A method for automatically configuring data center resources

Abstract

The present invention provides a method for automatic configuration of data center resources, which relates to the technical field of dynamic resource allocation. The method comprises: real-time monitoring of the load data of the data center, obtaining the real-time usage status of each resource, including the processor usage rate, memory occupancy rate, network bandwidth utilization rate and the usage of storage devices; according to the load data of the data center, real-time collection and analysis of the business demand data of the current running business and the business demand data to be launched, including business type, status, resource consumption, expected online time and expected resource demand, so as to obtain business analysis results. The present invention realizes the dynamic optimization configuration of data center resources through real-time monitoring, intelligent analysis and prediction, effectively improves resource utilization and business continuity, and reduces operating costs at the same time.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic resource allocation, and in particular to a method for automatic configuration of data center resources. Background Art

[0002] Traditional data centers are designed and built with the maximum load in mind and with room for future upgrades and improvements, but this approach may limit the flexibility and scalability of data centers. When new services need to be deployed quickly or resource configurations need to be adjusted, some traditional methods are slow to respond and cannot meet business needs. Some traditional data centers may have high operating costs due to low energy efficiency and high management complexity.

[0003] In addition, in terms of security, since some of them rely on manual operation and management, there may be data leakage or system failure caused by human errors or malicious behavior. The lack of effective automated tools and processes to ensure the consistency and security of configuration increases potential security risks. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method for automatic configuration of data center resources, which improves resource utilization, flexibility and response speed, while reducing management complexity and cost.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] A method for automatically configuring data center resources, the method comprising:

[0007] Monitor the load data of the data center in real time and obtain the real-time usage status of each resource, including processor usage, memory occupancy, network bandwidth utilization and storage device usage;

[0008] Based on the load data of the data center, the current running business and the business demand data to be launched are collected and analyzed in real time, including business type, status, resource consumption, expected launch time and expected resource demand, to obtain business analysis results;

[0009] According to the business analysis results, the real-time monitored load data and business demand data are input into the resource allocation prediction and optimization model for training and learning, and the resource demand in the future is predicted to obtain the final prediction results;

[0010] Based on the final prediction results, dynamically allocate and optimize data center resources, including adjusting the number of processor cores, memory allocation, and network bandwidth allocation;

[0011] Based on the dynamic allocation and optimization of data center resources, the corresponding resource configuration adjustments are automatically executed, and the operating status of the data center is continuously monitored to obtain new load data and business demand data in real time.

[0012] Furthermore, the data center load data collects and analyzes the current running business and the business demand data to be launched in real time, including business type, status, resource consumption, expected launch time and expected resource demand, to obtain business analysis results, including:

[0013] Integrate real-time business data and planned data of upcoming businesses into a unified data platform to obtain an integrated data set;

[0014] Clean and preprocess the integrated data set and extract key features to obtain the processed data set;

[0015] Based on the processed data set, analyze the resource usage patterns of currently running businesses, identify business types with high resource consumption or rapid growth, and mark them as critical businesses;

[0016] Prioritize the services to be launched based on the marked key services and business rules, expected revenue, and customer influencing factors, and output a business priority list;

[0017] Analyze the impact of different priority businesses on resources based on the business priority list and resource usage patterns to generate a business resource usage analysis report;

[0018] Use analysis reports based on business resources to obtain business analysis results.

[0019] Furthermore, based on the marked key businesses and business rules, expected revenues, and customer influencing factors, priority assessment is performed on the businesses to be launched, and a business priority list is output, including:

[0020] Setting evaluation criteria and metrics, including business urgency;

[0021] For each business that is about to be launched, collect relevant information, including business requirements documents and customer feedback, and use the relevant information as input data for evaluation criteria;

[0022] Score each business to be launched based on the set evaluation criteria and key business information;

[0023] During the scoring process, the business is assigned a corresponding score based on how well it matches each evaluation criterion;

[0024] Summarize the scores of each business to get a comprehensive score, and sort the businesses according to the comprehensive score, from high to low, to generate a business priority list.

[0025] Furthermore, based on the business analysis results, the real-time monitored load data and business demand data are input into the resource allocation prediction and optimization model for training and learning, and the resource demand in the future is predicted to obtain the final prediction results, including:

[0026] Based on the business analysis results, determine the characteristics related to resource demand forecasting, including historical load data, business growth trends, and business types;

[0027] Build resource allocation prediction and optimization models, and determine the input items of the optimization model, including real-time monitored load data, business demand data, and output items, including the predicted value of resource demand in the future period;

[0028] Divide historical load data and business demand data into training sets, validation sets, and test sets, and use the training set data to train the optimization model. During the training process, use the root mean square error to evaluate the prediction performance of the optimization model on the validation set to obtain the evaluation results.

[0029] According to the evaluation results, the optimization model is iteratively optimized, including adjusting parameters;

[0030] Repeat training, verification, and evaluation until the performance of the optimized model reaches the preset standard to obtain the final optimized model;

[0031] According to the final optimization model, the current real-time monitored load data and business demand data are used as input items to calculate the resource demand forecast value for a period of time in the future to obtain the final forecast result, including the resource demand forecast value, the estimated usage and growth trend of each type of resource.

[0032] Furthermore, the calculation formula of the root mean square error is:

[0033] R ;

[0034] Where R represents the root mean square error; Represents the total number of samples in the validation set; Indicates The actual value of samples; Indicates The predicted value of samples; Indicates The weight of the samples; represents the regularization coefficient; Indicates the number of model parameters; Indicates model parameters.

[0035] Furthermore, the calculation formula for the predicted value of resource demand in the future is:

[0036] ;

[0037] in, Indicates future time The resource demand forecast value; represents the intercept term of the model; An index representing the influencing factor; Indicates The weight of the influencing factors Indicates Factors affecting time Observed value Represents the deviation between the predicted value and the actual value.

[0038] Furthermore, according to the dynamic allocation and optimization of data center resources, the corresponding resource configuration is automatically adjusted, and the operation status of the data center is continuously monitored to obtain new load data and business demand data in real time, including:

[0039] Receive the final forecast results and automatically trigger the resource allocation adjustment process;

[0040] Determine the corresponding resource configuration adjustment plan based on the prediction results, including increasing or decreasing server resources, adjusting network bandwidth configuration, or reallocating storage resources;

[0041] Send the resource configuration adjustment plan to the management system of the data center and perform the resource configuration adjustment operation;

[0042] Continuously monitor the operating status of the data center, including key indicators of processor utilization, memory occupancy, network bandwidth utilization, and storage device usage, to obtain new load data and business demand data in real time, including real-time traffic, user request volume, and transaction processing volume.

