Data center energy consumption equipment scheduling method and system

Abstract

The invention belongs to the field of data processing, and particularly relates to a data center energy consumption device scheduling method and system, and the method comprises the steps: obtaining the real-time operation data of each device in a data center; transmitting the real-time operation data to a corresponding virtual equipment model in a cloud server, and simulating a real-time operation state of the data center through the virtual equipment model in the cloud server to obtain virtual energy consumption of the data center; and performing optimization scheduling on each device based on the virtual energy consumption, and continuously transmitting real-time operation data of each device after optimization scheduling to a cloud server for prediction until the predicted virtual energy consumption of the data center accords with a preset energy consumption management condition of the data center. According to the method, the real-time operation data is transmitted to the virtual equipment model in the cloud server, and optimal scheduling is performed based on the virtual energy consumption, so that the accuracy of energy consumption prediction is improved, the equipment scheduling is optimized, and the energy efficiency of the data center and the scheduling efficiency are improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method and system for scheduling energy-consuming equipment in a data center. Background Art

[0002] A data center is a facility that provides centralized services for storing, managing, and processing data. It usually contains a large number of servers, storage devices, network equipment, and other support systems to ensure the security, reliability, and availability of data.

[0003] In the related art, the energy consumption of equipment in data centers is one of the foundations of equipment scheduling. Data centers often contain different types of equipment and operating systems. Since there are a large number of devices inside the data center, a large amount of real-time data will be generated, and the complex interaction between the equipment and the system makes the calculation of equipment energy consumption complicated, resulting in a cumbersome process of calculating the equipment energy consumption in the data center, low accuracy, and low efficiency of equipment scheduling.

[0004] Therefore, it is urgent to design a data center energy consumption equipment scheduling solution to solve at least one of the above technical problems. Summary of the invention

[0005] The present application provides a method and system for scheduling energy-consuming equipment in a data center, which is used to solve the technical problem that the energy consumption calculation process of equipment in a data center is complicated and has low accuracy.

[0006] In a first aspect, the present application provides a data center energy consumption equipment scheduling method, which is applied to a data center, wherein the data center is deployed with multiple devices, and the data center energy consumption equipment scheduling method includes:

[0007] Acquire real-time operation data of each device in the data center; the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data;

[0008] The real-time operation data is transmitted to a corresponding virtual device model in the cloud server, and the real-time operation state of the data center is simulated by the virtual device model in the cloud server to obtain the virtual energy consumption of the data center;

[0009] Based on the virtual energy consumption, each device is optimized and scheduled, and the real-time operation data of each device after the optimization and scheduling is continuously transmitted to the cloud server for prediction, until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center.

[0010] In a second aspect, an embodiment of the present application provides a data center energy consumption equipment scheduling system, the data center energy consumption equipment scheduling system is applied to a data center, the data center is deployed with multiple devices, and the data center energy consumption equipment scheduling system includes:

[0011] An acquisition unit is configured to acquire real-time operation data of each device in the data center; the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data;

[0012] A prediction unit is configured to transmit the real-time operation data to a corresponding virtual device model in the cloud server, and simulate the real-time operation state of the data center through the virtual device model in the cloud server to obtain the virtual energy consumption of the data center;

[0013] The scheduling unit is configured to optimize the scheduling of each device based on the virtual energy consumption, and continue to transmit the real-time operation data of each device after the optimized scheduling to the cloud server for prediction until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center.

[0014] In a third aspect, an embodiment of the present application provides an intelligent computing platform, the intelligent computing platform comprising:

[0015] at least one processor, memory, and input-output unit;

[0016] The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the data center energy-consuming equipment scheduling method of the first aspect.

[0017] In a fourth aspect, a computer-readable storage medium is provided, which includes instructions, and when the instructions are executed on a computer, the computer executes the data center energy-consuming equipment scheduling method of the first aspect.

[0018] The technical solution provided by the embodiment of the present application is that the data center energy consumption equipment scheduling system is applied to the data center, and the data center is deployed with multiple devices. In this solution, first, the real-time operation data of each device in the data center is obtained; the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data. Then, the real-time operation data is transmitted to the corresponding virtual device model in the cloud server, and the real-time operation state of the data center is simulated by the virtual device model in the cloud server to obtain the virtual energy consumption of the data center. Finally, each device is optimized and scheduled based on the virtual energy consumption, and the real-time operation data of each device after the optimized scheduling is continuously transmitted to the cloud server for prediction until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center. The technical solution of the present application significantly improves the accuracy of energy consumption prediction, optimizes equipment scheduling, improves the energy efficiency of the data center, and ensures service quality and reliability by transmitting real-time operation data to the virtual device model in the cloud server and optimizing scheduling based on virtual energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0020] Figure 1 It is a flow chart of a method for scheduling energy-consuming equipment in a data center according to an embodiment of the present application;

[0021] Figure 2 It is a structural schematic diagram of a data center energy consumption equipment scheduling system according to an embodiment of the present application;

[0022] Figure 3 It is a structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. 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 this field without creative work are within the scope of protection of this application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0025] The embodiments of the present application provide a method and system for scheduling energy-consuming equipment in a data center.

