Network performance monitoring method and system based on artificial intelligence
The network performance monitoring system addresses data incompleteness and inefficiency by employing eBPF and AI-driven predictive analysis, ensuring real-time, comprehensive data collection and efficient network recovery.
Patent Information
- Application Number
- CN202510399364.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-01
- Publication Date
- 2025-07-15
AI Technical Summary
In the prior art, network performance monitoring methods have problems such as incomplete data acquisition, low processing efficiency and insufficient predictability.
The network performance monitoring system based on artificial intelligence is adopted, including eBPF data acquisition module, microservice processing unit, predictive maintenance module, alarm generation module, GSM push module, automated maintenance module and alarm analysis module. The data is collected in real time using eBPF technology, preprocessing and feature extraction is performed through the microservice processing unit, predictive analysis is applied to apply machine learning algorithms, and alarm information is promptly pushed through the GSM module to automatically perform maintenance operations.
It realizes more comprehensive data acquisition, improves processing efficiency and predictability, can timely identify network performance problems and automatically restore network performance, reducing false alarms and missed reports.
Smart Images
Figure CN120321094A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence technology, and in particular, to a network performance monitoring method and system based on artificial intelligence. Background Art
[0002] In today's digital age, the network has become the core infrastructure for information exchange and data transmission. With the continuous expansion of enterprise business and the increasing richness of Internet applications, the stability and efficiency of network performance are crucial for ensuring business continuity and user experience. Network performance not only concerns the speed and quality of data transmission, but also directly affects the operational efficiency of enterprises, customer satisfaction, and market competitiveness. Network performance monitoring has thus become a key link to ensure the stable operation of the network.
[0003] However, in the prior art, the network performance monitoring methods used have problems such as incomplete data collection, low processing efficiency, and insufficient predictability. Summary of the Invention
[0004] The purpose of the present invention is to provide a network performance monitoring method and system based on artificial intelligence, aiming to solve the technical problems of incomplete data collection, low processing efficiency, and insufficient predictability existing in the network performance monitoring methods used in the prior art.
[0005] To achieve the above object, a network performance monitoring system based on artificial intelligence adopted by the present invention includes an eBPF data collection module, a microservices processing unit, a predictive maintenance module, a rule library, an alarm generation module, a GSM push module, a maintenance terminal, an automated maintenance module, and an alarm analysis module;
[0006] The microservices processing unit is connected to the eBPF data collection module, the predictive maintenance module is connected to the microservices processing unit, the alarm generation module is connected to both the predictive maintenance module and the rule library, the GSM push module is connected to the alarm generation module, and the GSM module is also connected to the maintenance terminal and the automated maintenance module. The alarm analysis module is connected to both the alarm generation module and the rule library;
[0007] The eBPF data collection module is used to capture network traffic, system call information, and other relevant network performance data in real time and transmit them to the microservices processing unit;
[0008] The microservices processing unit is used to preprocess the data captured by the eBPF data collection module and extract useful network performance indicators and features;
[0009] The predictive maintenance module performs predictive analysis on the data transmitted by the microservices processing unit based on machine learning algorithms;
[0010] The alarm generation module generates alarm information based on the prediction results of the predictive maintenance module and the rules and thresholds in the rule library, and sends it to the GSM push module, the automated maintenance module, and the alarm analysis module;
[0011] The GSM push module is used to push alarm information to the maintenance terminal through the GSM network or other communication methods;
[0012] The automated maintenance module is used to automatically execute predefined maintenance operations according to the alarm information to quickly restore network performance, and the alarm analysis module is used to deeply analyze the alarm information to identify repeated alarms, false alarms, and missed alarms.
[0013] Among them, the microservice processing unit includes a data preprocessing module, a feature extraction module, and a dynamic policy adjustment module. The data preprocessing module is connected to the eBPF data acquisition module, the feature extraction module is connected to the data preprocessing module, and the dynamic policy adjustment module is connected to the feature extraction module.
[0014] Among them, the data preprocessing module includes a denoising sub-module and a normalization sub-module. The denoising sub-module is connected to the eBPF data acquisition module, and the normalization sub-module is connected to both the denoising sub-module and the feature extraction module;
[0015] The denoising sub-module uses methods such as filters, principal component analysis (PCA), dictionary-based learning methods, non-local means (NLM) algorithms, wavelet transforms, or deep learning-based algorithms to implement denoising;
[0016] The normalization sub-module uses standardization (Z-score standardization) or Min-Max scaling for normalization.
