Server power consumption balancing processing method, device and electronic equipment

Through AI models, the server power consumption is predicted and adjusted, and the problem of difficulty in accurately determining power consumption in the existing technology is solved, resource saving and server performance are achieved, and the problems of low energy efficiency utilization and high operational costs are solved.

CN118519784BActive Publication Date: 2024-11-15BEIJING TIANDI CHAOYUN TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410977493.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-15
Estimated Expiration
2044-07-22

AI Technical Summary

Technical Problem

Existing server technologies are difficult to accurately determine power consumption, resulting in waste of resources, low energy efficiency utilization, high operating costs, and difficult to achieve the rationality of resource allocation.

Method used

The AI ​​model is used to predict server power consumption, adjust power consumption parameters based on historical operating data, and monitor and optimize in real time to achieve power consumption equalization processing.

Benefits of technology

By predicting and adjusting power consumption, saving resources, reducing operating costs, improving server performance and stability, improving user experience, and achieving reasonable resource layout and energy conservation and emission reduction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118519784B_ABST
    Figure CN118519784B_ABST
Patent Text Reader

Abstract

The present application provides a method, device and electronic device for processing power consumption balance of a server. The method is applied to a baseboard management controller BMC; the BMC is configured with an AI model; the method includes: obtaining the operating parameters of the first server in the historical operating period; inputting the operating parameters of the first server into the AI ​​model, and outputting the predicted power consumption value of the server; using the predicted power consumption value as the current power consumption value, and performing the following power consumption adjustment monitoring steps: adjusting the server power consumption parameters according to the current power consumption value; real-time monitoring of the operating parameters of the second server after the power consumption parameters are adjusted; optimizing the predicted power consumption value output by the AI ​​model according to the operating parameters of the second server; using the optimized predicted power consumption value as the current power consumption value again, and continuing to perform the power consumption adjustment monitoring steps. The present application can achieve the effect of balancing the power consumption of the server, save resources and reduce operating costs, improve server performance and stability, and enhance user experience.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent technology, and in particular to a method, device and electronic device for balancing power consumption of a server. Background Art

[0002] At present, with the development of server technology, users have greatly increased their demand for intelligent server operation and maintenance. In the process of using the server, it is difficult to accurately determine the specific values ​​of some energy efficiency and power consumption in the server, and thus it is impossible to better realize efficient resource utilization to achieve energy conservation and emission reduction effects, which leads to a large waste of resources and increased manpower investment, resulting in a series of problems such as low energy efficiency utilization and unreasonable resource allocation in the server. Summary of the invention

[0003] The purpose of the present application is to provide a server power consumption balancing processing method, device and electronic device, which can predict the server power consumption based on the historical operation status, and after adjusting the server power consumption parameters based on the predicted power consumption, continuously adjust the predicted power consumption according to the real-time working data, so as to achieve the server power consumption balancing processing effect, save resources and reduce operating costs, improve server performance and stability, and enhance user experience.

[0004] In the first aspect, the present application provides a server power consumption balancing processing method, which is applied to a baseboard management controller BMC; the BMC is configured with an AI model; the method includes: obtaining the first server operating parameters of a historical operating period; the first server operating parameters include: server performance parameters and server power consumption parameters; inputting the first server operating parameters into the AI ​​model, and outputting a predicted power consumption value of the server; the predicted power consumption value includes: power supply input power; using the predicted power consumption value as the current power consumption value, and executing the following power consumption adjustment monitoring steps: adjusting the server power consumption parameters according to the current power consumption value; real-time monitoring of the second server operating parameters after the power consumption parameters are adjusted; optimizing the predicted power consumption values ​​output by the AI ​​model according to the second server operating parameters; using the optimized predicted power consumption value as the current power consumption value again, and continuing to execute the power consumption adjustment monitoring steps.

[0005] Furthermore, the above-mentioned historical operating time period is the time period of the previous week corresponding to the current moment; the step of inputting the first server working parameter into the AI ​​model and outputting the predicted power consumption value of the server includes: dividing the time period of the previous week into multiple sub-time periods on average; counting the average power consumption value in each sub-time period according to the first server working parameter; calculating the power consumption coefficient corresponding to each sub-time period based on the average power consumption value in each sub-time period; determining the predicted power consumption value of the server based on the average power consumption value in each sub-time period and the power consumption coefficient.