[0026] Specifically, in the data center energy consumption equipment scheduling scheme, the data center energy consumption equipment scheduling system can be applied to the data center, and the data center has multiple devices deployed. In this scheme, first, the real-time operation data of each device in the data center is obtained; the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data. Then, the real-time operation data is transmitted to the corresponding virtual device model in the cloud server, and the real-time operation status of the data center is simulated by the virtual device model in the cloud server to obtain the virtual energy consumption of the data center. Finally, based on the virtual energy consumption, each device is optimized and scheduled, and the real-time operation data of each device after the optimized scheduling is continuously transmitted to the cloud server for prediction until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center.

[0027] The technical solution of the present application can more accurately understand the current status and operation of the equipment by obtaining the real-time operation data (including equipment attribute information, real-time monitoring information, and equipment operation data) of each device in the data center. The collection and transmission of real-time data enables the energy consumption prediction model to be updated and adjusted based on the latest data, thereby improving the accuracy and reliability of the prediction. By simulating the real-time operation status of the data center through the virtual device model in the cloud server, the energy consumption characteristics of different devices under different load and environmental conditions can be more comprehensively evaluated. The virtual device model provides a fine simulation of the operation status of the data center, which can capture subtle changes in actual operation, thereby improving the accuracy of energy consumption prediction. Based on the virtual energy consumption, each device is optimized and scheduled, and the operation status of the device can be adjusted in real time to reduce the overall energy consumption. Through optimized scheduling, energy consumption can be minimized and operating costs can be reduced without affecting the quality of service. The real-time operation data of each device after optimized scheduling is continuously transmitted to the cloud server for prediction until the predicted virtual energy consumption meets the preset energy consumption management conditions to achieve continuous optimization. Continuous prediction and scheduling can dynamically adapt to changes in the operation status of the data center to ensure continuous energy consumption optimization effects. Through the application of real-time data collection and virtual device models, dynamic energy consumption management can be achieved, and the equipment operation strategy can be flexibly adjusted according to the actual operation conditions. Dynamic energy consumption management can significantly improve the energy efficiency of data centers and reduce unnecessary energy consumption. Optimized scheduling reduces energy consumption while ensuring the continuity and reliability of key services and equipment. Through reasonable scheduling, performance degradation and service interruption caused by equipment overload or excessive energy consumption can be avoided, and service quality and user experience can be improved. The technical solution of the present application significantly improves the accuracy of energy consumption prediction, optimizes equipment scheduling, improves the energy efficiency and scheduling efficiency of data centers, and ensures service quality and reliability by transmitting real-time operation data to the virtual device model in the cloud server and optimizing scheduling based on virtual energy consumption.

[0028] The data center energy consumption equipment scheduling scheme provided in the embodiment of the present application can also be executed by an electronic device, which can be a server, a server cluster, or a cloud server. The electronic device can also be a terminal device such as a mobile phone, a computer, a tablet computer, a wearable device, or a special device (such as a special terminal device with a data center energy consumption equipment scheduling system). These electronic devices can also be equipped with the chips introduced in the above embodiments. Alternatively, these electronic devices can also be installed with a service program for executing the data center energy consumption equipment scheduling scheme.

[0029] Figure 1 A schematic diagram of a data center energy consumption equipment scheduling method provided in an embodiment of the present application, such as Figure 1 , the method comprises the following steps:

[0030] 101. Acquire real-time operation data of each device in the data center.

[0031] Specifically, by deploying sensors and data acquisition systems in the data center, the operating data of each device is collected in real time and transmitted to the cloud server through the network for storage and management. Real-time operating data includes device attribute information, real-time monitoring information and device operation data, which provide the basis for subsequent energy consumption prediction and equipment scheduling, ensuring that the energy consumption management of the data center can be optimized based on the latest operating status. Specifically, various sensors and data acquisition devices are deployed in the data center to monitor the operating status of each device in real time. Sensor types include temperature sensors, humidity sensors, voltage sensors, current sensors, bandwidth monitoring devices, etc. Deploy data collectors (such as data acquisition gateways) to collect real-time data from various sensors and devices and perform preliminary processing. The data collector can also communicate with various devices and sensors through the network (such as Ethernet, InfiniBand, etc.) to obtain the operating data of the device in real time. The data collector caches the acquired real-time data in local storage for subsequent transmission and processing.

[0032] In the embodiment of the present application, the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data. Exemplarily, the device attribute information includes device ID (unique identifier), device type (such as server, storage device, network device, etc.), device model (for identifying specifications and performance), device location (such as rack number, cabinet number, etc.) and device status (such as running, idle, maintenance, etc.). Real-time monitoring information covers CPU usage, memory usage, disk I / O read and write rate, network inlet traffic, outlet traffic, bandwidth usage, computer room temperature, humidity, actual power consumption (watts) of the device, input voltage and input current. In addition, the device operation data includes power-on time, power-off time, number of restarts, maintenance records (including maintenance time and content), operation logs (recording daily operations and status changes) and alarm information (including abnormal status and fault information). These detailed real-time operation data provide a solid foundation for subsequent energy consumption prediction and equipment scheduling, ensuring that the energy consumption management of the data center can be optimized based on the latest operating status.

[0033] 102 , transmitting the real-time operation data to a corresponding virtual device model in the cloud server, and simulating the real-time operation status of the data center through the virtual device model in the cloud server to obtain the virtual energy consumption of the data center.