[0017] Among them, the artificial intelligence-based network performance monitoring system further includes a data visualization module, which is implanted in the maintenance terminal;
[0018] The data visualization module is used to display network performance data in an intuitive chart form.
[0019] Among them, the artificial intelligence-based network performance monitoring system further includes a security monitoring module, which is connected to the microservice processing unit;
[0020] The security monitoring module uses anomaly detection algorithms (such as isolation forest, LOF) to monitor abnormal traffic in the network; uses deep learning algorithms (such as CNN, RNN) to identify and classify malicious behaviors.
[0021] Among them, the artificial intelligence-based network performance monitoring system further includes a configuration management module and a log management module. The configuration management module is connected to the rule base, and the log management module is connected to the microservice processing unit.
[0022] Among them, the artificial intelligence-based network performance monitoring system further includes a login module, and the login module is connected to the configuration management module.
[0023] The present invention also provides an artificial intelligence-based network performance monitoring method, which is applied to the artificial intelligence-based network performance monitoring system as described above.
[0024] It includes the following steps:
[0025] First, use the eBPF data acquisition module to collect network data in real time, and preprocess and extract features from the data through the microservice processing unit:
[0026] After that, the predictive maintenance module applies machine learning algorithms to perform network performance prediction analysis, and the alarm generation module generates alarms according to the analysis results and the rule base, and pushes them to the maintenance terminal and the automated maintenance module through the GSM module;
[0027] The automated maintenance module automatically executes predefined maintenance operations according to the alarm information to quickly restore network performance.
[0028] A network performance monitoring method and system based on artificial intelligence. In specific use, the eBPF data collection module uses eBPF (Extended Berkeley Packet Filter) technology to collect raw network performance data such as data packets, system calls, and network traffic in real time at the network level. The eBPF technology ensures the real-time and accuracy of data collection with its high efficiency and low overhead characteristics. Subsequently, the microservice processing unit preprocesses the data captured by the eBPF data collection module and extracts useful network performance indicators and features. The extracted feature data is sent to the predictive maintenance module, which applies machine learning algorithms (such as time series analysis, anomaly detection, etc.) to perform predictive analysis on network performance. By analyzing historical data and real-time data, the predictive maintenance module can identify potential problems in network performance and provide a basis for subsequent alarm generation and response. When the alarm generation module combines the predictive maintenance module to detect network performance anomalies or potential problems, it will generate alarm information according to the preset rule library. The alarm information is pushed to relevant personnel or systems in a timely manner through the GSM push module (Global System for Mobile Communications) or other communication methods so that corresponding measures can be taken. And the automatic maintenance module is used to automatically execute predefined maintenance operations according to the alarm information to quickly restore network performance. The alarm analysis module is used to deeply analyze the alarm information to identify duplicate alarms, false alarms, and missed alarms, thus solving the technical problems of incomplete data collection, low processing efficiency, and insufficient predictability in the network performance monitoring methods used in the prior art.
[0029] The present invention uses eBPF technology to collect network data in real time. The eBPF technology can capture various types of data at the kernel level, including but not limited to network data packets, system calls, process information, etc. This means that the system can obtain more comprehensive and fine-grained network performance data, providing a solid foundation for subsequent analysis and prediction.
[0030] The predictive maintenance module applies machine learning algorithms to perform predictive analysis on network performance. The machine learning algorithms can automatically learn the patterns of historical data and identify potential problems in network performance. When it is predicted that the network performance may decline or fail, the system will generate alarm information in a timely manner and push it to relevant personnel or systems so that measures can be taken in advance to avoid the occurrence of problems. Description of the Drawings
[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.
[0032] Figure 1 is the principle block diagram of the first embodiment of the present invention.
[0033] Figure 2 is the principle block diagram of the second embodiment of the present invention.
[0034] Figure 3 is the principle block diagram of the third embodiment of the present invention.
[0035] 101 - eBPF data acquisition module, 102 - microservice processing unit, 103 - predictive maintenance module, 104 - rule library, 105 - alarm generation module, 106 - GSM push module, 107 - maintenance terminal, 108 - automated maintenance module, 109 - alarm analysis module, 110 - data visualization module, 111 - data preprocessing module, 112 - feature extraction module, 113 - dynamic policy adjustment module, 114 - denoising sub - module, 115 - normalization sub - module, 201 - security monitoring module, 202 - configuration management module, 203 - log management module, 204 - login module, 301 - permission management module, 302 - data storage module, 303 - data compression module. Detailed Embodiment
[0036] The following will describe in detail the embodiments of the present invention. The examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention and should not be construed as a limitation to the present invention.
