covariance information of the previous moment and the preset noise error in the historical prediction state information into the second prediction sub-model for prediction to obtain the theoretical error covariance information of the current moment.

[0146] S504, inputting the actual error noise information in the theoretical state information at the current moment and the actual state information at the current moment into the actual state prediction model for prediction, and obtaining the actual error covariance information at the current moment.

[0147] S505, inputting the theoretical error covariance information at the current moment in the theoretical state information at the current moment and the actual load information at the current moment into the gain prediction sub-model for prediction to obtain the target gain.

[0148] S506, input the target gain, the theoretical error covariance information at the current moment, and the actual state information at the current moment into the first optimization sub-model for prediction, to obtain the predicted load information at the current moment.

[0149] S507, inputting the target gain and the theoretical error covariance information at the current moment into the second optimization sub-model for prediction, and obtaining the prediction error covariance information at the current moment.

[0150] S508: Obtain the total power of the cooling system in the server system at the last moment.

[0151] S509: Input the current prediction state information, the historical prediction state information and the total power at the previous moment into the energy consumption prediction model to obtain the target power.

[0152] S510: Adjust the current power of the cooling system to the target power.

[0153] The methods described in the above steps are all described in the above embodiments. Please refer to the above description for details and will not be repeated here.

[0154] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0155] Based on the same inventive concept, the embodiment of the present application also provides an energy consumption regulating device for a server system for implementing the energy consumption regulating method for a server system involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of the energy consumption regulating device for one or more server systems provided below can refer to the limitations of the energy consumption regulating method for a server system above, and will not be repeated here.

[0156] In one embodiment, Figure 8 As shown, an energy consumption adjustment device for a server system is provided, comprising:

[0157] The acquisition module 11 is used to acquire the historical predicted state information of the server system and the actual state information at the current moment.

[0158] The prediction module 12 is used to input the historical prediction state information and the actual state information at the current moment into the state prediction model for prediction, and obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model.

[0159] The adjustment module 13 is used to adjust the energy consumption of the server system according to the current prediction state information and the historical prediction state information to obtain the adjustment result.

[0160] In one embodiment, the prediction module includes:

[0161] The first prediction unit is used to input the historical prediction state information into the theoretical state prediction model for prediction, so as to obtain the theoretical state information at the current moment.

[0162] The second prediction unit is used to input the actual error noise information in the theoretical state information at the current moment and the actual state information at the current moment into the actual state prediction model for prediction, so as to obtain the actual error covariance information at the current moment.

[0163] The third prediction unit is used to input the theoretical state information at the current moment, the actual error covariance information at the current moment and the actual state information at the current moment into the state optimization model for optimization to obtain the predicted state information at the current moment.

[0164] In one embodiment, the first prediction unit includes:

[0165] The first prediction sub-unit is used to input the historical prediction theoretical load information in the historical prediction state information into the first prediction sub-model for prediction, so as to obtain the theoretical load information at the current moment.

[0166] The second prediction sub-unit is used to input the prediction error covariance information of the previous moment and the preset noise error in the historical prediction state information into the second prediction sub-model for prediction to obtain the theoretical error covariance information of the current moment.

[0167] In one embodiment, the third prediction unit includes:

[0168] The third prediction subunit is used to input the theoretical error covariance information at the current moment in the theoretical state information at the current moment and the actual load information at the current moment into the gain prediction submodel for prediction to obtain the target gain.

[0169] The fourth prediction sub-unit is used to input the target gain, the theoretical error covariance information at the current moment, and the actual state information at the current moment into the first optimization sub-model for prediction, so as to obtain the predicted load information at the current moment.

[0170] The fifth prediction subunit is used to input the target gain and the theoretical error covariance information at the current moment into the second optimization submodel for prediction, so as to obtain the prediction error covariance information at the current moment.

[0171] In one embodiment, the adjustment module includes:

[0172] The acquisition unit is used to acquire the total power of the cooling system in the server system at the last moment.

[0173] The adjustment unit is used to adjust the energy consumption of the server system according to the current prediction state information, the historical prediction state information and the total power at the previous moment to obtain the adjustment result.

[0174] In one embodiment, the adjustment unit comprises:

[0175] The sixth prediction subunit is used to input the prediction state information at the current moment, the historical prediction state information and the total power at the previous moment into the energy consumption prediction model to obtain the target power.

