Operation and maintenance management system and method based on artificial intelligence

Abstract

The invention relates to the field of operation and maintenance management, in particular to an operation and maintenance management system and method based on artificial intelligence, and the system comprises an operation baseline modeling module, a resource consumption prediction module, an automatic resource allocation module, a data correlation analysis module, an abnormal data detection module and a fault path analysis module. According to the method, accurate data support is provided for resource management through real-time monitoring of operation data and baseline performance modeling, resource demands are analyzed and predicted in combination with a time sequence, shortage and excess in resource configuration are identified, automatic allocation of resources is achieved, optimal matching between resource supply and demands is ensured, and the optimal resource allocation efficiency is improved through construction of a dependency graph. The interaction between the data stream and the application is analyzed, abnormal data detection and fault path analysis are helped, fault problems are found in real time, reasons are identified, the downtime of the system is shortened, the continuity of services and the integrity of data are ensured, and the automation degree and the intelligent level of operation and maintenance management are improved.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance management, and in particular to an operation and maintenance management system and method based on artificial intelligence. Background Art

[0002] The field of operation and maintenance management technology focuses on maintaining and managing the normal operation of computer hardware and software systems, including network facilities, servers, databases, and various infrastructure components, ensuring the efficiency, stability, and security of various technical resources in providing services, and using monitoring, automation tools, troubleshooting, and continuous maintenance strategies, including configuration management, performance monitoring, automated testing, and fault recovery plans to prevent service interruptions and performance degradation. By integrating artificial intelligence and machine learning technologies, predictive maintenance and automated response can be achieved to improve the efficiency of operation and maintenance and reduce the burden on personnel.

[0003] Among them, the operation and maintenance management system focuses on assisting technical personnel in monitoring and maintaining the health of IT infrastructure, automating routine operation and maintenance tasks, including performance monitoring, fault detection, problem diagnosis and automatic repair, improving system stability and reliability, preventing potential technical problems, reducing system downtime, optimizing resource utilization, and achieving reduced operating costs and improved service quality. It usually integrates real-time data analysis, alarm systems, log management and configuration updates to help management formulate strategies and optimize operation and maintenance processes, improve system reliability, security and efficiency, reduce operating costs and improve service quality.

[0004] Traditional operation and maintenance management technology relies on preset rules and manual intervention, lacks sufficient flexibility and adaptability, and resource management cannot dynamically adapt to sudden changes in business needs. In cloud computing and big data, it leads to insufficient or excessive resource allocation, affecting service quality and user experience. It is inefficient when processing complex data relationships and performing fault analysis, and lacks the ability to fully understand system status and predict problems, which limits its application effect in rapidly developing IT environments. The lack of highly integrated data analysis and resource management tools leads to longer response times for the operation and maintenance team when dealing with equipment failures and performance issues, and increased problem-solving cycles. When performance degradation occurs in the data center, it takes a long time to diagnose the root cause of the problem, increasing the risk of business interruption. Summary of the invention

[0005] The present invention aims at the technical problems existing in the prior art and provides a

[0006] The technical solution of the present invention to solve the above technical problems is as follows: an operation and maintenance management system based on artificial intelligence: The operation baseline modeling module monitors the performance indicators of servers and network devices in real time based on server operation data, extracts key performance parameters, identifies the performance baselines of multiple devices, and generates a baseline performance model; The resource consumption prediction module calculates resource usage trends and generates resource usage trend information by performing time series analysis on resource usage data based on the baseline performance model; The resource automatic allocation module evaluates the matching degree between the current resource configuration and the demand in real time according to the resource usage trend information, adjusts the computing resource allocation, and generates the resource configuration optimization result; The data correlation analysis module uses the resource configuration optimization results to analyze the data flows and dependencies between multiple devices and applications, identify the mutual influence between key components, create a relationship chain for the interaction between multiple devices and applications, and generate a dependency graph; The abnormal data detection module uses the dependency graph to detect abnormal fluctuations in the data stream by comparing the deviation between the real-time data and the baseline performance, identifies and records abnormal data points, and generates abnormal activity detection records; The fault path analysis module analyzes the abnormal data points based on the abnormal activity detection records and combines the equipment fault data to identify the abnormal source and propagation path, matches the warning information and sends it to the management personnel, and generates a fault cause identification record. As a further solution of the present invention, the baseline performance model includes CPU usage records, memory consumption patterns and network bandwidth statistics, the resource usage trend information includes resource demand forecast graphs, expected resource shortage time points and resource surplus periods, the resource configuration optimization results include CPU allocation configuration, storage space configuration and network bandwidth optimization records, the dependency graph includes a data flow direction graph between devices, a call relationship graph between applications and key component mutual influence analysis results, the abnormal activity detection record includes a list of abnormal data points, abnormal fluctuation time mark information and abnormal impact range assessment results, and the fault cause identification record includes a list of devices where the fault occurred, fault propagation path analysis results and fault impact degree evaluation information.