[0006] Furthermore, the above-mentioned step of calculating the power consumption coefficient corresponding to each sub-time period based on the average power consumption value in each sub-time period includes: taking each sub-time period as the current sub-time period, and respectively calculating the power consumption difference between the average power consumption value of the next adjacent time period and the average power consumption value of the current sub-time period and the total power consumption; dividing the power consumption difference by the total power consumption to obtain the power consumption coefficient corresponding to the current sub-time period.

[0007] Furthermore, the step of determining the predicted power consumption of the server based on the average power consumption value and the power consumption coefficient in each sub-time period includes: calculating the predicted power consumption of the server according to the following specified formula:

[0008] Y=a0+a1×b1+a2×b2+.....a n ×b n ;

[0009] Where Y represents the predicted power consumption value of the server; a0 represents the power consumption value per unit spacing corresponding to the predicted power consumption value, which is a constant; b1-b n Respectively represent the average power consumption values ​​corresponding to the 1st sub-time period to the nth sub-time period; a1-a n They respectively represent the power consumption coefficients corresponding to the 1st sub-time period to the nth sub-time period.

[0010] Furthermore, the above-mentioned step of optimizing the predicted power consumption output by the AI ​​model according to the operating parameters of the second server includes: calculating the real-time power consumption value of the server according to the operating parameters of the second server; determining whether the real-time power consumption value exceeds the preset maximum power consumption value; if so, turning off the prediction function of the AI ​​model; if not, calculating the difference between the real-time power consumption value and the current power consumption value; and optimizing the predicted power consumption output by the AI ​​model according to the size relationship between the difference and multiple preset values.

[0011] Furthermore, the above-mentioned step of optimizing the predicted power consumption value output by the AI ​​model according to the size relationship between the difference and multiple preset values ​​includes: if the difference is less than the first preset value, determining that the power consumption adjustment is successful, and controlling the server to operate according to the current power consumption value; if the difference is greater than or equal to the first preset value, and less than the second preset value, determining the real-time power consumption value as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the second preset value, and less than the third preset value, determining the average of the real-time power consumption value and the current power consumption value as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the third preset value, treating the real-time power consumption value as invalid data, and continuing to execute the step of real-time monitoring of the working parameters of the second server after the power consumption parameters are adjusted; wherein, the first preset value, the second preset value, and the third preset value increase in sequence.

[0012] Furthermore, before the above step of inputting the first server operating parameters into the AI ​​model, it also includes: preprocessing the first server operating parameters, and the preprocessing includes: cleaning invalid data, integrating data, and normalizing data.

[0013] In the second aspect, the present application also provides a server power consumption balancing processing device, which is applied to a baseboard management controller BMC; the BMC is configured with an AI model; the device includes: a data acquisition module, used to obtain the first server operating parameters of the historical operating period; the first server operating parameters include: server performance parameters and server power consumption parameters; a prediction module, used to input the first server operating parameters into the AI ​​model, and output the predicted power consumption value of the server; the predicted power consumption value includes: power supply input power; a cycle optimization module, used to use the predicted power consumption value as the current power consumption value, and perform the following power consumption adjustment monitoring steps: adjust the server power consumption parameters according to the current power consumption value; monitor the second server operating parameters after the power consumption parameters are adjusted in real time; optimize the predicted power consumption value output by the AI ​​model according to the second server operating parameters; use the optimized predicted power consumption value as the current power consumption value again, and continue to perform the power consumption adjustment monitoring steps.

[0014] In a third aspect, the present application further provides an electronic device, comprising a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method described in the first aspect above.

[0015] In a fourth aspect, the present application further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method described in the first aspect above.

[0016] In the server power consumption balancing processing method, device and electronic device provided by the present application, firstly, the first server operating parameters of the historical operating period are obtained; the first server operating parameters include: server performance parameters and server power consumption parameters; the first server operating parameters are input into the AI ​​model, and the predicted power consumption value of the server, i.e., the power input power, is output; the predicted power consumption value is then used as the current power consumption value, and the following power consumption adjustment monitoring steps are performed: according to the current power consumption value, the server power consumption parameters are adjusted; real-time monitoring of the second server operating parameters after the power consumption parameters are adjusted; according to the second server operating parameters, the predicted power consumption value output by the AI ​​model is optimized; the optimized predicted power consumption value is used as the current power consumption value again, and the power consumption adjustment monitoring step is continued. In this method, the server power consumption can be predicted based on the historical operating conditions, and after adjusting the server power consumption parameters based on the predicted power consumption, the predicted power consumption can be adjusted continuously according to the real-time working conditions, so as to achieve the server power consumption balancing processing effect, save resources and reduce operating costs, improve server performance and stability, and enhance user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flow chart of a method for balancing power consumption of a server provided in an embodiment of the present application;