[0043] The above solution of the present invention includes at least the following beneficial effects:

[0044] By monitoring the real-time usage status of resources through the data center management system, changes in resource usage can be quickly captured; real-time collection and analysis of business demand data for running services and those about to be launched enables the system to quickly respond to changes in business demand; training and learning using resource allocation prediction and optimization models can predict resource demand in the future, thereby achieving advance planning and optimal allocation of resources. This prediction capability helps avoid situations where resources are in excess or insufficient, improves resource utilization, and reduces waste.

[0045] According to the final prediction results, the data center resources are dynamically allocated and adjusted, including the number of processor cores, memory allocation, network bandwidth allocation, etc., which enhances the flexibility and scalability of the data center. This dynamic adjustment capability enables the data center to better adapt to the rapid changes in business needs and improves the operational efficiency and service quality of the data center. The system automatically performs adjustments to the corresponding resource configuration based on the dynamic allocation and optimization results, reducing the need for manual intervention; automated configuration not only improves configuration efficiency, but also reduces the possibility of human error, and improves the stability and reliability of the data center. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flowchart of a method for automatic configuration of data center resources provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0048] like Figure 1 As shown, an embodiment of the present invention provides a method for automatically configuring data center resources, the method comprising the following steps:

[0049] Step 11: monitor the load data of the data center in real time and obtain the real-time usage status of each resource, including processor usage, memory occupancy, network bandwidth utilization, and storage device usage;

[0050] Step 12: Based on the load data of the data center, collect and analyze the demand data of the currently running services and the services to be launched in real time, including the service type, status, resource consumption, expected launch time and expected resource demand, to obtain the service analysis results;

[0051] Step 13: Based on the business analysis results, the real-time monitored load data and business demand data are input into the resource allocation prediction and optimization model for training and learning, and the resource demand in the future is predicted to obtain the final prediction result;

[0052] Step 14: dynamically allocate and optimize data center resources based on the final prediction results, including adjusting the number of processor cores, memory allocation, and network bandwidth allocation;

[0053] Step 15, according to the dynamic allocation and optimization of the data center resources, the corresponding resource configuration is automatically adjusted, and the operation status of the data center is continuously monitored to obtain new load data and business demand data in real time.

[0054] In an embodiment of the present invention, by real-time monitoring of the resource usage status of the data center, such as processor usage, memory occupancy, etc., the administrator can obtain instant feedback on the performance of the data center. This transparency helps to quickly identify and resolve potential performance bottlenecks or improper resource usage issues. Continuous resource status monitoring can detect abnormal situations, so that preventive measures can be taken before problems occur, reduce unexpected downtime, and improve the availability of the data center. Real-time collection and analysis of current and future business demand data enables the data center to more accurately understand the development trends and resource requirements of the business, so as to make more timely responses. By conducting demand analysis on the business that is about to go online, the data center can conduct forward-looking resource planning to ensure that sufficient resources are available when the new business goes online, avoiding service delays or interruptions due to insufficient resources.

[0055] Using resource allocation prediction and optimization models, data centers can predict future resource needs based on historical and real-time data, providing a scientific basis for resource allocation decisions. Accurate resource demand prediction helps data centers find a balance between avoiding excess and insufficient resources, thereby reducing operational risks and costs. Dynamically adjusting resources based on the final forecast results can ensure that data center resources are always utilized and improve overall operational efficiency. Dynamic resource allocation enables data centers to respond more flexibly to changes in business needs, whether it is a sudden increase or decrease in business volume, and can make adjustments quickly. The system automatically performs resource configuration adjustments based on resource allocation results, reducing the need for manual intervention and thus reducing the risks caused by human operational errors. By continuously monitoring the operating status of the data center and acquiring new data in real time, the system can continuously optimize resource allocation to ensure that the data center is always in the best performance state.

[0056] In a preferred embodiment of the present invention, the above step 11 monitors the load data of the data center in real time to obtain the real-time usage status of each resource, including the processor usage rate, memory occupancy rate, network bandwidth utilization rate and storage device usage, which may include:

[0057] Sensors, such as temperature sensors, humidity sensors, voltage and current sensors, are deployed on key equipment in the data center to collect the operating status data of the equipment in real time. The data collector is responsible for receiving the raw data from the sensor and performing preliminary processing and formatting. For example, DHT11, DHT22, etc. are data collectors. The data collector sends the processed data to the monitoring platform of the data center management system through a communication module (such as Ethernet). The communication module ensures the real-time and accurate transmission of data. Standardized data transmission protocols, such as MQTT, are used to ensure the consistency and reliability of data during transmission. After receiving the raw data, the monitoring platform performs data cleaning (removing outliers, noise, etc.) and data aggregation (integrating data from multiple data sources together) to provide an accurate and comprehensive view of resource usage status.

[0058] The monitoring platform displays the usage status of each resource in real time through a graphical interface, such as bar charts, line charts, pie charts, etc. Set resource usage thresholds (such as processor usage exceeding 80%). When actual usage exceeds these thresholds, the system automatically triggers an alarm and notifies management personnel through emails, text messages, etc., so that timely measures can be taken to prevent potential problems.

[0059] When it comes to specific applications, assume that a data center management system is monitoring a computer room containing 100 servers. Temperature sensors and voltage sensors are deployed on each server to monitor the temperature and voltage changes of the server in real time. The data collector collects data from the sensor every 5 seconds and transmits the data to the database of the monitoring platform via Ethernet. The data processing module of the monitoring platform cleans and aggregates the received data and calculates key indicators such as the average temperature, maximum temperature, and voltage fluctuation of each server. The average temperature of each server is ;in, Indicates the number of temperature samples collected in one data collection cycle; Indicates The highest temperature is ;in, Indicates The voltage fluctuation is calculated as ;in, Indicates voltage samples. These indicators are displayed to managers in the form of charts through a visual interface. When the temperature of a server exceeds a preset threshold (such as 65 degrees Celsius), the system automatically sends an alarm email to managers, prompting them to check whether the cooling system of the server is working properly.