[0007] As a further solution of the present invention, the operating baseline modeling module includes: The performance data collection submodule monitors the performance indicators of multiple servers and network devices in real time based on server operation data, including CPU usage, memory consumption, and network bandwidth utilization, extracts key performance parameters, and generates key performance indicator data; The device performance analysis submodule records the performance baselines of multiple devices based on the key performance indicator data and generates a performance baseline identification record by identifying the performance thresholds and behavior patterns of multiple devices under normal operating conditions; The baseline model building submodule builds a baseline performance model based on the performance baseline identification record and utilizes a variety of key performance parameters to reflect the operating status of the equipment in real time and generate a baseline performance model.

[0008] As a further solution of the present invention, the resource consumption prediction module includes: The resource data extraction submodule extracts the usage data of various resources in real time based on the baseline performance model, including the real-time resource allocation status and actual usage, and generates resource usage analysis information; The demand trend analysis submodule calculates resource usage trends based on the resource usage analysis information and uses time series analysis to evaluate resource requirements at multiple time points, identify resource shortages and surpluses, and generate resource demand evaluation results; The resource trend prediction submodule predicts the changing trend of resource utilization rate based on the resource demand assessment result, records the peak and valley time periods, and generates resource utilization trend information.

[0009] As a further solution of the present invention, the automatic resource allocation module includes: The matching evaluation submodule evaluates the matching degree between the current resource configuration and the actual demand in real time based on the resource usage trend information and the allocation status of CPU, storage usage and network bandwidth, and generates resource matching evaluation information; The allocation demand analysis submodule identifies resource shortage and surplus situations based on the resource matching evaluation information, analyzes the computing resources that need to be adjusted, and generates adjustment demand calculation results; The configuration real-time adjustment submodule executes the adjustment demand calculation result, adjusts CPU allocation, storage usage and network bandwidth configuration, and evaluates the matching degree between the adjusted resource configuration and the predicted business demand, and generates a resource configuration optimization result.

[0010] As a further solution of the present invention, the data association analysis module includes: The data flow analysis submodule monitors and analyzes the data interaction and flow between multiple devices and applications in real time based on the resource configuration optimization result, records the frequency and amount of data transmission, and generates data flow monitoring information; The dependency analysis submodule analyzes and identifies the interdependencies between various key components based on the data flow monitoring information, and generates an interaction relationship data set; The relationship graph construction submodule analyzes the interactions and impact paths between devices based on the interaction relationship analysis results, creates a relationship chain for the interaction relationship between devices and applications, and generates a dependency graph.

[0011] As a further solution of the present invention, the abnormal data detection module includes: The data residual calculation submodule calculates the deviation between the real-time performance data and the baseline performance data based on the dependency graph, including CPU usage, memory consumption, and network bandwidth utilization, and generates a data deviation calculation result; The abnormal threshold setting submodule sets the abnormal detection threshold and evaluates the data deviation based on the data deviation calculation result to generate an abnormal threshold comparison result; The abnormal data recording submodule records the occurrence time, equipment, performance indicator type and numerical information of multiple abnormal data points based on the abnormal threshold comparison result, and generates an abnormal activity detection record.

[0012] As a further solution of the present invention, the fault path analysis module includes: The fault data comparison submodule analyzes the causes of abnormal data points based on the abnormal activity detection record and generates a fault analysis result by comparing the abnormal activity detection record with a variety of known equipment fault data. The fault path tracing submodule analyzes the fault propagation path based on the fault analysis results and the mutual influence relationship between multiple key components, identifies the key components and connection points involved, and generates fault propagation path information; The warning information sending submodule matches the warning information based on the fault propagation path information, the fault source, the impact range and the fault time, and sends it to the management personnel to generate a fault cause identification record.