[0019] Figure 2 A flowchart of a prediction process provided in an embodiment of the present application;

[0020] Figure 3 A structural block diagram of a server power consumption balancing processing device provided in an embodiment of the present application;

[0021] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] The technical solution of the present application will be described clearly and completely below in conjunction with the embodiments. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0023] The existing operation and maintenance methods for server equipment have the following technical problems:

[0024] How to achieve balanced design of BMC power consumption to improve energy efficiency, reduce resource waste, reduce operating costs, and reduce impact on the environment.

[0025] Based on this, the embodiment of the present application provides a server power consumption balancing processing method, device and electronic device, which can predict the server power consumption based on the historical operation status, and after adjusting the server power consumption parameters based on the predicted power consumption, continuously adjust the predicted power consumption according to the real-time working data, so as to achieve the server power consumption balancing processing effect, save resources and reduce operating costs, improve server performance and stability, and enhance user experience. To facilitate understanding of this embodiment, a server power consumption balancing processing method disclosed in the embodiment of the present application is first introduced in detail.

[0026] Figure 1 A flowchart of a method for balancing power consumption of a server provided in an embodiment of the present application, the method is applied to a BMC (Baseboard Management Controller); the BMC is configured with an AI model; the method specifically comprises the following steps:

[0027] Step S102, obtaining the operating parameters of the first server in the historical operating period;

[0028] Each server will be burned with the factory BMC firmware package by default when it leaves the factory; after the BMC login is successful, the server working parameter data will be collected. The server working parameters may include: server performance parameters and server power consumption parameters; server performance parameters include: fan speed, CPU usage, memory usage, disk IO, network bandwidth and other performance indicators; server power consumption parameters include: voltage, current and other power consumption data.

[0029] The above-mentioned historical operation time period can be a specified time period before the current moment, such as the previous week, the previous three days, etc. In the embodiment of the present application, the historical operation time period is set to the previous week.

[0030] When the server just starts running and there is no historical running data, the server power consumption parameters (such as voltage and current values) are controlled and adjusted with a preset power consumption value (for example, 100W) so that the server runs according to the preset power consumption value. When the running time reaches a certain time, for example, one week in this embodiment, the server working parameters within one week are sent to the AI ​​model so that the AI ​​model can predict the power consumption, as follows:

[0031] Step S104, inputting the operating parameters of the first server into the AI ​​model, and outputting the predicted power consumption value of the server;

[0032] The AI ​​model here is an intelligent model that can predict the power consumption required by the server in the future based on the server operating parameters in the historical operating period. In this embodiment, the AI ​​model can determine the average power consumption value of each sub-period in the historical operating period based on the server operating parameters in the historical operating period, and further, calculate based on the average power consumption value of each sub-period to obtain the predicted power consumption value. The predicted power consumption value is determined based on the change pattern of power consumption data in the historical operating period, and is relatively more in line with the power consumption requirements of the server in the future time period.

[0033] Step S106, taking the predicted power consumption value as the current power consumption value, and executing the following power consumption adjustment monitoring steps: adjusting the server power consumption parameters according to the current power consumption value; monitoring the second server operating parameters after the power consumption parameters are adjusted in real time; optimizing the predicted power consumption value output by the AI ​​model according to the second server operating parameters; taking the optimized predicted power consumption value as the current power consumption value again, and continuing to execute the power consumption adjustment monitoring steps.

[0034] Based on the above predicted power consumption value, adjust the server power consumption parameters, such as voltage and current, and then obtain the adjusted server working parameters in real time. Verify the above predicted power consumption value according to the adjusted server working parameters to determine whether the adjustment is effective. If not, optimize the predicted power consumption value, guide the adjustment of the server power consumption parameters through the optimized predicted power consumption value, and continue the monitoring feedback optimization process. If the adjustment is effective, no further optimization is required.