[0034] Further optionally, in 102, the real-time operation data is transmitted to the corresponding virtual device model in the cloud server, and the real-time operation status of the data center is simulated by the virtual device model in the cloud server to obtain the virtual energy consumption of the data center. The corresponding device type can also be determined based on the device attribute information; a virtual device prototype matching the device type is selected from the device library; and the device attribute information is matched to the corresponding setting items in the virtual device prototype to construct a corresponding virtual device model in the cloud server.

[0035] In an embodiment of the present application, the device library is dynamically configured according to the deployment information of the data center; the device library adaptively adjusts the device prototypes in the inventory as the functions of the data center are expanded or reduced.

[0036] First, determine the corresponding device type based on the device type in the device attribute information (such as server, storage device, network device, etc.). Select a virtual device prototype that matches the device type from the device library. The device library is dynamically configured based on the deployment information of the data center, and will adaptively adjust the device prototypes in the inventory as the data center expands or shrinks its functions.

[0037] The device library is dynamically configured according to the actual deployment of the data center and contains virtual device prototypes of various device types. According to the type of device, a matching virtual device prototype is selected from the device library. For example, if the device type is a server, a server-type virtual device prototype is selected.

[0038] Finally, the device attribute information (such as device ID, model, location, etc.) is matched to the corresponding setting items in the virtual device prototype. By matching the device attribute information to the setting items in the virtual device prototype, the corresponding virtual device model in the cloud server is constructed.

[0039] Take the deployment of a new server as an example, assuming that the device type is a server. Assume that the new server device ID is S12345. Assume that the device model is Xeon E5-2600, and the device deployment location is rack 1 and cabinet 2. Based on these assumptions, select a server virtual device prototype that matches the model Xeon E5-2600 from the device library. Then, match the device ID, model, location and other information to the virtual device prototype to construct the corresponding new server virtual device model in the cloud server.

[0040] Taking the expansion of network device functions as an example, assume that the device attribute information is the newly added network device ID, namely N67890, the device model is Cisco ASR9000, and the device deployment location is rack 3 and cabinet 5. Based on these assumptions, the device library is dynamically updated according to the deployment information of the data center, and the virtual device prototype of the Cisco ASR9000 router is added. Then, the device ID, model, location and other information are matched to the newly added virtual device prototype to construct the corresponding new router virtual device model in the cloud server.

[0041] By determining the device type based on device attribute information, selecting a matching virtual device prototype from the device library, and matching the device attribute information to the virtual device prototype, a virtual device model in the cloud server is constructed, which can significantly improve the accuracy of the model, the accuracy of energy consumption prediction, and the flexibility and reliability of the system. This method of constructing a virtual device model with dynamic configuration and adaptive adjustment provides a technical implementation basis for energy consumption management and equipment scheduling in data centers.

[0042] In an optional embodiment of 102, the real-time operating state of the virtual device model corresponding to each device is predicted through the operating state prediction layer. Furthermore, the sub-virtual energy consumption corresponding to each virtual device model is predicted based on the real-time operating state of each virtual device model through the device energy consumption prediction layer. Among them, the sub-virtual energy consumption corresponding to each device is predicted based on the energy consumption model matching each virtual device model. Finally, through the total energy consumption prediction layer, the sub-virtual energy consumption corresponding to each virtual device model is aggregated, and the aggregation result is optimized based on the actual layout and / or application scenario of the data center to obtain the virtual energy consumption of the data center.

[0043] Specifically, 102 predicts the virtual energy consumption of the data center through a hierarchical prediction method, which includes an operation status prediction layer, a device energy consumption prediction layer, and a total energy consumption prediction layer. The operation status prediction layer predicts the real-time operation status of the virtual device model of each device. This layer mainly predicts the status based on real-time monitoring information (such as CPU usage, memory usage, disk I / O, network traffic, etc.) and device operation data (such as boot time, number of restarts, operation logs, etc.). The operation status prediction layer can dynamically update the status of the virtual device model according to the latest real-time operation data to ensure the accuracy of the prediction.

[0044] In the device energy consumption prediction layer, based on the real-time operating status of each virtual device model output by the operating status prediction layer, the sub-virtual energy consumption corresponding to each virtual device model is predicted. The sub-virtual energy consumption corresponding to each device is predicted based on the energy consumption model matched by each virtual device model. The energy consumption model is usually constructed based on historical data and device characteristics, and can reflect the energy consumption characteristics of the device under different operating conditions.

[0045] In the total energy consumption prediction layer, the sub-virtual energy consumption corresponding to each virtual device model is aggregated. This process takes into account the actual layout and / or application scenarios of the data center, and optimizes the aggregation results to reflect the overall virtual energy consumption. Based on the actual layout and / or application scenarios of the data center, the aggregation results are optimized, for example, considering the geographical location of the computer room, the mutual influence between devices, etc., to obtain a more accurate total virtual energy consumption.

[0046] In this way, by dynamically updating the operating status of the virtual device model and optimizing the aggregation results based on the actual layout and application scenarios, the accuracy of the virtual energy consumption prediction is significantly improved. Prediction based on a refined energy consumption model can more accurately reflect the energy consumption characteristics of the equipment under different operating conditions, further improving the prediction accuracy. Through a hierarchical prediction method, it is possible to flexibly respond to changes in equipment in the data center, whether it is new equipment or equipment retirement. Through real-time monitoring and dynamic optimization, the system can detect potential problems in advance and make optimization adjustments, improving the reliability and stability of the data center. Accurate virtual energy consumption prediction provides a reliable basis for energy consumption management in the data center and helps managers develop more effective energy consumption optimization strategies. Based on the energy consumption prediction results, the scheduling strategy of equipment in the data center can be optimized to improve energy utilization efficiency and reduce operating costs.