[0013] An operation and maintenance management method based on artificial intelligence, characterized in that it is executed according to the operation and maintenance management system based on artificial intelligence according to any one of claims 1 to 8, comprising the following steps: S1: Based on server operation data, it monitors multiple performance indicators in real time, identifies the performance baselines of multiple devices, combines time series analysis, calculates resource usage trends, predicts resource requirements, and generates resource usage trend information; S2: using the resource usage trend information, evaluating the matching degree between the current resource configuration and the resource demand, adjusting and optimizing the allocation parameters of the CPU, memory, and network bandwidth, and generating a resource configuration optimization result; S3: using the resource configuration optimization result, analyzing the data flow and dependency relationship between devices and applications, constructing the interaction relationship chain and impact path between devices, and generating a dependency relationship map; S4: By comparing the deviation of real-time data with baseline performance through the dependency graph, abnormal fluctuations in the data stream are detected in real time, abnormal data points are recorded, and abnormal activity detection records are generated; S5: Based on the abnormal activity detection record, combined with the equipment failure data, analyze the source and propagation path of the abnormal data point, identify the cause of the failure, match the warning information and send it to the management personnel, and generate a failure cause identification record.

[0014] The beneficial effects of the present invention are: through real-time monitoring of operation data and baseline performance modeling, accurate data support is provided for resource management, resource demand is predicted in combination with time series analysis, shortages and surpluses in resource allocation are identified, and automatic allocation of resources is realized to ensure the optimal match between resource supply and demand; through the construction of dependency graphs, the interaction between data flows and applications is analyzed to assist in abnormal data detection and fault path analysis, fault problems are discovered in real time and their causes are identified, system downtime is reduced, business continuity and data integrity are ensured, and the automation and intelligence level of operation and maintenance management is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a system flow chart of the present invention; Figure 2 It is a schematic diagram of the system framework of the present invention; Figure 3 It is a flowchart of the operation baseline modeling module of the present invention; Figure 4 It is a flow chart of the resource consumption prediction module of the present invention; Figure 5 This is a flow chart of the automatic resource allocation module of the present invention; Figure 6 It is a flow chart of the data association analysis module of the present invention; Figure 7 This is a flow chart of the abnormal data detection module of the present invention; Figure 8 This is a flow chart of the fault path analysis module of the present invention; Fig. 9 It is a schematic diagram of the method steps of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only 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 those skilled in the art without creative work are within the scope of protection of this application.

[0017] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0018] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0019] Example 1 See also Figure 1 to Figure 2 The present invention provides a technical solution, an operation and maintenance management system based on artificial intelligence includes: The operation baseline modeling module monitors the performance indicators of servers and network devices in real time based on server operation data, extracts key performance parameters, identifies the performance baselines of multiple devices, and generates a baseline performance model; The resource consumption prediction module is based on the baseline performance model. It calculates the resource usage trend and generates resource usage trend information by performing time series analysis on resource usage data. The automatic resource allocation module evaluates the matching degree between current resource configuration and demand in real time based on resource usage trend information, adjusts computing resource allocation, and generates resource configuration optimization results; The data correlation analysis module uses the resource configuration optimization results to analyze the data flow and dependencies between multiple devices and applications, identify the mutual influence between key components, create a relationship chain for the interaction between multiple devices and applications, and generate a dependency graph; The abnormal data detection module uses the dependency graph to detect abnormal fluctuations in the data stream by comparing the deviation between real-time data and baseline performance, identifying and recording abnormal data points, and generating abnormal activity detection records; The fault path analysis module analyzes abnormal data points based on abnormal activity detection records and equipment fault data, identifies the source and propagation path of the abnormality, matches the warning information and sends it to the management personnel, and generates a fault cause identification record. The baseline performance model includes CPU usage records, memory consumption patterns and network bandwidth statistics. Resource usage trend information includes resource demand forecast graphs, expected resource shortage time points and resource surplus periods. Resource configuration optimization results include CPU allocation configuration, storage space configuration and network bandwidth optimization records. Dependency graphs include data flow direction graphs between devices, call relationship graphs between applications and key component mutual influence analysis results. Abnormal activity detection records include a list of abnormal data points, abnormal fluctuation time mark information and abnormal impact range assessment results. Fault cause identification records include a list of devices where faults occurred, fault propagation path analysis results and fault impact degree evaluation information.

[0020] See also Figures 2 to 3 , running the baseline modeling module includes: The performance data collection submodule monitors the performance indicators of multiple servers and network devices in real time based on server operation data, including CPU usage, memory consumption, and network bandwidth utilization, extracts key performance parameters, and generates key performance indicator data; In the performance data collection submodule, based on the real-time data provided by the server and network equipment, data collection is used to continuously monitor and record the CPU usage, memory consumption and network bandwidth utilization of each device. After the target data is initially cleaned and formatted, each performance indicator is stored in a time series to ensure the timeliness and accuracy of the data. Timestamps are used in the data storage process to facilitate subsequent data tracking and historical comparison. By calculating the moving average and fluctuation range of each performance indicator, key parameters affecting system performance, such as average load and peak response time, are extracted to generate performance monitoring logs. The target log includes instantaneous performance values ​​and statistical summaries, providing a basic data source for performance analysis.