[0060] In a preferred embodiment of the present invention, the above step 12 collects and analyzes the demand data of the currently running services and the services to be launched in real time based on the load data of the data center, including the service type, status, resource consumption, expected launch time and expected resource demand, to obtain the service analysis results, which may include:

[0061] Step 122, integrating the real-time business data and the planned data of the business to be launched into a unified data platform to obtain an integrated data set;

[0062] Step 123, cleaning and preprocessing the integrated data set, extracting key features, to obtain a processed data set;

[0063] Step 124, analyzing the resource usage pattern of the currently running business based on the processed data set, identifying the business types with large resource consumption or fast growth, and marking them as key businesses;

[0064] Step 125, based on the marked key businesses and business rules, expected revenues and customer influencing factors, priority assessment is performed on the businesses to be launched, and a business priority list is output;

[0065] Step 126, analyzing the impact of different priority services on resources according to the service priority list and resource usage pattern to generate a service resource usage analysis report;

[0066] Step 127, obtaining a business analysis result according to the business resource usage analysis report.

[0067] In an embodiment of the present invention, real-time business data from different sources (such as databases, log files, APIs, etc.) and planned data of the upcoming business are integrated into a unified data platform through a data extraction, transformation, and loading (ETL) process. These data include key information such as business type, status, resource consumption, expected time to go online, and expected resource requirements. The integrated data set is cleaned to remove duplicate, erroneous or incomplete data. Data preprocessing, such as standardization and normalization, is performed to eliminate noise and outliers in the data. Key features, such as resource consumption patterns and growth trends of business types, are extracted to form a processed data set.

[0068] Use data analysis techniques (such as statistical analysis) to conduct in-depth analysis on the processed data set. First, calculate the resource consumption indicators of each business type, such as average value, peak value, etc. The calculation formula for the average value is A ;in, Indicates the total number of collections; Indicates Resource consumption data collected. ;in, Indicates Collect resource consumption data. Compare resource consumption indicators with set thresholds to identify business types with large or fast-growing resource consumption. Mark the identified business types with large or fast-growing resource consumption as key businesses. Verify the results of the marked key businesses to ensure their accuracy and rationality. If mislabeling or omissions are found, adjust the analysis method and threshold settings in a timely manner to optimize the analysis process. Output the analysis results in the form of a report, including the resource consumption of each business type, the identification results of key businesses, and the corresponding analysis charts and instructions. Combine the marked key businesses, business rules (such as service level agreements SLA, business importance, etc.), expected benefits, and customer influencing factors (such as customer satisfaction, market demand, etc.), to build a priority assessment model.

[0069] Prioritize the businesses that are about to be launched and output a business priority list to guide resource allocation and launch plans. Analyze the impact of different priority businesses on resources based on the business priority list and the resource usage patterns of currently running businesses. Generate business resource usage analysis reports using visualization tools (such as charts, dashboards, etc.) to display resource consumption, growth trends, and estimated resource requirements for each business type. Based on the business resource usage analysis report, obtain business analysis results, including resource optimization suggestions, business adjustment strategies, etc. Provide analysis results to decision-makers and related teams to support business decision-making and resource management.

[0070] By analyzing business data in real time and predicting resource needs, resources can be allocated and adjusted more accurately to avoid waste and shortage of resources. Based on comprehensive business analysis results, decision-makers can make more informed business decisions, such as adjusting business strategies and optimizing product portfolios. By prioritizing key businesses and meeting customer needs, customer satisfaction and loyalty can be improved, thereby increasing market share and revenue. Real-time monitoring and analysis of business data can help companies identify potential risks and problems in a timely manner, and take appropriate countermeasures to reduce business risks. By building a unified data platform and analysis process, companies can be driven to transform into a data-driven culture and improve overall operational efficiency and innovation capabilities.

[0071] When specifically applied, the above content specifically includes:

[0072] Collect data about currently running businesses in real time, including business type, status, resource consumption (such as CPU, memory, storage, etc.). Collect planning data for businesses that are about to go online, including business type, expected time to go online, expected resource requirements, etc. Integrate this data into a unified data platform, such as a data warehouse or big data lake. Clean the integrated data set to remove duplicate, erroneous, or incomplete data.

[0073] Extract key features such as business type, resource consumption pattern, expected launch time, etc. for deeper analysis. Analyze resource usage patterns of currently running businesses using statistical analysis and machine learning algorithms.

[0074] Identify business types that consume large resources or grow rapidly. These business types usually have a significant impact on the stability of the system and are therefore marked as critical businesses. Prioritize the businesses that are about to be launched based on the marked critical businesses, business rules (such as business importance, urgency, etc.), expected benefits, and customer influencing factors (such as customer satisfaction, user growth, etc.). Output a business priority list for resource allocation and scheduling. Based on the business priority list and resource usage patterns, analyze the impact of different priority businesses on resources. Generate a business resource usage analysis report, which should include resource demand forecasts, resource bottleneck identification, resource optimization recommendations, etc. Based on the business resource usage analysis report, obtain business analysis results. These results can be used to guide decisions such as resource allocation, business scheduling, and system optimization.

[0075] Assume that an e-commerce platform is running multiple businesses, including product search, order processing, user recommendation, etc. At the same time, the platform also plans to launch a new promotion business. Collect the operation data of services such as product search, order processing, and user recommendation in real time, including business type, status, resource consumption, etc. Collect the planned data of the new promotion business, including business type (promotion), expected launch time (next Wednesday), expected resource requirements (large amounts of CPU and memory resources), etc. Clean and preprocess the collected data to extract key features, such as business type, resource consumption pattern, etc. Analysis shows that the product search business consumes a lot of resources during peak hours and grows rapidly, so it is marked as a key business.

[0076] Based on key businesses (product search), business rules (promotions have a significant impact on user experience and sales), expected benefits (promotions are expected to bring a large increase in sales), and customer influencing factors (users look forward to promotions), the priority of the new promotion business is evaluated as high. Analyzing the impact of the new promotion business on resources, it is found that after going online, this business may compete with the product search business for a large amount of CPU and memory resources, resulting in resource bottlenecks. Generate a business resource usage analysis report and recommend increasing resource quotas or optimizing resource scheduling strategies in advance to ensure the stable operation of the new promotion business and key businesses. According to the business resource usage analysis report, the business analysis results are obtained: the new promotion business has a high priority, but attention should be paid to resource competition with the product search business. It is recommended to plan and optimize resources in advance to ensure stable business operation and user experience.