[0047] Taking the energy consumption prediction of a server cluster as an example, the real-time operating status of each server in the server cluster is predicted through real-time monitoring information (such as CPU usage, memory usage) and device operation data (such as boot time, number of restarts). Based on the predicted real-time operating status, the sub-virtual energy consumption corresponding to each server is predicted using a matching energy consumption model. For example, when server A is under high load, its sub-virtual energy consumption is predicted to be 1000W. The sub-virtual energy consumption corresponding to each server (such as 1000W for server A, 800W for server B, etc.) is aggregated, and the aggregation result is optimized based on the actual layout of the server cluster (such as the location of the computer room, the heat dissipation effect between devices, etc.) to obtain the total virtual energy consumption of the entire server cluster.

[0048] Through a hierarchical prediction method, including the operation status prediction layer, the equipment energy consumption prediction layer and the total energy consumption prediction layer, a refined prediction of the virtual energy consumption of the data center is achieved. This implementation principle significantly improves the accuracy of energy consumption prediction, enhances the flexibility and reliability of the system, and provides strong support for energy consumption management and equipment scheduling in the data center.

[0049] Further optionally, the steps of constructing an energy consumption model matching each virtual device model can be implemented as follows: obtaining historical energy consumption data of each virtual device model, and fitting a corresponding historical energy consumption change curve based on the historical energy consumption data; selecting an original energy consumption model matching each virtual device model based on the device properties matching each virtual device model; and adaptively adjusting the original energy consumption model of each virtual device model according to the historical energy consumption change curve to obtain an energy consumption model matching each virtual device model.

[0050] In a further optional step, the implementation process of constructing the energy consumption model matching each virtual device model includes obtaining historical energy consumption data, fitting the historical energy consumption change curve, selecting the original energy consumption model, and adaptively adjusting the original energy consumption model.

[0051] First, obtain the historical energy consumption data of each virtual device model. These data usually include the energy consumption record, running time, load condition, ambient temperature, etc. of the device. Preprocess the collected historical energy consumption data, such as data cleaning, missing value filling, data standardization, etc., to ensure data quality. Then, based on the historical energy consumption data, use appropriate mathematical models (such as linear regression, polynomial regression, time series model, etc.) to fit the historical energy consumption change curve. Analyze the fitted historical energy consumption change curve to understand the law and trend of the energy consumption of the device over time. In this way, by fitting the historical energy consumption change curve based on the historical energy consumption data and using appropriate mathematical models (such as linear regression, polynomial regression, time series model, etc.), the energy consumption change law of the device can be accurately captured. This process involves multiple steps such as data collection and preprocessing, model selection, parameter estimation and model verification to ensure that the model can accurately reflect the actual energy consumption characteristics of the device and provide a reliable basis for energy consumption management and optimized scheduling of the data center. Then, according to the device attributes (such as device type, model, specification, etc.) matched by each virtual device model, select the matching original energy consumption model from the model library. The model library contains energy consumption models for various types of equipment, such as server models, storage device models, etc. The original energy consumption model can be an energy consumption model obtained from energy consumption data provided by the equipment manufacturer, public research data, or historical data. Finally, according to the fitted historical energy consumption change curve, the parameters of the original energy consumption model are adapted. This step includes calibrating the key parameters of the model (such as energy consumption coefficient, load sensitivity parameters, etc.) to more accurately reflect the actual energy consumption characteristics of the equipment. Use additional historical energy consumption data to verify the adjusted energy consumption model to ensure the accuracy and reliability of its prediction results.

[0052] In this way, by fitting the historical energy consumption change curve based on the historical energy consumption data, the energy consumption change law of the equipment can be captured more accurately. The original energy consumption model is adjusted to ensure that it can more accurately reflect the actual energy consumption characteristics of the equipment and improve the prediction accuracy of the model. The energy consumption model can be adjusted based on the dynamic changes of historical data, so that it can flexibly respond to changes in the operating status and environmental conditions of the equipment. Through equipment attribute matching, the energy consumption model library can support multiple equipment types to ensure that the model can cover all types of equipment in the data center. By fine-tuning the energy consumption model, the energy consumption of the equipment can be predicted more accurately, providing a reliable basis for energy consumption management and optimized scheduling. Through verification and calibration, the prediction results of the energy consumption model are ensured to be highly consistent with the actual energy consumption data, improving the reliability of the prediction. The use of accurate energy consumption prediction provides strong support for the energy consumption management of the data center and helps managers formulate more effective energy consumption optimization strategies. Through accurate energy consumption prediction and optimized scheduling, the energy consumption of the data center can be significantly reduced, the operating costs can be reduced, and the goal of energy conservation and emission reduction can be achieved.

[0053] Taking the construction of the server energy consumption model as an example, the historical energy consumption data of server A is collected, including energy consumption records, running time, load conditions, ambient temperature, etc. The historical energy consumption data of server A is fitted with a time series model to obtain the historical energy consumption change curve, and analyze the change pattern of energy consumption over time. According to the device attributes of server A (such as model: Xeon E5-2600), a matching original energy consumption model is selected from the model library. Based on the fitted historical energy consumption change curve, the parameters of the original energy consumption model are adjusted, such as the energy consumption coefficient and load sensitivity parameters, to more accurately reflect the actual energy consumption characteristics of server A. Additional historical energy consumption data is used to verify the adjusted energy consumption model to ensure the accuracy and reliability of its prediction results.