[0021] The equipment performance analysis submodule records the performance baselines of multiple devices based on key performance indicator data by identifying the performance thresholds and behavior patterns of multiple devices under normal operating conditions, and generates performance baseline identification records; In the equipment performance analysis submodule, based on the collected key performance indicator data, data analysis tools such as R language or Python's Pandas library are used to analyze the data, calculate the historical average and standard deviation of each performance indicator, determine the normal fluctuation range of performance, and analyze the variability of performance data through the statistical test ANOVA to determine whether there is a significant performance degradation or improvement. Cluster analysis methods such as K-means or DBSCAN are applied to identify the behavior patterns of equipment under specific performance parameters. The target pattern helps identify the performance threshold of the equipment under normal operating conditions. The performance baseline of each device is then recorded and labeled, including the model, configuration and corresponding performance threshold of the device, and a performance baseline identification record is generated. The target record details the expected performance level of each device under normal operating conditions.

[0022] The baseline model building submodule builds a baseline performance model based on performance baseline identification records and uses multiple key performance parameters to reflect the operating status of the equipment in real time and generate a baseline performance model; In the baseline model building submodule, based on the identified performance baseline records, predictive modeling techniques, such as linear regression or neural network models, are used to build a baseline performance model for real-time monitoring of device performance. The model combines multiple key performance parameters, such as CPU load, memory usage, and network traffic. During the model building process, each parameter is assigned an appropriate weight, and the target weight is obtained through regression analysis of historical performance data to ensure that the model can accurately reflect the performance of different devices under various workloads. Model training uses cross-validation methods to optimize parameter settings to avoid overfitting and ensure that the model has good generalization capabilities. The constructed baseline performance model can reflect the current operating status and health of the device in real time, providing instant performance monitoring support for operation and maintenance management.

[0023] See also Figures 2 to 4 ,The resource consumption prediction module includes: The resource data extraction submodule extracts the usage data of various resources in real time based on the baseline performance model, including the real-time resource allocation status and actual usage, and generates resource usage analysis information; In the resource data extraction submodule, according to the setting of the baseline performance model, the resource usage of servers and network devices is monitored and extracted in real time, including CPU allocation status, memory usage, and actual utilization of network bandwidth. High-frequency sampling technology is used in the data collection process to ensure the timeliness and accuracy of the data. Data is collected through a distributed monitoring system. The technologies used include SNMP protocol and remote monitoring services. The target technology supports the collection of performance data from different devices and platforms. The collected data is preliminarily cleaned and screened to exclude any abnormal or invalid data, and formatted to provide standardized input for subsequent analysis and generate resource usage analysis information. The target information records the resource usage status of each device in detail, providing a data basis for resource allocation optimization.

[0024] The demand trend analysis submodule uses time series analysis based on resource usage analysis information to calculate resource usage trends, evaluate resource demand at multiple time points, identify resource shortages and surpluses, and generate resource demand assessment results; In the demand trend analysis submodule, the collected resource usage analysis information is used to apply time series analysis technology to process the target data. The ARIMA model is used to predict and analyze the time series feature data. Through model identification and parameter adjustment, the seasonal changes and trends of resource usage data are analyzed, and the resource demand at each time point in the future is predicted. In the process, the mean, standard deviation and prediction error of resource usage are calculated, and resource shortages and surpluses are identified. The target situation is usually manifested as peaks and troughs in resource usage. The analysis results help system administrators understand the dynamic changes in resource usage and generate resource demand assessment results. The target results provide decision support for resource management and allocation.

[0025] The resource trend prediction submodule predicts the changing trend of resource utilization based on the resource demand assessment results, records the peak and valley periods, and generates resource utilization trend information; In the resource trend prediction submodule, based on the resource demand assessment results, the changing trend of resource utilization is predicted, and the decision tree regression model in machine learning technology is used for trend prediction. The decision tree model establishes decision rules based on historical resource usage data to predict possible future resource usage situations, including predictions for peak and valley periods. The target prediction results help operation and maintenance managers better plan resources to ensure a rapid response when resource demand increases and avoid waste or shortage of resources. The generated resource usage trend information describes in detail the predicted peak and valley periods of resource usage, providing a basis for the subsequent formulation of resource allocation strategies.