[0035] The server power consumption balancing processing method provided in the embodiment of the present application can predict the server power consumption based on the historical operating conditions, and after adjusting the server power consumption parameters based on the predicted power consumption, continuously adjust the predicted power consumption according to the real-time working conditions, thereby achieving the server power consumption balancing processing effect, saving resources and reducing operating costs, improving server performance and stability, and enhancing user experience.

[0036] The embodiment of the present application also provides another method for balancing the power consumption of a server, which is implemented on the basis of the above embodiment; this embodiment focuses on describing the power consumption prediction process and the cycle optimization process.

[0037] In one embodiment, the historical operating period is the period of the previous week corresponding to the current moment; the step of inputting the working parameters of the first server into the AI ​​model and outputting the predicted power consumption value of the server includes the following steps, see Figure 2 As shown:

[0038] Step S202, divide the time period of the previous week into multiple sub-time periods evenly; for example, divide the time period into units of 15 minutes to obtain n sub-time periods.

[0039] Step S204, counting the average power consumption value in each sub-time period according to the working parameters of the first server;

[0040] The first server operating parameters include: fan speed, CPU usage, hard disk IO, voltage, current, etc. The BMC data refresh frequency is once every 10 seconds. The instantaneous power is calculated based on the server operating parameters collected every 10 seconds; further, the average power is calculated based on the instantaneous power every 10 seconds within the 15-minute sub-time period, and the average power is used as the average power consumption value in the sub-time period. In this way, the average power consumption value corresponding to the server operating parameters in each sub-time period can be counted, such as b1, b2, b3...b n Indicates the average power consumption values ​​corresponding to n sub-time periods.

[0041] Step S206, calculating the power consumption coefficient corresponding to each sub-time period based on the average power consumption value in each sub-time period;

[0042] Specifically, take each sub-time period as the current sub-time period, calculate the power consumption difference between the average power consumption value of the next adjacent time period and the average power consumption value of the current sub-time period and the total power consumption; divide the power consumption difference by the total power consumption to obtain the power consumption coefficient corresponding to the current sub-time period.

[0043] The specific formula is as follows: n = (b n -b n-1 ) / (b n +b n-1 ); where b0=0 is preset;

[0044] For example, b1=100, b2=150, b3=100, b4=200,

[0045] Then: a1=(b1-b0) / (b1+b0) =1; a2=(b2-b1) / (b1+b2) = 0.2, a3= (b3-b2) / (b2+b3) =0.1....

[0046] Step S208: determining a predicted power consumption value of the server based on the average power consumption value and the power consumption coefficient in each sub-time period.

[0047] Calculate the predicted power consumption of the server according to the following given formula:

[0048] Y=a0+a1×b1+a2×b2+.....a n ×b n ;

[0049] Where Y represents the predicted power consumption value of the server; a0 represents the power consumption value per unit spacing corresponding to the predicted power consumption value, which is a constant; b1-b n Respectively represent the average power consumption values ​​corresponding to the 1st sub-time period to the nth sub-time period; a1-a n They respectively represent the power consumption coefficients corresponding to the 1st sub-time period to the nth sub-time period.

[0050] The above-mentioned step of optimizing the predicted power consumption output by the AI ​​model according to the working parameters of the second server includes: calculating the real-time power consumption value of the server according to the working parameters of the second server; determining whether the real-time power consumption value exceeds the preset maximum power consumption value; wherein, the preset maximum power consumption value = power supply rated power × load factor × 0.9.

[0051] If so, turn off the prediction function of the AI ​​model; this situation is extremely abnormal and requires an alarm. While turning off the prediction function of the AI ​​model, you can also send an alarm message to relevant personnel so that they can handle it in a timely manner.

[0052] If not, calculate the difference between the real-time power consumption value and the current power consumption value; based on the size relationship between the difference and multiple preset values, optimize the predicted power consumption output by the AI ​​model.