[0054] In this way, by acquiring historical energy consumption data, fitting historical energy consumption change curves, selecting original energy consumption models, and adaptively adjusting the original energy consumption models, energy consumption models that match each virtual device model are constructed, which significantly improves the accuracy, flexibility, and adaptability of the energy consumption model, enhances the reliability of energy consumption prediction, and provides strong support for energy consumption management and optimized scheduling of data centers.

[0055] In the above optional embodiment, optimizing the aggregation result based on the actual layout and / or application scenario of the data center to obtain the virtual energy consumption of the data center can be further implemented as follows:

[0056] First, construct the objective function corresponding to the data center; the objective function at least includes: minimizing the energy consumption function and maximizing the equipment utilization function. Then, based on the actual layout and / or application scenario of the data center, use the objective function to perform hybrid energy consumption optimization processing on the aggregation results to obtain a hybrid energy consumption optimization result; the hybrid energy consumption optimization result is the optimal solution or approximate optimal solution corresponding to the objective function in the actual layout and / or application scenario. Finally, analyze the hybrid energy consumption optimization result of the aggregation results to obtain the virtual energy consumption of the data center.

[0057] In the above optional embodiment, by constructing an objective function and performing hybrid energy consumption optimization processing under actual layout and application scenarios, the optimization calculation of the virtual energy consumption of the data center is further realized. Specifically, an objective function corresponding to the data center is constructed, and the objective function at least includes a minimization energy consumption function and a maximization equipment utilization function. The minimization energy consumption function aims to minimize the total energy consumption of the data center. It is usually expressed as minimizing the total energy consumption of all equipment. The maximization equipment utilization function aims to maximize the utilization of the equipment and ensure that the equipment is used in an efficient working state.

[0058] Based on the actual layout of the data center (such as the location of the computer room, the heat dissipation effect between devices, etc.) and application scenarios (such as load fluctuations, business needs, etc.), the constructed objective function is used to optimize the aggregated results for hybrid energy consumption. The optimization algorithm (such as genetic algorithm, particle swarm optimization, linear programming, etc.) is used to solve the optimal solution or approximate optimal solution of the objective function. The optimization algorithm needs to consider the trade-offs between multiple objectives to find a solution that can minimize energy consumption and maximize equipment utilization.

[0059] Analyze the hybrid energy consumption optimization results to obtain the virtual energy consumption of the data center. This step involves interpreting and analyzing the output of the optimization algorithm to ensure the rationality and accuracy of the results. Use historical data or simulation data to verify the optimization results to ensure their effectiveness in actual operation.

[0060] In one example, constructing the objective function corresponding to the data center includes:

[0061] Construct a minimum energy consumption function corresponding to the virtual device model i, and optimize the minimum energy consumption function using the weight of the influencing factor; wherein the weight of the influencing factor is determined by the importance of different influencing shadows to the overall energy consumption of the virtual device model i in the actual layout and / or application scenario; wherein the minimum energy consumption function is expressed as the following formula:

[0062]

[0063] Among them, T i represents the minimized energy consumption function corresponding to the virtual device model i, p i1,……,p in They represent the weights of the impact factors of different device attributes or monitoring data on the virtual device model i in the actual layout and / or application scenario, x i Indicates the real-time running status of virtual device model i, time i represents the running time of virtual device model i, and n is the total number of influencing factors of virtual device model i.

[0064] In another example, a maximized equipment utilization function corresponding to the virtual equipment model i is constructed; wherein the maximized equipment utilization function is expressed as the following formula:

[0065]

[0066] Among them, S i represents the maximum equipment utilization function corresponding to virtual equipment model i, x i Indicates the real-time running status of virtual device model i, d i represents the resource requirements of virtual device model i, c i It represents the resource capacity of virtual device model i, and m is the total number of virtual devices.

[0067] In this way, by simultaneously considering minimizing energy consumption and maximizing equipment utilization, the energy consumption management strategy of the data center can be optimized more comprehensively. Based on the optimization processing of the actual layout and application scenarios, the energy consumption management strategy can be dynamically adjusted to make it more suitable for the actual operating environment. The optimization algorithm can flexibly respond to changes in the layout and application scenarios of the data center to ensure the continued effectiveness of the energy consumption management strategy. By adaptively adjusting the objective function and optimization strategy, the system can better respond to the needs of different equipment and application scenarios. Through hybrid energy consumption optimization processing, the energy consumption of the data center can be significantly reduced and energy utilization efficiency can be improved. By maximizing equipment utilization, it can ensure that the equipment operates in a high-efficiency state and reduce resource waste. By optimizing energy consumption and equipment utilization, a more effective equipment scheduling strategy can be formulated to improve the overall operating efficiency of the data center. Through efficient energy consumption management and optimized scheduling, the energy consumption of the data center can be significantly reduced, operating costs can be reduced, and the goal of energy conservation and emission reduction can be achieved.