[0026] See also Figures 2 to 5 , the automatic resource allocation module includes: The matching evaluation submodule evaluates the matching degree between the current resource configuration and actual demand in real time based on the resource usage trend information, according to the allocation status of CPU, storage usage and network bandwidth, and generates resource matching evaluation information; In the matching evaluation submodule, the system uses dynamic resource management technology to monitor resource usage status in real time, including CPU occupancy, storage space usage, and network bandwidth consumption. Based on resource usage trend information, real-time data analysis is performed through decision support systems such as SAP HANA to calculate the deviation between current resource configuration and actual demand. The linear regression model is used to analyze the correlation between each resource item and business demand to ensure the accuracy and reliability of the calculation results. Through target analysis, the system evaluates the actual utilization efficiency and potential bottlenecks of resources and generates resource matching evaluation information. This information details the utilization status and optimization suggestions of each resource, helping administrators make more reasonable resource allocation decisions.

[0027] The allocation demand analysis submodule identifies resource shortages and surpluses based on resource matching assessment information, analyzes the computing resources that need to be adjusted, and generates adjustment demand calculation results; In the allocation demand analysis submodule, the generated resource matching evaluation information is used to analyze the deficiencies in resource allocation. Network flow analysis is used to visualize the resource flow and identify the flow direction and shortage nodes of key resources. Through Monte Carlo simulation, the system simulates the business impact under different resource allocation schemes and identifies shortages and surpluses in resource allocation. After the analysis is completed, the system automatically calculates the amount of resources that need to be optimized or increased based on business priorities and resource consumption rates, and generates adjustment demand calculation results. This result provides a scientific basis for resource allocation, ensuring that resource allocation is more accurate and efficient.

[0028] Configure the real-time adjustment submodule to perform the adjustment demand calculation results, adjust the CPU allocation, storage usage and network bandwidth configuration, and evaluate the matching degree between the adjusted resource configuration and the predicted business demand to generate the resource configuration optimization results; In the configuration real-time adjustment submodule, resource configuration adjustment is performed under the guidance of the adjustment demand calculation results. Optimization tools based on genetic algorithms, such as MATLAB's optimization toolbox, are used to dynamically adjust CPU allocation, storage space, and network bandwidth to meet changing business needs. During the adjustment process, the system monitors the effect of each configuration adjustment, uses real-time feedback to adjust algorithm parameters, and optimizes resource allocation strategies. After the adjustment is completed, the system re-evaluates the match between the adjusted resource configuration and the predicted business needs to ensure that the optimized resources can support future business expansion and peak demand, and generate resource configuration optimization results. This result directly affects the stability of business operations and the overall performance of the system.

[0029] See also Figures 2 to 6 , the data correlation analysis module includes: The data flow analysis submodule monitors and analyzes the data interaction and flow between multiple devices and applications in real time based on the resource configuration optimization results, records the frequency and amount of data transmission, and generates data flow monitoring information; In the data flow analysis submodule, network monitoring tools such as Wireshark and SolarWinds are deployed based on the resource configuration optimization results. The target tools can capture and analyze data packets transmitted through the network in real time, including the source address, destination, size and transmission time of the data packets. The system analyzes the captured data in real time, calculates the frequency and transmission volume of each data flow, and uses data mining techniques such as flow clustering algorithms to classify and mark data flows, identify high-traffic and low-traffic data paths, identify data hotspots and potential bottlenecks in the network, and generate data flow monitoring information. The target information contains key network traffic statistics and potential problems, providing a basis for network optimization and resource adjustment.

[0030] The dependency analysis submodule analyzes and identifies the interdependencies between various key components based on data flow monitoring information and generates an interaction relationship dataset; In the dependency analysis submodule, the system applies graph analysis tools such as GraphX ​​based on previously generated data flow monitoring information, executes graph algorithms including node degree analysis and path search, identifies and understands the dependencies between devices and applications, and builds a dependency network of key components during the analysis, including which components are highly interdependent and which components have the greatest impact on network performance. Through target analysis, the system is able to generate an interaction relationship dataset that records the dependency strength and interaction frequency between various devices and applications, providing key insights into system stability and optimization.