[0077] In another preferred embodiment of the present invention, the above step 125, based on the marked key businesses and business rules, expected revenue and customer influencing factors, performs priority evaluation on the businesses to be launched, and outputs a business priority list, which may include:

[0078] Step 1255, setting evaluation criteria and metrics, including the urgency of the business;

[0079] Step 1256, for each business to be launched, collect relevant information, including business requirement documents and customer feedback, and use the relevant information as input data for evaluation criteria;

[0080] Step 1257, marking the business types identified as those with large resource consumption or rapid growth as key business information, and scoring each business to be launched based on the set evaluation criteria and key business information;

[0081] Step 1258, during the scoring process, assign a corresponding score according to the degree of match between the business and each evaluation criterion;

[0082] Step 1259, summarize the scores of each business to obtain a comprehensive score, and sort the businesses according to the comprehensive score, from high to low, to generate a business priority list.

[0083] In the embodiment of the present invention, the urgency of launching the service is evaluated, such as whether it involves seasonal activities, market response speed requirements, etc. The economic benefits, market share growth, user activity improvement, etc. that may be brought about by the service launch are analyzed. The potential impact of the service on customer experience, satisfaction, retention rate, etc. is considered.

[0084] Evaluate the resources required for the launch of the service (such as servers, storage, bandwidth, etc.) and its potential impact on existing system resources. For each service to be launched, collect its business requirements documents to understand the specific functions, technical requirements and expected goals of the service. Conduct expected benefit analysis, including market research, user research, etc., to quantify the potential benefits after the service is launched. Collect customer feedback to understand user needs and expectations for similar services and possible market reactions. Based on the previous resource usage pattern analysis, mark the business types that have been identified as high resource consumption or fast growth as key businesses. For each service to be launched, score each item according to the set evaluation criteria and key business information. During the scoring process, the business is assigned a corresponding score according to the degree of match between the business and each evaluation criterion. For example, a business with high business urgency, high expected benefits, positive customer impact and reasonable resource requirements will receive a high score. Sum up the scores of each service to obtain a comprehensive score. Sort the services according to the comprehensive score, from high to low, to generate a business priority list.

[0085] By defining the evaluation criteria and metrics, the priority of the services to be launched can be quickly and accurately evaluated, thus improving decision-making efficiency. By considering the resource demand as an evaluation criterion, system resources can be allocated and scheduled more effectively, avoiding resource waste and bottlenecks. Taking customer impact as one of the evaluation criteria can ensure that the services can better meet customer needs after they are launched, thus improving customer satisfaction and loyalty. By comprehensively considering factors such as business urgency, expected benefits, and resource requirements, the development of different services can be more effectively coordinated, thus promoting the overall coordinated development of the services. By evaluating and analyzing the priority and resource requirements of the services to be launched, the operation and resource consumption of the services after they are launched can be better predicted.

[0086] For example, in step 1258, for each service that is about to go online (assuming that service ), its score under each evaluation criterion can be calculated by the following formula :

[0087] ;

[0088] in, represents winter sales; represents spring sales; represents summer sales; represents fall sales; Indicates customer feedback score; represents the historical sales growth rate; Indicates changes in competitor market share; Indicates changes in competitor prices; Indicates changes in competitor advertising expenditures; , , , , Represents the coefficient.

[0089] When specifically applied, the above content specifically includes:

[0090] Check whether the business is related to a specific season or festival, such as Spring Festival promotions, summer travel products, etc. Assess the urgency of the market demand for the business and the dynamics of competitors to determine whether it is necessary to launch quickly to seize the market opportunity. Predict the growth of economic indicators such as sales and profits after the business is launched through market research. Analyze whether the business can attract new users, expand market share, and increase brand awareness after it is launched. Evaluate the role of the business in promoting user activity, such as increasing the number of user visits and increasing user stay time. Analyze the improvement or potential problems of the customer experience after the business is launched, such as interface friendliness and ease of operation. Evaluate the impact of the business on customer satisfaction and retention rate, such as whether it can improve user satisfaction and reduce user churn. List in detail the infrastructure resources such as servers, storage, bandwidth, etc. required for the business launch, and evaluate its potential impact on existing system resources. Based on business characteristics and historical data, predict the resource consumption of the business after it is launched so as to prepare in advance. For each business to be launched, collect its detailed requirements documents, including functional requirements, technical requirements, expected goals, etc. Through market research and user research, quantify the potential benefits of the business after it goes online, including economic benefits, market share growth, and increased user activity. Collect user needs and expectations for similar businesses, as well as possible market reactions, to evaluate the customer acceptance and market potential of the business. Based on the previous analysis of resource usage patterns, mark business types with high resource consumption or rapid growth as key businesses as the focus of evaluation. For each business to be launched, score each item according to the set evaluation criteria and key business information. During the scoring process, assign a corresponding score based on the degree of match between the business and each evaluation criterion. Summarize the scores of each business to obtain a comprehensive score. Sort the businesses according to the comprehensive score, from high to low, to generate a business priority list.

[0091] Suppose an e-commerce platform plans to launch two new businesses: one is a promotion for New Year goods during the Spring Festival, and the other is a fashion and trendy goods section for young users. The New Year goods promotion is closely related to the seasonal event of the Spring Festival, and the market demand is strong, so it has a high urgency. Although the fashion and trendy goods section is not restricted by season, there are similar sections launched in the market, and the young user group has a high demand for new and trendy goods, so it also has a certain urgency. The New Year goods promotion is expected to bring significant sales growth and increase the platform's share in the New Year goods market. The fashion and trendy goods section is expected to attract more young users, increase user activity and overall platform revenue, and enhance the platform's brand influence among young user groups. The New Year goods promotion needs to optimize the platform's shopping experience, such as providing convenient payment and logistics services to meet users' shopping needs during the Spring Festival. The fashion and trendy goods section needs to pay attention to the aesthetics and shopping habits of young users, provide personalized recommendations and high-quality customer service to improve customer satisfaction and retention. Promotions may require additional server and bandwidth resources to cope with the surge in user visits during the Spring Festival. The fashion and trendy product area may require more storage resources to store product images and videos, as well as a more powerful recommendation system to provide personalized services. Collect the requirements documents of the two businesses to understand the specific content of the promotion, the setting of the product area, and the expected goals. Through market research and user research, predict the potential benefits of the two businesses after they go online, including sales growth, market share increase, and user activity increase. Collect user needs and expectations for New Year promotions and fashion and trendy product areas, as well as possible market reactions. Based on the analysis of resource usage patterns, mark the New Year promotion as a key business because it consumes a lot of resources and is closely related to user peaks.