[0068] Based on the above example, in the above steps, based on the actual layout and / or application scenario of the data center, the objective function is used to perform hybrid energy consumption optimization processing on the aggregation results to obtain a hybrid energy consumption optimization result. This can be further implemented as: based on the actual layout and / or application scenario of the data center, the objective function is used to perform a hybrid solution on the aggregation results to obtain a hybrid energy consumption optimization result.

[0069] Among them, the hybrid solution process of the aggregation results is expressed as the following formula:

[0070]

[0071] Among them, O i represents the hybrid energy consumption optimization result of virtual device model i, w p represents the weight of the solution of the minimized energy consumption function corresponding to the virtual device model i, p i1 ,……,p in They represent the weights of the impact factors of different device attributes or monitoring data on virtual device model i, x i Indicates the real-time running status of virtual device model i, time i represents the running time of virtual device model i, n is the total number of influencing factors of virtual device model i, x i Indicates the real-time running status of virtual device model i, d i represents the resource requirements of virtual device model i, c i represents the resource capacity of virtual device model i, m is the total number of virtual devices, and w dc Represents the solution weight of the maximized equipment utilization function of virtual equipment model i.

[0072] By constructing the objective function and performing hybrid energy consumption optimization processing in actual layout and application scenarios, the accuracy of data center energy consumption management can be significantly improved, the flexibility and adaptability of the system can be enhanced, energy efficiency and equipment utilization can be improved, scheduling strategies can be optimized, and the goal of energy conservation and emission reduction can be achieved. This implementation principle and method provides strong support for energy consumption management and optimized scheduling of data centers.

[0073] 103. Optimize and schedule each device based on the virtual energy consumption, and continue to transmit the real-time operation data of each device after the optimization and scheduling to the cloud server for prediction, until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center.

[0074] Specifically, the particle swarm optimization algorithm (PSO) can be used to optimize the scheduling of each device based on the virtual energy consumption. In the particle swarm optimization algorithm, each particle represents a device scheduling scheme. During initialization, a group of particles are randomly generated, each particle contains the scheduling parameters of the device. Each particle can be represented as a vector x i =(x i1 , x i2 , ..., x in ), where x ijRepresents the scheduling parameters of the jth device of the ith particle. Scheduling parameters can be the running time, power configuration, load distribution, etc. of the device. The fitness function is used to evaluate the pros and cons of each particle (scheduling scheme). The fitness function is usually constructed based on virtual energy consumption and other performance indicators (such as equipment utilization). Energy consumption evaluation is to calculate the total energy consumption of each scheduling scheme based on the virtual energy consumption model. Equipment utilization evaluation is to calculate the equipment utilization of each scheduling scheme. The fitness function can be expressed as a multi-objective optimization problem, such as minimizing energy consumption and maximizing equipment utilization.

[0075] Then, the particle's velocity is updated according to the particle's current velocity, historical optimal position, and global optimal position. This calculation can be expressed as: vi_{ij}(t+1)=w\cdot vi_{ij}(t)+c_1\cdot r_1\cdot(pbest_{ij}-x_{ij}(t))+c_2\cdot r_2\cdot(gbest_j-x_{ij}(t)). Among them, vi_{ij}(t) is the velocity of the ith particle in the jth dimension, w is the inertia weight, c_1 and c_2 are acceleration constants, r_1 and r_2 are random numbers, pbest_{ij} is the historical optimal position of the ith particle, and gbest_j is the global optimal position. Then, the particle's position is updated according to its velocity. That is, x_{ij}(t+1)=x_{ij}(t)+vi_{ij}(t+1).

[0076] Next, each particle updates its historical optimal position according to the current fitness function value. The global optimal position is updated according to the historical optimal positions of all particles. Repeat steps 2 to 4 until the termination condition is met. The termination condition can be that the preset maximum number of iterations is reached, or the change in the fitness function value is less than a certain threshold. Finally, the optimal scheduling scheme is selected according to the global optimal position. The optimal scheduling scheme is applied to each device for actual optimization scheduling. The real-time operating data of each device after optimization scheduling is transmitted to the cloud server. Based on the transmitted real-time data, a prediction model (such as a time series model) is used to predict the virtual energy consumption of the data center. If the predicted virtual energy consumption does not meet the preset energy consumption management conditions, return to step 1 and re-optimize the scheduling.

[0077] By optimizing the scheduling of each device based on virtual energy consumption using a particle swarm optimization algorithm, the operation strategy of the device can be dynamically adjusted and optimized to improve the accuracy of energy management and the flexibility of the system. This process involves multiple steps such as particle initialization, fitness function calculation, particle position and velocity update, historical optimal position and global optimal position update, termination condition judgment, optimized scheduling execution, and real-time data transmission and prediction, ensuring efficient energy management and optimized scheduling of the data center under the optimal scheduling scheme.