[0031] The relationship graph construction submodule analyzes the interactions and impact paths between devices based on the interaction relationship analysis results, creates a relationship chain for the interaction relationship between devices and applications, and generates a dependency graph; In the relationship graph construction submodule, the system uses the interaction relationship analysis results and network visualization tools such as Gephi to build a dependency graph. The graph shows the interaction network between devices and applications, identifying key data flows and impact paths. Through this visualization, managers can intuitively see the connection between components and their impact on the overall performance of the system, making it easier to quickly identify and resolve potential problem points. The generated dependency graph not only provides an intuitive image of the interaction between devices, but also includes critical paths that may affect service quality, allowing the operation and maintenance team to respond quickly when problems arise and optimize system configuration to meet future business needs.

[0032] See also Figures 2 to 7 , the abnormal data detection module includes: The data residual calculation submodule calculates the deviation between real-time performance data and baseline performance data based on the dependency graph, including CPU usage, memory consumption, and network bandwidth utilization, and generates data deviation calculation results; In the data residual calculation submodule, the interaction data between devices and applications obtained from the dependency graph is combined with the real-time monitoring performance data to calculate the deviation between key performance indicators such as CPU usage, memory consumption and network bandwidth utilization and the preset baseline performance data. The difference analysis method is used in the calculation process, including the difference calculation of real-time and historical data for each performance indicator, and the application of statistical methods such as standard deviation and mean difference to evaluate the degree of data deviation, generate detailed data deviation calculation results, and the target results accurately reflect the specific difference between the current system performance status and the expected performance, providing accurate data support for anomaly detection.

[0033] The anomaly threshold setting submodule sets the anomaly detection threshold and evaluates the data deviation based on the data deviation calculation results to generate anomaly threshold comparison results; In the anomaly threshold setting submodule, the system sets the anomaly detection threshold based on the data deviation calculation results. In the process, the deviation distribution of various performance indicators is analyzed, and the statistical decision theory is used to determine the anomaly detection threshold to ensure that the target threshold can cover most normal fluctuations and sensitively respond to possible abnormal conditions. The cumulative distribution function is used to set the threshold to ensure that the set threshold can maximize the accuracy and response speed of anomaly detection within a certain confidence interval. The generated anomaly threshold comparison result provides a threshold setting report, including the threshold of each indicator and the theoretical basis and expected effect of the target threshold setting.

[0034] The abnormal data recording submodule records the occurrence time, equipment, performance indicator type and numerical information of multiple abnormal data points based on the abnormal threshold comparison results, and generates abnormal activity detection records; In the abnormal data recording submodule, the system monitors and records all abnormal data points that exceed the threshold in real time based on the set abnormal threshold comparison results, including the occurrence time of each abnormal event, the affected equipment, the relevant performance indicator type and its specific value. Database management systems such as MySQL are used for data recording to ensure data integrity and query efficiency. Through detailed records of abnormal data, the system can provide a comprehensive abnormal activity detection record. The target record not only includes the specific information of the abnormal event, but also provides sufficient data to support subsequent fault analysis and system optimization measures, providing important decision-making support for the operation and maintenance team in fault response and prevention.

[0035] See also Figures 2 to 8, the fault path analysis module includes: The fault data comparison submodule is based on the abnormal activity detection records, and by comparing with a variety of known equipment fault data, it analyzes the causes of abnormal data points and generates fault analysis results; In the fault data comparison submodule, the system compares and analyzes the collected abnormal activity detection records with the historically recorded equipment fault database. The database includes a wide range of fault cases and their related performance data, so that the comparative analysis can accurately identify the potential causes of abnormal data points. The analysis process uses pattern recognition and machine learning classification algorithms, such as support vector machines and random forests. The target algorithm can accurately classify new abnormal situations based on historical data patterns, analyze the cause of each abnormal data point, such as hardware failure, configuration error, or external network attack, and generate fault analysis results. The target results provide an important basis for fault response and prevention.

[0036] The fault path tracing submodule analyzes the fault propagation path based on the fault analysis results and the mutual influence relationship between multiple key components, identifies the key components and connection points involved, and generates fault propagation path information; In the fault path tracing submodule, the system uses graph theory and network analysis methods, such as the shortest path algorithm, to track the fault propagation path based on the fault analysis results. The analysis process examines the mutual influence relationship between multiple key components in detail, and identifies the key nodes and connection points in the fault propagation process. The target nodes and connection points are usually the weak links or risk areas of the system, and a fault propagation path diagram is constructed. The diagram shows the starting point and diffusion route of the fault, revealing the fault diffusion mechanism and impact range. The generated fault propagation path information is of great value for formulating targeted maintenance strategies and preventive measures.