[0092] In a preferred embodiment of the present invention, the above step 13, based on the business analysis results, inputs the real-time monitored load data and business demand data into the resource allocation prediction and optimization model for training and learning, and predicts the resource demand in the future period to obtain the final prediction result, which may include:

[0093] Step 133, determining features related to resource demand forecasting based on the business analysis results, including historical load data, business growth trends, and business types;

[0094] Step 134, constructing a resource allocation prediction and optimization model, and determining the input items of the optimization model, including real-time monitored load data, business demand data, and output items, including resource demand forecast values ​​for a period of time in the future;

[0095] Step 135, dividing the historical load data and the business demand data into a training set, a validation set, and a test set, and using the training set data to train the optimization model. During the training process, the root mean square error is used to evaluate the prediction performance of the optimization model on the validation set to obtain an evaluation result;

[0096] Step 136, iteratively optimizing the optimization model according to the evaluation results, including adjusting parameters;

[0097] Step 137, repeating the training, verification and evaluation until the performance of the optimization model reaches a preset standard to obtain a final optimization model;

[0098] Step 138, based on the final optimization model, the current real-time monitored load data and business demand data are used as input items to calculate the resource demand forecast value for a period of time in the future to obtain the final forecast result, including the resource demand forecast value, the estimated usage and growth trend of each type of resource.

[0099] In an embodiment of the present invention, the system load data over a period of time is collected and analyzed, including CPU usage, memory usage, disk I / O, etc., to understand the historical consumption pattern of system resources. The growth trend of the business is analyzed, including user growth, transaction volume growth, etc., to predict the demand for resources of the future business. The differences in resource requirements of different types of business, such as computationally intensive, are identified to more accurately predict resource requirements. The calculation formula for the intensive type is: ;in, Indicates the total CPU resources required for intensive services; Indicates the basic CPU requirement; Indicates the user growth; Indicates the CPU consumption per user; Indicates the growth of transaction volume; Indicates the CPU consumption per transaction. By considering the basic CPU demand, the additional CPU consumption caused by user growth, and the additional CPU consumption caused by transaction volume growth, the total CPU resources required by the system in a certain period of time are dynamically calculated. This helps system administrators or operation and maintenance personnel to plan and allocate CPU resources in advance according to business growth trends and resource consumption patterns to ensure system stability and performance. Real-time monitoring of load data (such as current CPU usage, memory usage, etc.) and business demand data (such as expected user growth, transaction volume, etc.). Resource demand forecasts for a period of time in the future, including the expected usage and growth trends of various types of resources.

[0100] Select time series analysis to build an optimization model. Divide historical load data and business demand data into training set, validation set, and test set. The training set is used to train the model, the validation set is used to evaluate the performance of the optimization model, and the test set is used to finally verify the model effect. Use the training set data to train the model, and by continuously adjusting the model parameters, the optimization model can more accurately predict resource demand. During the training process, use evaluation indicators such as root mean square error (R) to evaluate the prediction performance of the model on the validation set. If the performance is poor, perform iterative optimization. According to the evaluation results, adjust the parameters of the model, such as changing the learning rate, increasing the number of iterations, etc., to improve the prediction performance of the model. Repeat the training, validation, and evaluation process until the model performance meets the preset standard, such as R is reduced to an acceptable range. The current real-time monitored load data and business demand data are used as input items and input into the final optimization model. The model outputs the predicted value of resource demand in the future, including the expected usage of various types of resources, growth trends, etc.

[0101] Based on the forecast results, a reasonable resource allocation plan is determined to ensure that the system can run stably and meet business needs. By accurately predicting resource requirements in the future, system resources can be allocated and utilized more effectively to avoid resource waste and shortages. Based on the resource demand forecast results, system configuration and resource expansion can be adjusted in a timely manner to ensure that the system can maintain stable operation under high load. Through reasonable resource allocation and optimization, unnecessary resource procurement and operation and maintenance costs can be reduced, and overall economic benefits can be improved. Accurately predicting resource requirements and preparing in advance can respond to business changes and market demands more quickly and improve business competitiveness.

[0102] When specifically applied, the above content specifically includes:

[0103] Collect system load data over a period of time (such as one year), including CPU usage, memory usage, disk I / O, etc., as well as corresponding business data, such as user growth, transaction volume growth, etc.

[0104] Monitor the current system load data and business demand data in real time as a supplement to the model input. Clean the data to remove outliers and missing values. Sort the data by time series and divide it into training set, validation set, and test set. For example, 70% of the data can be used as the training set, 15% of the data as the validation set, and 15% of the data as the test set. According to the characteristics of the data and the forecasting requirements, select a suitable time series analysis model, such as ARIMA, and use the training set data to train the model. By continuously adjusting the model parameters (such as learning rate, number of iterations, etc.), the model can more accurately predict resource requirements. During the training process, use evaluation indicators such as root mean square error (R) to evaluate the prediction performance of the model on the validation set. If the model performance is poor, iterative optimization is performed. According to the evaluation results, adjust the parameters of the model, such as changing the learning rate, increasing the number of iterations, etc. Repeat the training, validation, and evaluation process until the model performance reaches the preset standard, such as R is reduced to an acceptable range.

[0105] The current real-time monitored load data and business demand data are used as input items and input into the optimized resource allocation prediction and optimization model. The model outputs the resource demand forecast value for a period of time in the future (such as one month), including the expected usage and growth trend of each type of resource. According to the forecast results, a reasonable resource allocation plan is determined. For example, if it is predicted that the CPU usage rate will increase significantly in the future, it is necessary to increase CPU resources in advance; if it is predicted that the memory usage will reach a critical value, it is necessary to consider expanding the memory capacity.

[0106] Suppose there is an online shopping platform that needs to predict the CPU resource demand for the next month. Here are the specific steps: Collect CPU usage data, user growth data, and transaction volume data for the past year.

[0107] Clean and sort the data, and divide it into training set, validation set, and test set. Select the LSTM model for time series analysis. Use the training set data to train the LSTM model and adjust the model parameters to minimize R. Evaluate the model performance on the validation set, and perform iterative optimization if the performance is poor. Input the current real-time monitored CPU usage, user growth data, and transaction volume data into the optimized LSTM model. The model outputs the predicted value of CPU resource demand for the next month, including the daily average CPU usage, peak CPU usage, etc. Determine the resource allocation plan based on the prediction results. For example, if the peak CPU usage is predicted to reach 90% on a certain day in the future, you need to increase CPU resources in advance to ensure system stability.