[0176] The regulating subunit is used to regulate the current power of the cooling and heat dissipation system to the target power.

[0177] Each module in the energy consumption adjustment device of the server system can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0178] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of energy consumption regulation of the server system described in any of the above embodiments are implemented.

[0179] The above embodiment provides a computer device, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0180] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of adjusting the energy consumption of the server system described in any of the above embodiments are implemented.

[0181] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0182] In one embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps of energy consumption regulation of the server system described in any of the above embodiments.

[0183] The above embodiment provides a computer program product, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0184] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0185] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0186] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for adjusting energy consumption of a server system, characterized in that: The method comprises: Obtain the historical predicted status information and the actual status information of the server system at the current moment; Input the historical predicted state information and the actual state information at the current moment into a state prediction model for prediction, and obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model; According to the predicted state information at the current moment and the historical predicted state information, the energy consumption of the server system is adjusted to obtain an adjustment result.

2. The method according to claim 1, characterized in that The step of inputting the historical predicted state information and the actual state information at the current moment into a state prediction model for prediction to obtain the predicted state information at the current moment includes: Inputting the historical prediction state information into the theoretical state prediction model for prediction to obtain the theoretical state information at the current moment; Inputting the theoretical state information at the current moment and the actual error noise information in the actual state information at the current moment into the actual state prediction model for prediction to obtain the actual error covariance information at the current moment; The theoretical state information at the current moment, the actual error covariance information at the current moment, and the actual state information at the current moment are input into the state optimization model for optimization to obtain the predicted state information at the current moment.

3. The method according to claim 2, characterized in that The theoretical state information at the current moment includes the theoretical load information at the current moment and the theoretical error covariance information at the current moment, the theoretical state prediction model includes a first prediction sub-model and a second prediction sub-model, and the historical prediction state information is input into the theoretical state prediction model for prediction to obtain the theoretical state information at the current moment, including: Inputting the historical predicted theoretical load information in the historical predicted state information into the first prediction sub-model for prediction to obtain the theoretical load information at the current moment; The prediction error covariance information at the previous moment and the preset noise error in the historical prediction state information are input into the second prediction sub-model for prediction to obtain the theoretical error covariance information at the current moment.

4. The method according to claim 2, characterized in that: The state optimization model includes a gain prediction sub-model, a first optimization sub-model and a second optimization sub-model, the predicted state information at the current moment includes the predicted load information at the current moment and the predicted error covariance information at the current moment, the theoretical state information at the current moment, the actual error covariance information at the current moment and the actual state information at the current moment are input into the state optimization model for optimization to obtain the predicted state information at the current moment, including: Inputting the theoretical error covariance information at the current moment in the theoretical state information at the current moment and the actual load information at the current moment into the gain prediction sub-model for prediction to obtain a target gain; Inputting the target gain, the theoretical error covariance information at the current moment, and the actual state information at the current moment into the first optimization sub-model for prediction, to obtain the predicted load information at the current moment; The target gain and the theoretical error covariance information at the current moment are input into the second optimization sub-model for prediction to obtain the predicted error covariance information at the current moment.

5. The method according to any one of claims 1 to 4, characterized in that: The step of adjusting the energy consumption of the server system according to the current prediction state information and the historical prediction state information to obtain an adjustment result includes: Obtaining the total power of the cooling system in the server system at the last moment; According to the predicted state information at the current moment, the historical predicted state information and the total power at the previous moment, the energy consumption of the server system is adjusted to obtain an adjustment result.

6. The method according to claim 5, characterized in that The step of adjusting the energy consumption of the server system according to the predicted state information at the current moment, the historical predicted state information and the total power at the previous moment to obtain an adjustment result includes: Input the current prediction state information, the historical prediction state information and the total power at the previous moment into the energy consumption prediction model to obtain the target power; The current power of the cooling and heat dissipation system is adjusted to the target power.

7. An energy consumption regulating device for a server system, characterized in that: The device comprises: An acquisition module is used to acquire the historical predicted status information of the server system and the actual status information at the current moment; A prediction module, used to input the historical prediction state information and the actual state information at the current moment into a state prediction model for prediction, so as to obtain the predicted state information at the current moment; the state prediction model includes a theoretical state prediction model, an actual state prediction model and a state optimization model; The adjustment module is used to adjust the energy consumption of the server system according to the current prediction state information and the historical prediction state information to obtain an adjustment result.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images