[0037] The warning information sending submodule matches the warning information based on the fault propagation path information, the fault source, the impact range and the fault time, and sends it to the management personnel to generate the fault cause identification record; In the early warning information sending submodule, the determined fault propagation path information is used, combined with the nature, impact scope and specific occurrence time of the fault source, to automatically match the relevant early warning information through the early warning system. In the process, an event-driven message publishing system, such as Apache Kafka, is used to ensure that the information is quickly and accurately conveyed to the responsible managers and technical teams. The system adjusts the priority and response measures of the message according to the severity and urgency of the fault to ensure that the management can make decisions and deploy countermeasures quickly. The generated fault cause identification record describes the full picture of the fault event in detail, including the fault point, possible causes and recommended response strategies, providing important support for reducing system downtime and reducing potential losses.

[0038] See also Fig. 9 , an operation and maintenance management method based on artificial intelligence, characterized in that it is executed according to the operation and maintenance management system based on artificial intelligence according to any one of claims 1-8, comprising the following steps: S1: Based on server operation data, it monitors multiple performance indicators in real time, identifies the performance baselines of multiple devices, combines time series analysis, calculates resource usage trends, predicts resource requirements, and generates resource usage trend information; S2: Use resource usage trend information to evaluate the matching degree between current resource configuration and resource demand, adjust and optimize the allocation parameters of CPU, memory, and network bandwidth, and generate resource configuration optimization results; S3: Use the resource configuration optimization results to analyze the data flow and dependencies between devices and applications, build the interaction relationship chain and impact path between devices, and generate a dependency graph; S4: Through the dependency graph, by comparing the deviation of real-time data with baseline performance, abnormal fluctuations in the data stream are detected in real time, abnormal data points are recorded, and abnormal activity detection records are generated; S5: Based on the abnormal activity detection records and combined with the equipment failure data, analyze the source and propagation path of the abnormal data points, identify the cause of the failure, match the warning information and send it to the management personnel, and generate the failure cause identification record.

[0039] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0040] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0041] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0042] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0043] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0044] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0045] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An operation and maintenance management system based on artificial intelligence, characterized in that: The system comprises: The operation baseline modeling module monitors the performance indicators of servers and network devices in real time based on server operation data, extracts key performance parameters, identifies the performance baselines of multiple devices, and generates a baseline performance model; The resource consumption prediction module calculates resource usage trends and generates resource usage trend information by performing time series analysis on resource usage data based on the baseline performance model; The resource automatic allocation module evaluates the matching degree between the current resource configuration and the demand in real time according to the resource usage trend information, adjusts the computing resource allocation, and generates the resource configuration optimization result; The data correlation analysis module uses the resource configuration optimization results to analyze the data flows and dependencies between multiple devices and applications, identify the mutual influence between key components, create a relationship chain for the interaction between multiple devices and applications, and generate a dependency graph; The abnormal data detection module uses the dependency graph to detect abnormal fluctuations in the data stream by comparing the deviation between the real-time data and the baseline performance, identifies and records abnormal data points, and generates abnormal activity detection records; The fault path analysis module analyzes the abnormal data points based on the abnormal activity detection records and combines the equipment fault data to identify the abnormal source and propagation path, matches the warning information and sends it to the management personnel, and generates a fault cause identification record.

2. The artificial intelligence-based operation and maintenance management system according to claim 1, characterized in that: The baseline performance model includes CPU usage records, memory consumption patterns and network bandwidth statistics; the resource usage trend information includes resource demand forecast graphs, expected resource shortage time points and resource surplus periods; the resource configuration optimization results include CPU allocation configuration, storage space configuration and network bandwidth optimization records; the dependency graph includes a data flow direction graph between devices, a call relationship graph between applications and key component mutual influence analysis results; the abnormal activity detection record includes a list of abnormal data points, abnormal fluctuation time mark information and abnormal impact range assessment results; the fault cause identification record includes a list of devices where the fault occurred, fault propagation path analysis results and fault impact degree evaluation information.

3. The artificial intelligence-based operation and maintenance management system according to claim 1, characterized in that: The running baseline modeling module includes: The performance data collection submodule monitors the performance indicators of multiple servers and network devices in real time based on server operation data, including CPU usage, memory consumption, and network bandwidth utilization, extracts key performance parameters, and generates key performance indicator data; The device performance analysis submodule records the performance baselines of multiple devices based on the key performance indicator data and generates a performance baseline identification record by identifying the performance thresholds and behavior patterns of multiple devices under normal operating conditions; The baseline model building submodule builds a baseline performance model based on the performance baseline identification record and utilizes a variety of key performance parameters to reflect the operating status of the equipment in real time and generate a baseline performance model.