[0108] In a preferred embodiment of the present invention, the calculation formula of the root mean square error is:

[0109] R ;

[0110] Where R represents the root mean square error; Represents the total number of samples in the validation set; Indicates The actual value of samples; Indicates The predicted value of samples; Indicates The weight of the samples; represents the regularization coefficient; Indicates the number of model parameters; Indicates model parameters.

[0111] In this embodiment of the present invention, sample data of the validation set is collected, including actual values and predicted values , and the weight of each sample . Determine the model parameters and the regularization coefficient For each sample in the validation set, the square of the difference between the actual value and the predicted value is calculated, that is, . Multiply these squared differences by the corresponding sample weights , and sum them to get the sum of squares of weighted prediction errors: For each parameter in the model, calculate its square, that is , multiplying these square sums by the regularization coefficient , get the regularization term of the model parameters The sum of squares of the weighted prediction errors is added to the regularization term of the model parameters to obtain the total error term . Divide the total error term by the sum of the sample weights , to get the average error. Take the square root of the average error to get the root mean square error (R).

[0112] R not only considers the prediction accuracy of the model (measured by the sum of squares of weighted prediction errors), but also introduces regularization terms to penalize the complexity of model parameters, thereby avoiding the problem of model overfitting.

[0113] By introducing the regularization term, R encourages the model to keep the parameters simple while fitting the training data, which helps improve the model's generalization ability on unknown data. Allows us to flexibly adjust according to the importance of different samples. R is an intuitive metric, and the smaller its value, the better the predictive performance of the model. This makes it easy to compare the predictive performance of different models or different parameter settings.

[0114] In a preferred embodiment of the present invention, the calculation formula for the resource demand forecast value in the future period is:

[0115] ;

[0116] in, Indicates future time The resource demand forecast value; represents the intercept term of the model; An index representing the influencing factor; Indicates The weight of the influencing factors Indicates Factors affecting time Observed value Represents the deviation between the predicted value and the actual value.

[0117] In the embodiment of the present invention, it is determined that the future time to be predicted is Resource requirements Identify and select 10 key factors that will impact future resource requirements ( =1, 2, ..., 10), these factors are historical resource usage, economic conditions, population changes, etc. Collect time series data, including historical resource demand and time series values ​​of various influencing factors. Clean the data, handle missing values ​​and outliers, and ensure the integrity and accuracy of the data. Standardize or normalize the data as needed. According to the prediction target and influencing factors. In this model, is the dependent variable, is the intercept term, It is The weight of the influencing factors, It is Factors affecting time The observed value of is the deviation between the predicted value and the actual value. The model is expressed in matrix form. Let is the resource demand vector (containing historical data and placeholders for future time points that need to be predicted), is the influencing factor matrix (including the intercept term and the time series values ​​of 10 influencing factors), is the parameter vector to be estimated (including the intercept term and 10 weights), is the error vector. The parameter vector is estimated by minimizing the sum of squares of the deviations between the predicted values ​​and the actual values. First, calculate the influencing factor matrix Transpose T. The equation is: ; According to the equation, we get the parameter vector Using the estimated parameter vector and future influencing factors here Refers to the value at a future time point, substituted into the model formula , calculate the future time The resource demand forecast .

[0118] The accuracy of resource demand forecasts can be improved by comprehensively considering multiple influencing factors and using historical data for parameter estimation. Accurate resource demand forecasts help decision makers such as enterprises and governments determine reasonable resource allocation plans and avoid resource waste or shortages. Based on the forecast results, the allocation and use of resources can be optimized, resource utilization efficiency can be improved, and costs can be reduced. Forecast models can help identify potential resource risks, such as supply disruptions and demand surges, so that measures can be taken in advance for risk management. Accurate resource demand forecasts help balance the relationship between economic development and resource protection.

[0119] In a preferred embodiment of the present invention, the above step 14, dynamically allocating and optimizing the resources of the data center according to the final prediction result, including adjusting the number of processor cores, memory allocation, and network bandwidth allocation, may include:

[0120] Dynamically adjust the allocation of processor cores based on the results of the prediction model. For example, when a high load period is predicted, the number of processor cores allocated to a specific application can be increased to ensure performance. Through the load balancing algorithm, tasks are evenly distributed to multiple processor cores to avoid overloading some cores while others are idle. Dynamically adjust memory allocation based on the application's memory demand prediction. For memory-intensive applications, more memory resources can be allocated. Regularly recycle unused memory resources and reallocate them to applications that need them to improve memory utilization. Use the prediction model to predict future network traffic so that network bandwidth allocation can be adjusted in a timely manner. Set bandwidth priorities for different applications to ensure that critical applications can still obtain sufficient bandwidth when the network is congested. Through the data center management system, key indicators such as processor utilization, memory usage, and network bandwidth consumption can be monitored in real time. Dynamically adjust resource allocation based on real-time monitoring data and prediction results. For example, when it is detected that the actual resource demand of an application exceeds the predicted value, the resources allocated to the application can be increased immediately. Establish a feedback mechanism to collect user feedback on the effectiveness of resource allocation. This feedback can be used to improve the prediction model and dynamic allocation strategy. Regularly evaluate and optimize forecasting models and dynamic allocation strategies to adapt to changes in data center resource demands.

[0121] When specifically applied, the above content specifically includes:

[0122] Use a load balancing algorithm, such as round-robin, to evenly distribute tasks across multiple processor cores to avoid overloading some cores while leaving others idle. Specifically, get the number of processor cores available in the system, which can be done through the API provided by the operating system. Set a polling pointer to track the processor core to which the task should be currently assigned. Set the initial value to 0. Create a task queue to store the tasks to be assigned. Tasks can be computational tasks, I / O operations, or other operations that require processor resources. Take a task from the task queue. Based on the current value of the polling pointer, assign the task to the corresponding processor core. Update the polling pointer to point to the next processor core. If the polling pointer has reached the maximum number of cores, reset it to 0 to form a loop. Submit the task to the assigned processor core for execution.

[0123] If the system supports multi-threading or multi-processing, you can use the corresponding mechanism to bind tasks to a specified core for execution. During the task execution, monitor the load of each processor core. If a core is found to be overloaded or idle for too long, you can dynamically adjust the polling strategy, such as increasing or decreasing the amount of tasks assigned to the core.