[0053] According to the size relationship between the difference value and multiple preset values, the steps of optimizing the predicted power consumption value output by the AI ​​model may include the following situations:

[0054] (1) If the difference is less than the first preset value, it is determined that the power consumption adjustment is successful, and the server is controlled to operate according to the current power consumption value;

[0055] (2) If the difference is greater than or equal to the first preset value and less than the second preset value, the real-time power consumption value is determined as the new predicted power consumption value output by the AI ​​model;

[0056] (3) If the difference is greater than or equal to the second preset value and less than the third preset value, the average of the real-time power consumption value and the current power consumption value is determined as the new predicted power consumption value output by the AI ​​model;

[0057] (4) If the difference is greater than or equal to the third preset value, the real-time power consumption value is regarded as invalid data, and the step of real-time monitoring of the operating parameters of the second server after the power consumption parameters are adjusted is continued; wherein the first preset value, the second preset value, and the third preset value are increased in sequence.

[0058] Here, a variety of different processing methods are provided according to the size of the difference, which can improve the accuracy of the predicted power consumption value. Through the above-mentioned simple prediction optimization methods, there is no need to analyze and predict based on a large amount of data in the previous week every time, which can improve the efficiency of power consumption prediction adjustment.

[0059] In a preferred embodiment, before the step of inputting the first server operating parameters into the AI ​​model, it also includes: preprocessing the first server operating parameters, and the preprocessing includes: cleaning invalid data, integrating data, and normalizing data.

[0060] In specific implementation, a series of server operating parameter data obtained will be collected and integrated into AI data; through cleaning, integration and standardization, it can be understood and used by subsequent AI models. The cleaning, integration and standardization methods are as follows:

[0061] In a feasible implementation, the instantaneous power consumption data is recorded every 10 seconds; the power consumption data of the previous week is screened, for example, the power consumption in the shutdown state tends to 0 as invalid data. Normalization processing, new_value = (value -min) / (max - min), where value is the original screened valid data, min and max are the minimum and maximum valid values ​​of the original data set respectively.

[0062] Use the processed data to train and integrate an AI model, such as a deep learning model or a reinforcement learning model. This model can predict power consumption values ​​based on historical operating period data, and can also optimize the predicted power consumption values ​​output by the model based on the current server operating parameters obtained in real time, and adjust the server power consumption settings accordingly.

[0063] The AI ​​model needs to continuously receive feedback data from the BMC to verify the effect of its power consumption adjustment. Based on the feedback data, the AI ​​model can further optimize its prediction and decision-making capabilities. When data is collected, processed, and AI model trained, failures will be recorded and entered into the model, allowing the AI ​​model to continue to grow and optimize.

[0064] For example, after adjusting the server power consumption parameters according to the power consumption values ​​predicted by the model, the adjusted server working parameter data will be continuously recorded and further converted into power consumption data. If the power consumption data is not much different from the power consumption value predicted by the aforementioned model, it is considered that the power consumption can meet the current usage, and this adjustment is considered to be in line with expectations.

[0065] If it is predicted that the load of a server will increase, the model can increase its power consumption in advance to improve performance. Conversely, if it is predicted that the load will decrease, the model can reduce power consumption to save energy. The specific way to adjust the server power consumption parameters is: when the server is not shut down, the power input power is adjusted, that is, the voltage and current are adjusted; when the server is shut down, the CPU performance, such as the CPU frequency, is adjusted.

[0066] A specific embodiment is listed below:

[0067] 1. BMC firmware preparation:

[0068] Before the server leaves the factory, the original BMC firmware package is burned into the BMC hardware to ensure that the BMC firmware supports data collection, remote access, and basic hardware monitoring functions.

[0069] 2. Initial setup:

[0070] After the server is deployed, authenticate and log in through the OpenBMC interface. Configure the BMC to allow remote access to ensure security and stability.

[0071] 3. Data acquisition:

[0072] After logging in to the BMC, start the data collection module to regularly collect server performance indicators, including CPU usage, memory usage, disk I / O, network bandwidth, etc. At the same time, collect power consumption and heat dissipation data, such as voltage, current, temperature, etc.

[0073] 4. Data preprocessing:

[0074] Perform preliminary screening of the collected data to remove invalid or erroneous data. Integrate the data to ensure its consistency and completeness. Perform standardization, such as normalization, so that the AI ​​model can better process and analyze the data.

[0075] 5. AI model selection and training:

[0076] Select an appropriate AI model, such as a deep learning model or a reinforcement learning model. Use the preprocessed data to train the AI ​​model. During the training process, ensure that the model can learn the relationship between the server operating parameters and power consumption.

[0077] 6. Model verification and feedback:

[0078] Implement AI models in BMC to monitor server performance in real time. Collect actual power consumption data after model prediction and compare it with the model prediction results. Provide feedback and adjust the AI ​​model based on the difference between actual data and predicted data.