[0078] In another example, a genetic optimization algorithm can also be used to optimize the scheduling of each device. Specifically, the genetic optimization algorithm (GA) based on virtual energy consumption is an optimization method that simulates natural selection and genetic mechanisms and is widely used in equipment scheduling optimization. The algorithm gradually iterates and optimizes the scheduling plan by initializing the population, calculating the fitness function, selecting, crossing, and mutating, and finally finds the individual with the highest fitness as the optimal scheduling plan. In the initialization stage, a group of individuals are randomly generated, each of which represents an equipment scheduling plan, including scheduling parameters such as the equipment's operating time, power configuration, and load distribution. When calculating the fitness function, the pros and cons of each scheduling plan are evaluated based on multi-objective indicators such as the virtual energy consumption model and equipment utilization. The selection operation uses the roulette selection method to select individuals with higher fitness according to the fitness function value for the next step. The crossover operation simulates gene recombination and generates new individuals through single-point crossover; the mutation operation introduces new genes and changes certain gene positions in individuals through random mutation. When updating the population, the original individuals with lower fitness are replaced with new individuals to form a new population. Repeat the above steps until the termination condition is met, and the optimal scheduling solution is obtained and applied to the equipment. Finally, through real-time data transmission and prediction, the scheduling strategy is dynamically adjusted to ensure efficient energy consumption management and optimized scheduling of the data center. This process not only simulates the evolutionary mechanism of nature, but also takes into account the multi-objective optimization of energy consumption and equipment utilization, providing a scientific and effective method for equipment scheduling.

[0079] In an embodiment of the present application, by transmitting real-time operating data to the virtual device model in the cloud server and performing optimized scheduling based on virtual energy consumption, the accuracy of energy consumption prediction is significantly improved, equipment scheduling is optimized, the energy efficiency of the data center is improved, and service quality and reliability are guaranteed.

[0080] In another embodiment of the present application, a data center energy consumption equipment scheduling system is provided. The data center energy consumption equipment scheduling system is applied to a data center. The data center is deployed with multiple devices. Figure 2 The data center energy consumption equipment scheduling system includes the following units:

[0081] An acquisition unit is configured to acquire real-time operation data of each device in the data center; the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data;

[0082] A prediction unit is configured to transmit the real-time operation data to a corresponding virtual device model in the cloud server, and simulate the real-time operation state of the data center through the virtual device model in the cloud server to obtain the virtual energy consumption of the data center;

[0083] The scheduling unit is configured to optimize the scheduling of each device based on the virtual energy consumption, and continue to transmit the real-time operation data of each device after the optimized scheduling to the cloud server for prediction until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center.

[0084] Further optionally, before the prediction unit transmits the real-time operation data to the corresponding virtual device model in the cloud server and simulates the real-time operation state of the data center through the virtual device model in the cloud server to obtain the virtual energy consumption of the data center, it is further configured to:

[0085] Determine a corresponding device type based on the device attribute information;

[0086] A virtual device prototype matching the device type is selected from a device library; the device library is dynamically configured according to the deployment information of the data center; the device library adaptively adjusts the device prototypes in the inventory as the function of the data center is expanded or reduced;

[0087] The device attribute information is matched to the corresponding setting items in the virtual device prototype to construct a corresponding virtual device model in the cloud server.

[0088] Further optionally, the prediction unit simulates the real-time operation status of the data center through a virtual device model in the cloud server to obtain the virtual energy consumption of the data center, and is configured as follows:

[0089] Through the operation status prediction layer, the real-time operation status of the virtual device model corresponding to each device is predicted;

[0090] Through the device energy consumption prediction layer, based on the real-time operating status of each virtual device model, the sub-virtual energy consumption corresponding to each virtual device model is predicted; wherein the sub-virtual energy consumption corresponding to each device is predicted based on the energy consumption model matched by each virtual device model;

[0091] The sub-virtual energy consumptions corresponding to the various virtual device models are aggregated through the total energy consumption prediction layer, and the aggregation results are optimized based on the actual layout and / or application scenarios of the data center to obtain the virtual energy consumption of the data center.

[0092] Further optionally, the prediction unit, the step of constructing the energy consumption model matching each virtual device model, is configured as follows:

[0093] Obtaining historical energy consumption data of each virtual device model, and fitting a corresponding historical energy consumption change curve based on the historical energy consumption data;

[0094] Based on the device attributes matched with each virtual device model, an original energy consumption model matched with each virtual device model is selected;

[0095] According to the historical energy consumption change curve, the original energy consumption model of each virtual device model is adaptively adjusted to obtain an energy consumption model matching each virtual device model.

[0096] Further optionally, the prediction unit optimizes the aggregation result based on the actual layout and / or application scenario of the data center to obtain the virtual energy consumption of the data center, and is configured to:

[0097] Constructing an objective function corresponding to the data center; the objective function at least includes: minimizing energy consumption function and maximizing equipment utilization function;

[0098] Based on the actual layout and / or application scenario of the data center, the objective function is used to perform hybrid energy consumption optimization processing on the aggregation result to obtain a hybrid energy consumption optimization result; the hybrid energy consumption optimization result is the optimal solution or approximate optimal solution corresponding to the objective function in the actual layout and / or application scenario;

[0099] The mixed energy consumption optimization result of the aggregation result is analyzed to obtain the virtual energy consumption of the data center.

[0100] Further optionally, the prediction unit constructs the objective function corresponding to the data center and is configured to:

[0101] Construct a minimum energy consumption function corresponding to the virtual device model i, and optimize the minimum energy consumption function using the weight of the influencing factor; wherein the weight of the influencing factor is determined by the importance of different influencing shadows to the overall energy consumption of the virtual device model i in the actual layout and / or application scenario; wherein the minimum energy consumption function is expressed as the following formula:

[0102]

[0103] Among them, T i represents the minimized energy consumption function corresponding to the virtual device model i, p i1 ,……,p in They represent the weights of the impact factors of different device attributes or monitoring data on the virtual device model i in the actual layout and / or application scenario, x i Indicates the real-time running status of virtual device model i, time i represents the running time of virtual device model i, and n is the total number of influencing factors of virtual device model i;

[0104] Construct the maximum equipment utilization function corresponding to the virtual equipment model i; wherein the maximum equipment utilization function is expressed as the following formula:

[0105]

[0106] Among them, S i represents the maximum equipment utilization function corresponding to virtual equipment model i, x i Indicates the real-time running status of virtual device model i, d i represents the resource requirements of virtual device model i, c i It represents the resource capacity of virtual device model i, and m is the total number of virtual devices.