[0124] Assume that the number of users accessing an online game surges between 8 and 10 p.m. on weekends, causing the processor to be overloaded. Using the prediction model, the administrator can increase the number of processor cores for the game in advance and configure a load balancing strategy to ensure smooth game operation.

[0125] Predict future memory requirements based on historical memory usage and business growth trends of applications. Allocate more memory resources to memory-intensive applications to ensure that they have sufficient memory available during operation.

[0126] Regularly scan memory usage, reclaim unused memory resources, and reallocate them to applications that need them, improving memory utilization.

[0127] For example, a big data analysis platform consumes a lot of memory when processing large amounts of data. The administrator allocates sufficient memory resources to the platform in advance based on the prediction model and sets up a memory recycling mechanism to ensure efficient use of memory. Use the prediction model to predict future network traffic, especially traffic during peak hours and emergencies. Set higher bandwidth priorities for critical applications to ensure that they can still obtain sufficient bandwidth resources when the network is congested. Dynamically adjust network bandwidth allocation based on real-time traffic data and prediction results to avoid bandwidth waste and bottlenecks. For example, during a large online event, it is predicted that the number of website visits will increase significantly and the demand for network bandwidth will surge. The administrator increases the allocation of network bandwidth in advance and sets high priorities for critical businesses to ensure unimpeded network during the event.

[0128] Use the data center management system to monitor key indicators such as processor usage, memory usage, and network bandwidth consumption in real time. Set up an alarm mechanism to promptly notify the administrator to handle when an indicator exceeds the preset threshold. Dynamically adjust resource allocation based on real-time monitoring data and forecast results. For example, when it is detected that the actual resource demand of an application exceeds the forecast value, immediately increase the resources allocated to the application to ensure its normal operation. Use automated scripts to quickly adjust resources and improve the efficiency and accuracy of resource allocation. Collect user feedback on the effectiveness of resource allocation through user surveys, system logs, and other methods.

[0129] Regularly analyze user feedback to identify problems and improvements in resource allocation. Based on user feedback and real-time monitoring data, continuously optimize the prediction model and dynamic allocation strategy. For example, adjust the parameters of the prediction model, improve the load balancing algorithm, etc. Regularly evaluate and plan data center resources to ensure that resources can meet the needs of future business development. At the same time, adjust the resource allocation strategy in a timely manner according to business changes to improve resource utilization and business response speed.

[0130] In a preferred embodiment of the present invention, the above step 15, according to the dynamic allocation and optimization of the data center resources, automatically performs the adjustment of the corresponding resource configuration, and continuously monitors the operation status of the data center to obtain new load data and business demand data in real time, may include:

[0131] Step 155, receiving the final prediction result and automatically triggering the resource configuration adjustment process;

[0132] Step 156, determining a corresponding resource configuration adjustment plan according to the prediction result, including increasing or decreasing server resources, adjusting network bandwidth configuration, or reallocating storage resources;

[0133] Step 157, sending the resource configuration adjustment plan to the management system of the data center, and performing the resource configuration adjustment operation;

[0134] Step 158, continuously monitor the operating status of the data center, including key indicators such as processor usage, memory occupancy, network bandwidth utilization, storage device usage, etc., to obtain new load data and business demand data in real time, including real-time traffic, user request volume, and transaction processing volume.

[0135] In an embodiment of the present invention, according to the prediction results, the system intelligently determines the corresponding resource configuration adjustment plan. These plans include increasing or decreasing server resources, adjusting network bandwidth configuration, reallocating storage resources, etc. When determining the plan, the system will consider multiple factors, such as cost-effectiveness, resource utilization, business continuity, etc., to ensure the rationality and feasibility of the plan. The system sends the resource configuration adjustment plan to the management system of the data center. After receiving the instruction, the data center management system will automatically perform the resource configuration adjustment operation. These operations involve the start and stop of the server, the reallocation of network bandwidth, the division of storage resources, etc. Continuously monitor the operating status of the data center, including key indicators such as processor usage, memory occupancy, network bandwidth utilization, and storage device usage. These indicators are obtained in real time through the data center management system. The collected monitoring data will be analyzed by the system in real time to identify any anomalies or trends. These data can also be used to update the prediction model to improve the accuracy of future predictions. The system dynamically adjusts the resource configuration strategy based on the monitoring results and business demand data.

[0136] By dynamically allocating and optimizing data center resources, the system ensures that resources are fully utilized when needed and avoids waste. This helps reduce data center operating costs and improves overall efficiency. The system automatically monitors and adjusts resource configuration to ensure that the data center maintains stable operation when business needs change. This helps reduce the risk of business interruption caused by insufficient or overloaded resources. By adjusting resource allocation in real time, the system ensures that the data center always provides sufficient processing power, network bandwidth, and storage resources to meet user needs. This helps improve user experience and satisfaction. The automated resource configuration adjustment process simplifies data center management. The system automatically completes resource allocation and adjustment tasks without manual intervention by administrators. This helps reduce management costs and improve management efficiency. The system can dynamically adjust resource configuration according to changes in business needs, providing strong support for business growth.

[0137] When specifically applied, the above content specifically includes:

[0138] The system conducts in-depth analysis of the forecast results to understand the demand trend of data center resources in the future. The forecast results include expected changes in key indicators such as processor utilization, memory occupancy, network bandwidth utilization, and storage device usage. Based on the forecast results, the system will intelligently determine the corresponding resource configuration adjustment plan. The plan will determine the number of server resources that need to be increased or decreased, the adjustment range of network bandwidth, and the reallocation method of storage resources. When determining the plan, the system will comprehensively consider various factors, such as cost-effectiveness (ensuring that the adjustment plan is economically reasonable), resource utilization (improving resource utilization efficiency and avoiding waste), business continuity (ensuring that the adjustment process will not affect business operations), etc. The system sends the resource configuration adjustment plan to the data center management system. After receiving the instruction, the data center management system will automatically perform resource configuration adjustment operations. These operations involve starting and stopping servers (increasing or reducing the number of servers according to demand), reallocation of network bandwidth (adjusting the allocation ratio of network bandwidth to meet the needs of different businesses), and partitioning of storage resources (repartitioning or expanding storage devices). The system will continuously monitor the operating status of the data center, including key indicators such as processor utilization, memory occupancy, network bandwidth utilization, and storage device usage.