[0079] 7. Dynamic adjustment of power consumption:

[0080] The AI ​​model dynamically adjusts the power consumption settings based on the current server performance indicators and the predicted load. If an increase in load is predicted, the model will increase power consumption in advance to avoid performance bottlenecks. For example, the power consumption curve rises straight up and changes greatly. The power consumption input will be increased in advance in the next few days. If a decrease in load is predicted, the model will reduce power consumption to save energy.

[0081] 8. Continuous optimization and iteration:

[0082] The AI ​​model continuously receives new data to continuously optimize its learning and prediction capabilities. Regularly retrain the AI ​​model to adapt to changes in server performance. Record all adjustment processes and results, including success and failure cases, for future analysis and improvement.

[0083] 9. Exception handling and safety measures:

[0084] When the AI ​​model adjusts power consumption, set a safety threshold to prevent over-adjustment from causing system instability. When an abnormal situation occurs, such as the AI ​​model not working properly, the BMC should have the ability to restore to the default settings, such as the BMC will restore the dynamic adjustment function of the AI ​​model until it is manually turned on next time. Ensure the security of the entire system and prevent unauthorized access and operation.

[0085] Through the above steps, the BMC power consumption balancing design method based on AI technology can be ensured. Through data collection, data processing, model training, power consumption adjustment, feedback and optimization, AI power consumption balancing design can be performed, which greatly increases the energy efficiency utilization of the server and realizes the reasonable layout of resources. With the improvement of autonomous learning of AI models, operating costs are reduced, the impact on the environment is reduced, the investment in manpower is reduced, the maintenance workload of the server is reduced, the intelligence level of the server is improved, and the performance and stability of the server are improved.

[0086] The server power consumption balancing processing method provided in the embodiment of the present application is actually a BMC power consumption balancing design method implemented based on AI technology, including an AI model prediction process, an AI power consumption adjustment process, and an AI feedback and optimization process; the AI ​​model predicts future power consumption based on data from historical operating periods, and adjusts power consumption according to the prediction results. When power consumption needs to be increased, server performance can be improved; when power consumption needs to be reduced, energy can be saved; through AI feedback and optimization, the model's learning, prediction, and decision-making capabilities can be further optimized to achieve a server power consumption balancing effect.

[0087] The above server power consumption balancing processing method can achieve the following effects:

[0088] 1. Improve the energy efficiency of servers and achieve reasonable resource layout.

[0089] 2. Reduce server operating costs and reduce resource waste.

[0090] 3. Reduce the impact on the environment and achieve energy conservation and emission reduction.

[0091] 4. Improve server performance and stability and enhance user experience.

[0092] 5. Reduce manpower input and realize intelligent operation and maintenance of servers.

[0093] Based on the above method embodiment, the embodiment of the present application also provides a server power consumption balancing processing device, which is applied to a baseboard management controller BMC; the BMC is configured with an AI model; see Figure 3 As shown, the device includes: a data acquisition module 32, used to obtain the first server operating parameters of the historical operating period; a prediction module 34, used to input the first server operating parameters into the AI ​​model and output the predicted power consumption value of the server; a loop optimization module 36, used to use the predicted power consumption value as the current power consumption value, and perform the following power consumption adjustment monitoring steps: adjust the server power consumption parameters according to the current power consumption value; monitor the second server operating parameters after the power consumption parameters are adjusted in real time; optimize the predicted power consumption value output by the AI ​​model according to the second server operating parameters; use the optimized predicted power consumption value again as the current power consumption value, and continue to perform the power consumption adjustment monitoring steps.

[0094] Furthermore, the above-mentioned historical operating time period is the time period of the previous week corresponding to the current moment; the prediction module 34 is used to evenly divide the time period of the previous week into multiple sub-time periods; according to the working parameters of the first server, the average power consumption value in each sub-time period is counted; based on the average power consumption value in each sub-time period, the power consumption coefficient corresponding to each sub-time period is calculated; based on the average power consumption value in each sub-time period and the power consumption coefficient, the predicted power consumption value of the server is determined.

[0095] Furthermore, the above-mentioned prediction module 34 is used to take each sub-time period as the current sub-time period, and calculate the power consumption difference between the average power consumption value of the next adjacent time period and the average power consumption value of the current sub-time period and the total power consumption; and divide the power consumption difference by the total power consumption to obtain the power consumption coefficient corresponding to the current sub-time period.