[0107] Further optionally, the prediction unit, based on the actual layout and / or application scenario of the data center, uses the objective function to perform hybrid energy consumption optimization processing on the aggregation result to obtain a hybrid energy consumption optimization result, and is configured to:

[0108] Based on the actual layout and / or application scenario of the data center, the objective function is used to perform a hybrid solution on the aggregation results to obtain a hybrid energy consumption optimization result; wherein the hybrid solution process of the aggregation results is expressed as the following formula:

[0109]

[0110] Among them, O i represents the hybrid energy consumption optimization result of virtual device model i, w p represents the weight of the solution of the minimized energy consumption function corresponding to the virtual device model i, p i1 ,……,p in They represent the weights of the impact factors of different device attributes or monitoring data on virtual device model i, x i Indicates the real-time running status of virtual device model i, time i represents the running time of virtual device model i, n is the total number of influencing factors of virtual device model i, x i Indicates the real-time running status of virtual device model i, d i represents the resource requirements of virtual device model i, c i represents the resource capacity of virtual device model i, m is the total number of virtual devices, and w dc Represents the solution weight of the maximized equipment utilization function of virtual equipment model i.

[0111] In an embodiment of the present application, the system significantly improves the accuracy of energy consumption prediction, optimizes equipment scheduling, improves the energy efficiency of the data center, and ensures service quality and reliability by transmitting real-time operation data to the virtual device model in the cloud server and performing optimized scheduling based on virtual energy consumption.

[0112] In another embodiment of the present application, an intelligent computing platform is further provided, comprising: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0113] Memory, used to store computer programs;

[0114] The processor is used to implement the data center energy consumption equipment scheduling method described in the method embodiment when executing the program stored in the memory.

[0115] The communication bus 1140 mentioned in the above electronic device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 1140 may be divided into an address bus, a data bus, a control bus, etc.

[0116] The embodiment of the present application provides a data center energy consumption equipment scheduling method for constructing a low energy consumption operator.

[0117] For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0118] The communication interface 1120 is used for communication between the above electronic device and other devices.

[0119] The memory 1130 may include a random access memory (RAM) or a non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0120] The above-mentioned processor 1110 can be a general-purpose processor, including an artificial intelligence processor, a graphics processing unit (GPU), an artificial intelligence processor card (Machine Learning Unit, MLU), 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.

[0121] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed, can implement each step that can be performed by an electronic device in the above method embodiment.

Claims

1. A method for scheduling energy-consuming equipment in a data center, characterized in that: Applied to a data center, the data center is deployed with multiple devices, and the data center energy consumption equipment scheduling method includes: Acquire real-time operation data of each device in the data center; the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data; The real-time operation data is transmitted to a corresponding virtual device model in the cloud server, and the real-time operation state of the data center is simulated by the virtual device model in the cloud server to obtain the virtual energy consumption of the data center; Based on the virtual energy consumption, each device is optimized and scheduled, and the real-time operation data of each device after the optimization and scheduling is continuously transmitted to the cloud server for prediction, until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center.

2. The method for scheduling energy-consuming equipment in a data center according to claim 1, characterized in that: Before transmitting the real-time operation data to the corresponding virtual device model in the cloud server and simulating the real-time operation state of the data center through the virtual device model in the cloud server to obtain the virtual energy consumption of the data center, the method further includes: Determine a corresponding device type based on the device attribute information; A virtual device prototype matching the device type is selected from a device library; the device library is dynamically configured according to the deployment information of the data center; the device library adaptively adjusts the device prototypes in the inventory as the function of the data center is expanded or reduced; The device attribute information is matched to the corresponding setting items in the virtual device prototype to construct a corresponding virtual device model in the cloud server.

3. A data center energy consumption equipment scheduling system, characterized in that: Applied to a data center, the data center is deployed with multiple devices, and the data center energy consumption equipment scheduling system includes: An acquisition unit is configured to acquire real-time operation data of each device in the data center; the real-time operation data at least includes: device attribute information, real-time monitoring information, and device operation data; A prediction unit is configured to transmit the real-time operation data to a corresponding virtual device model in the cloud server, and simulate the real-time operation state of the data center through the virtual device model in the cloud server to obtain the virtual energy consumption of the data center; The scheduling unit is configured to optimize the scheduling of each device based on the virtual energy consumption, and continue to transmit the real-time operation data of each device after the optimized scheduling to the cloud server for prediction until the predicted virtual energy consumption of the data center meets the preset energy consumption management conditions of the data center.

4. An intelligent computing platform, characterized in that: The intelligent computing platform includes: at least one processor, memory, and input-output unit; The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the data center energy-consuming equipment scheduling method according to any one of claims 1 to 2.

5. A readable storage medium, characterized in that: The readable storage medium includes instructions for executing the data center energy-consuming equipment scheduling method according to any one of claims 1 to 2.

Images