[0139] The collected monitoring data will be analyzed by the system in real time to identify any anomalies or trends. If anomalies or trends are found, the system will issue an alarm in time and dynamically adjust the resource allocation strategy based on the monitoring results and business demand data.

[0140] Suppose a data center predicts that due to the launch of new services in the next week, the demand for network bandwidth will increase significantly, while the demand for storage resources will decrease. Based on this forecast result, the system determines the following resource allocation adjustment plan: Increase the allocation ratio of network bandwidth to ensure that the new service can run smoothly. The specific operation is: reallocate part of the network bandwidth originally allocated to other services to the new service that is about to be launched to meet its high bandwidth demand. Repartition the storage device to release some unused storage space for other business needs. At the same time, optimize storage resources to improve storage efficiency.

[0141] After the adjustment operation is performed, the operating status of the data center is continuously monitored. If abnormal fluctuations in network bandwidth or storage resources are found, the system will immediately analyze and dynamically adjust the resource allocation strategy based on the actual situation. For example, if the actual network bandwidth used by the new business is lower than the predicted value, the system will reallocate part of the network bandwidth to other businesses; if storage resources are tight, the system will consider adding storage devices or optimizing storage strategies.

[0142] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for automatically configuring data center resources, characterized in that: The method comprises: Step 11: monitor the load data of the data center in real time and obtain the real-time usage status of each resource, including processor usage, memory occupancy, network bandwidth utilization, and storage device usage; Step 12: Based on the load data of the data center, collect and analyze the demand data of the currently running services and the services to be launched in real time, including the service type, status, resource consumption, expected launch time and expected resource demand, to obtain the service analysis results; Step 13: Based on the business analysis results, the real-time monitored load data and business demand data are input into the resource allocation prediction and optimization model for training and learning, and the resource demand in the future is predicted to obtain the final prediction result; Step 14: dynamically allocate and optimize data center resources based on the final prediction results, including adjusting the number of processor cores, memory allocation, and network bandwidth allocation; Step 15, according to the dynamic allocation and optimization of data center resources, automatically perform the adjustment of corresponding resource configuration, and continuously monitor the operation status of the data center to obtain new load data and business demand data in real time; In step 12, based on the load data of the data center, the demand data of the currently running services and the services to be launched are collected and analyzed in real time, including the service type, status, resource consumption, expected launch time and expected resource demand, to obtain the service analysis results, including: Step 122, integrating the real-time business data and the planned data of the business to be launched into a unified data platform to obtain an integrated data set; Step 123, cleaning and preprocessing the integrated data set, extracting key features, to obtain a processed data set; Step 124, analyzing the resource usage pattern of the currently running business based on the processed data set, identifying the business types with large resource consumption or fast growth, and marking them as key businesses; Step 125, based on the marked key businesses and business rules, expected revenues and customer influencing factors, priority assessment is performed on the businesses to be launched, and a business priority list is output; Step 126, analyzing the impact of different priority services on resources according to the service priority list and resource usage pattern to generate a service resource usage analysis report; Step 127, obtaining a business analysis result according to the business resource usage analysis report; In step 125, priority evaluation is performed on the services to be launched based on the marked key services and business rules, expected revenues, and customer influencing factors, and a service priority list is output, including: Step 1255, setting evaluation criteria and metrics, including the urgency of the business; Step 1256, for each business to be launched, collect relevant information, including business requirement documents and customer feedback, and use the relevant information as input data for evaluation criteria; Step 1257, marking the business types identified as those with large resource consumption or rapid growth as key business information, and scoring each business to be launched based on the set evaluation criteria and key business information; Step 1258, during the scoring process, assign a corresponding score according to the degree of match between the business and each evaluation criterion; Step 1259, summing up the scores of each service to obtain a comprehensive score, and sorting the services according to the comprehensive score, from high to low, to generate a service priority list; In step 13, based on the business analysis results, the real-time monitored load data and business demand data are input into the resource allocation prediction and optimization model for training and learning, and the resource demand in the future is predicted to obtain the final prediction results, including: Step 133, determining features related to resource demand forecasting based on the business analysis results, including historical load data, business growth trends, and business types; Step 134, constructing a resource allocation prediction and optimization model, and determining the input items of the optimization model, including real-time monitored load data, business demand data, and output items, including resource demand forecast values ​​for a period of time in the future; Step 135, divide the historical load data and business demand data into a training set, a validation set and a test set, and use the training set data to train the optimization model. During the training process, the root mean square error is used to evaluate the prediction performance of the optimization model on the validation set to obtain an evaluation result; the calculation formula of the root mean square error is: Where R represents the root mean square error; n represents the total number of samples in the validation set; y v represents the actual value of the vth sample; represents the predicted value of the vth sample; w v represents the weight of the vth sample; λ represents the regularization coefficient; m represents the number of model parameters; θ j represents the jth model parameter; Step 136, iteratively optimizing the optimization model according to the evaluation results, including adjusting parameters; Step 137, repeating the training, verification and evaluation until the performance of the optimization model reaches a preset standard to obtain a final optimization model; Step 138, according to the final optimization model, the current real-time monitored load data and business demand data are used as input items to calculate the resource demand forecast value in the future period to obtain the final forecast result, including the resource demand forecast value, the estimated usage and growth trend of each type of resource; The calculation formula for the forecast value of resource demand in the future is: Among them, R t+k represents the predicted value of resource demand at the future time t+k; β0 represents the intercept term of the model; i represents the index of the influencing factor; β h represents the weight of the hth influencing factor; X h,t represents the observed value of the g-th influencing factor at time t; ∈ represents the deviation between the predicted value and the actual value.

2. A method for automatic configuration of data center resources according to claim 1, characterized in that: Step 15, according to the dynamic allocation and optimization of data center resources, automatically perform the adjustment of corresponding resource configuration, and continuously monitor the operation status of the data center to obtain new load data and business demand data in real time, including: Step 155, receiving the final prediction result and automatically triggering the resource configuration adjustment process; Step 156, determining a corresponding resource configuration adjustment plan according to the prediction result, including increasing or decreasing server resources, adjusting network bandwidth configuration, or reallocating storage resources; Step 157, sending the resource configuration adjustment plan to the management system of the data center, and performing the resource configuration adjustment operation; Step 158, continuously monitor the operating status of the data center, including key indicators of processor usage, memory occupancy, network bandwidth utilization, and storage device usage, to obtain new load data and business demand data in real time, including real-time traffic, user request volume, and transaction processing volume.

Images