[0096] Furthermore, the prediction module 34 is used to calculate the predicted power consumption of the server according to the following specified formula:

[0097] Y=a0+a1×b1+a2×b2+.....a n ×b n ;

[0098] Where Y represents the predicted power consumption value of the server; a0 represents the power consumption value per unit spacing corresponding to the predicted power consumption value, which is a constant; b1-b n Respectively represent the average power consumption values ​​corresponding to the 1st sub-time period to the nth sub-time period; a1-a n They respectively represent the power consumption coefficients corresponding to the 1st sub-time period to the nth sub-time period.

[0099] Furthermore, the loop optimization module 36 is used to calculate the real-time power consumption value of the server according to the working parameters of the second server; determine whether the real-time power consumption value exceeds the preset maximum power consumption value; if so, turn off the prediction function of the AI ​​model; if not, calculate the difference between the real-time power consumption value and the current power consumption value; and optimize the predicted power consumption output by the AI ​​model according to the size relationship between the difference and multiple preset values.

[0100] Furthermore, the root loop optimization module 36 is used to determine that the power consumption adjustment is successful if the difference is less than a first preset value, and control the server to operate according to the current power consumption value; if the difference is greater than or equal to the first preset value and less than the second preset value, determine the real-time power consumption value as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the second preset value and less than the third preset value, determine the average of the real-time power consumption value and the current power consumption value as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the third preset value, treat the real-time power consumption value as invalid data, and continue to execute the step of real-time monitoring of the working parameters of the second server after the power consumption parameters are adjusted; wherein, the first preset value, the second preset value, and the third preset value increase in sequence.

[0101] Furthermore, the above-mentioned device also includes: a preprocessing module, which is used to preprocess the first server working parameters before the step of inputting the first server working parameters into the AI ​​model, and the preprocessing includes: cleaning invalid data, integrating data, and normalizing data.

[0102] The device provided in the embodiment of the present application has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the embodiment of the device, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0103] The present application also provides an electronic device, such as Figure 4 As shown, it is a schematic diagram of the structure of the electronic device, wherein the electronic device includes a processor 41 and a memory 40, the memory 40 stores computer executable instructions that can be executed by the processor 41, and the processor 41 executes the computer executable instructions to implement the above method.

[0104] exist Figure 4 In the illustrated embodiment, the electronic device further includes a bus 42 and a communication interface 43 , wherein the processor 41 , the communication interface 43 and the memory 40 are connected via the bus 42 .

[0105] Among them, the memory 40 may include a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 43 (which can be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 42 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 42 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0106] The processor 41 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 41. The above processor 41 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor 41 reads the information in the memory and completes the steps of the method of the above embodiment in combination with its hardware.

[0107] An embodiment of the present application also provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the above method. The specific implementation can be found in the aforementioned method embodiment, which will not be repeated here.

[0108] The computer program products of the methods, devices, and electronic devices provided in the embodiments of the present application include a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be found in the method embodiments, which will not be repeated here.

[0109] Unless otherwise specifically stated, the relative steps, numerical expressions and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present application.

[0110] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0111] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0112] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for balancing power consumption of a server, characterized in that: The method is applied to a baseboard management controller BMC; The BMC is configured with an AI model; the method comprises: Acquire a first server operating parameter during a historical operation period; the first server operating parameter includes: a server performance parameter and a server power consumption parameter; Inputting the first server operating parameter into the AI ​​model, and outputting the predicted power consumption value of the server; the predicted power consumption value includes: power supply input power; The predicted power consumption value is used as the current power consumption value, and the following power consumption adjustment monitoring steps are performed: According to the current power consumption value, adjust the power consumption parameters of the server; Real-time monitoring of the operating parameters of the second server after the power consumption parameters are adjusted; According to the working parameters of the second server, calculate the real-time power consumption value of the server; determine whether the real-time power consumption value exceeds the preset maximum power consumption value; wherein the preset maximum power consumption value = power supply rated power × load factor × 0.9; if yes, turn off the prediction function of the AI ​​model; if not, calculate the difference between the real-time power consumption value and the current power consumption value; if the difference is less than the first preset value, determine that the power consumption adjustment is successful, and control the server to run according to the current power consumption value; if the difference is greater than or equal to the first preset value and less than the second preset value, determine the real-time power consumption value as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the second preset value and less than the third preset value, determine the average of the real-time power consumption value and the current power consumption value as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the third preset value, treat the real-time power consumption value as invalid data, and continue to execute the step of real-time monitoring the working parameters of the second server after the power consumption parameters are adjusted; wherein the first preset value, the second preset value, and the third preset value increase in sequence; The optimized predicted power consumption value is used again as the current power consumption value, and the power consumption adjustment and monitoring step is continued.