4. The artificial intelligence-based operation and maintenance management system according to claim 1, characterized in that: The resource consumption prediction module includes: The resource data extraction submodule extracts the usage data of various resources in real time based on the baseline performance model, including the real-time resource allocation status and actual usage, and generates resource usage analysis information; The demand trend analysis submodule calculates resource usage trends based on the resource usage analysis information and uses time series analysis to evaluate resource requirements at multiple time points, identify resource shortages and surpluses, and generate resource demand evaluation results; The resource trend prediction submodule predicts the changing trend of resource utilization rate based on the resource demand assessment result, records the peak and valley time periods, and generates resource utilization trend information.

5. The artificial intelligence-based operation and maintenance management system according to claim 1, characterized in that: The automatic resource allocation module includes: The matching evaluation submodule evaluates the matching degree between the current resource configuration and the actual demand in real time based on the resource usage trend information and the allocation status of CPU, storage usage and network bandwidth, and generates resource matching evaluation information; The allocation demand analysis submodule identifies resource shortage and surplus situations based on the resource matching evaluation information, analyzes the computing resources that need to be adjusted, and generates adjustment demand calculation results; The configuration real-time adjustment submodule executes the adjustment demand calculation result, adjusts CPU allocation, storage usage and network bandwidth configuration, and evaluates the matching degree between the adjusted resource configuration and the predicted business demand, and generates a resource configuration optimization result.

6. The artificial intelligence-based operation and maintenance management system according to claim 1, characterized in that: The data relevance analysis module includes: The data flow analysis submodule monitors and analyzes the data interaction and flow between multiple devices and applications in real time based on the resource configuration optimization result, records the frequency and amount of data transmission, and generates data flow monitoring information; The dependency analysis submodule analyzes and identifies the interdependencies between various key components based on the data flow monitoring information, and generates an interaction relationship data set; The relationship graph construction submodule analyzes the interactions and impact paths between devices based on the interaction relationship analysis results, creates a relationship chain for the interaction relationship between devices and applications, and generates a dependency graph.

7. The artificial intelligence-based operation and maintenance management system according to claim 1, characterized in that: The abnormal data detection module comprises: The data residual calculation submodule calculates the deviation between the real-time performance data and the baseline performance data based on the dependency graph, including CPU usage, memory consumption, and network bandwidth utilization, and generates a data deviation calculation result; The abnormal threshold setting submodule sets the abnormal detection threshold and evaluates the data deviation based on the data deviation calculation result to generate an abnormal threshold comparison result; The abnormal data recording submodule records the occurrence time, equipment, performance indicator type and numerical information of multiple abnormal data points based on the abnormal threshold comparison result, and generates an abnormal activity detection record.

8. The artificial intelligence-based operation and maintenance management system according to claim 1, characterized in that: The fault path analysis module includes: The fault data comparison submodule analyzes the causes of abnormal data points based on the abnormal activity detection record and generates a fault analysis result by comparing the abnormal activity detection record with a variety of known equipment fault data. The fault path tracing submodule analyzes the fault propagation path based on the fault analysis results and the mutual influence relationship between multiple key components, identifies the key components and connection points involved, and generates fault propagation path information; The warning information sending submodule matches the warning information based on the fault propagation path information, the fault source, the impact range and the fault time, and sends it to the management personnel to generate a fault cause identification record.

9. An operation and maintenance management method based on artificial intelligence, characterized in that: The operation and maintenance management system based on artificial intelligence according to any one of claims 1 to 8 comprises the following steps: Based on server operation data, it monitors multiple performance indicators in real time, identifies the performance baselines of multiple devices, and combines time series analysis to calculate resource usage trends, predict resource requirements, and generate resource usage trend information; Using the resource usage trend information, evaluate the matching degree between current resource configuration and resource demand, adjust and optimize the allocation parameters of CPU, memory, and network bandwidth, and generate resource configuration optimization results; Using the resource configuration optimization results, analyzing the data flows and dependencies between devices and applications, building the interaction relationship chain and impact path between devices, and generating a dependency graph; By comparing the deviation of real-time data with baseline performance through the dependency graph, abnormal fluctuations in the data stream are detected in real time, abnormal data points are recorded, and abnormal activity detection records are generated; Based on the abnormal activity detection record, combined with the equipment failure data, the source and propagation path of the abnormal data point are analyzed, the cause of the failure is identified, the warning information is matched and sent to the management personnel, and the failure cause identification record is generated.

Images