2. The method according to claim 1, characterized in that The historical operation time period is a time period of the previous week corresponding to the current moment; the step of inputting the first server operating parameter into the AI ​​model and outputting the predicted power consumption value of the server includes: Divide the previous week's time period into multiple sub-time periods evenly; Counting the average power consumption value in each sub-time period according to the working parameters of the first server; Based on the average power consumption value in each sub-time period, calculate the power consumption coefficient corresponding to each sub-time period; Based on the average power consumption value and the power consumption coefficient in each sub-time period, a predicted power consumption value of the server is determined.

3. The method according to claim 2, characterized in that The step of calculating the power consumption coefficient corresponding to each sub-time period based on the average power consumption value in each sub-time period includes: Taking each sub-time period as the current sub-time period, respectively calculating the power consumption difference and the power consumption sum between the average power consumption value of the next adjacent time period and the average power consumption value of the current sub-time period; The power consumption coefficient corresponding to the current sub-time period is obtained by dividing the power consumption difference by the power consumption sum.

4. The method according to claim 2, characterized in that: The step of determining the predicted power consumption of the server based on the average power consumption value and the power consumption coefficient in each sub-time period includes: Calculate the predicted power consumption of the server according to the following given formula: Y=a0+a1×b1+a2×b2+.....a n ×b n ; Where Y represents the predicted power consumption value of the server; a0 represents the power consumption value per unit spacing corresponding to the predicted power consumption value, which is a constant; b1-b n Respectively represent the average power consumption values ​​corresponding to the 1st sub-time period to the nth sub-time period; a1-a n They respectively represent the power consumption coefficients corresponding to the 1st sub-time period to the nth sub-time period.

5. The method according to claim 1, characterized in that Before the step of inputting the first server working parameter into the AI ​​model, the method further includes: The working parameters of the first server are preprocessed, and the preprocessing includes: cleaning invalid data, integrating data, and normalizing data.

6. A server power consumption balancing processing device, characterized in that: The device is applied to a baseboard management controller BMC; The BMC is configured with an AI model; the device comprises: A data acquisition module, used to acquire the first server operating parameters in the historical operation period; the first server operating parameters include: server performance parameters and server power consumption parameters; A prediction module, used to input the working parameters of the first server into the AI ​​model and output a predicted power consumption value of the server; the predicted power consumption value includes: power supply input power; A loop optimization module is used to use the predicted power consumption value as the current power consumption value and perform the following power consumption adjustment monitoring steps: adjust the server power consumption parameters according to the current power consumption value; monitor the second server working parameters after the power consumption parameters are adjusted in real time; calculate the real-time power consumption value of the server according to the second server working parameters; determine whether the real-time power consumption value exceeds the preset maximum power consumption value; wherein the preset maximum power consumption value = power supply rated power × load factor × 0.9; if yes, turn off the prediction function of the AI ​​model; if not, calculate the difference between the real-time power consumption value and the current power consumption value; if the difference is less than the first preset value, determine that the power consumption adjustment is successful, and control the server to operate according to the current power consumption value; if the difference If the real-time power consumption value is greater than or equal to the first preset value and less than the second preset value, the real-time power consumption value is determined as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the second preset value and less than the third preset value, the average of the real-time power consumption value and the current power consumption value is determined as the new predicted power consumption value output by the AI ​​model; if the difference is greater than or equal to the third preset value, the real-time power consumption value is regarded as invalid data, and the step of real-time monitoring of the working parameters of the second server after the power consumption parameters are adjusted is continued; wherein, the first preset value, the second preset value, and the third preset value increase in sequence; the optimized predicted power consumption value is used again as the current power consumption value, and the power consumption adjustment monitoring step is continued.

7. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the method according to any one of claims 1 to 5.