[0037] The first embodiment of the present application is as follows:
[0038] Please refer to Figure 1 , Figure 1 is the principle block diagram of the first embodiment of the present invention.
[0039] The present invention provides an artificial intelligence-based network performance monitoring system, including an eBPF data acquisition module 101, a microservices processing unit 102, a predictive maintenance module 103, a rule base 104, an alarm generation module 105, a GSM push module 106, a maintenance terminal 107, an automated maintenance module 108, an alarm analysis module 109, and a data visualization module 110; the microservices processing unit 102 includes a data preprocessing module 111, a feature extraction module 112, and a dynamic policy adjustment module 113, and the data preprocessing module 111 includes a denoising sub-module 114 and a normalization sub-module 115. The foregoing solution solves the technical problems of incomplete data acquisition, low processing efficiency, and insufficient predictability existing in the network performance monitoring methods used in the prior art.
[0040] For this specific embodiment, the eBPF data acquisition module 101 is used to capture network traffic, system call information, and other relevant network performance data in real time, and transmit them to the microservices processing unit 102;
[0041] The microservices processing unit 102 is used to preprocess the data captured by the eBPF data acquisition module 101 and extract useful network performance indicators and features;
[0042] The predictive maintenance module 103 performs predictive analysis on the data transmitted by the microservices processing unit 102 based on machine learning algorithms;
[0043] The alarm generation module 105 generates alarm information according to the prediction results of the predictive maintenance module 103 and the rules and thresholds in the rule base 104, and sends it to the GSM push module 106, the automated maintenance module 108, and the alarm analysis module 109;
[0044] The GSM push module 106 is used to push alarm information to the maintenance terminal 107 through the GSM network or other communication methods;
[0045] The automated maintenance module 108 is used to automatically execute predefined maintenance operations according to the alarm information to quickly restore network performance, and the alarm analysis module 109 is used to deeply analyze the alarm information to identify repeated alarms, false alarms, and missed alarms.
[0046] Among them, the microservice processing unit 102 is connected to the eBPF data acquisition module 101, the predictive maintenance module 103 is connected to the microservice processing unit 102, the alarm generation module 105 is connected to both the predictive maintenance module 103 and the rule library 104, the GSM push module 106 is connected to the alarm generation module 105, and the GSM module is also connected to the maintenance terminal 107 and the automated maintenance module 108. The alarm analysis module 109 is connected to both the alarm generation module 105 and the rule library 104. During specific use, the eBPF data acquisition module 101 uses eBPF (Extended Berkeley Packet Filter) technology to collect raw network performance data such as data packets, system calls, and network traffic in real time at the network layer. The eBPF technology ensures the real-time and accuracy of data acquisition with its high efficiency and low overhead characteristics. Subsequently, the microservice processing unit 102 preprocesses the data captured by the eBPF data acquisition module 101 and extracts useful network performance indicators and features. The extracted feature data is sent to the predictive maintenance module 103, which applies machine learning algorithms (such as time series analysis, anomaly detection, etc.) to perform predictive analysis on network performance. By analyzing historical data and real-time data, the predictive maintenance module 103 can identify potential problems in network performance and provide a basis for subsequent alarm generation and response. When the alarm generation module 105 combines with the predictive maintenance module 103 to detect network performance anomalies or potential problems, it generates alarm information according to the preset rule library 104. The alarm information is timely pushed to relevant personnel or systems through the GSM push module 106 (Global System for Mobile Communications) or other communication methods so that corresponding measures can be taken. And the automated maintenance module 108 is used to automatically execute predefined maintenance operations according to the alarm information to quickly restore network performance. The alarm analysis module 109 is used to deeply analyze the alarm information to identify repeated alarms, false alarms, and missed alarms, thus solving the technical problems of incomplete data collection, low processing efficiency, and insufficient predictability in the network performance monitoring methods used in the prior art.
[0047] Secondly, the data preprocessing module 111 is connected to the eBPF data acquisition module 101, the feature extraction module 112 is connected to the data preprocessing module 111, and the dynamic policy adjustment module is connected to the feature extraction module 112;
[0048] The data preprocessing module 111 processes the collected data such as denoising and cleaning, improving the data quality and reducing the computational amount of subsequent analysis. The feature extraction module 112 extracts key features using an efficient algorithm, further simplifying the data processing flow.
[0049] Meanwhile, the denoising sub-module 114 is connected to the eBPF data acquisition module 101. The normalization sub-module 115 is connected to both the denoising sub-module 114 and the feature extraction module 112. The denoising sub-module 114 implements denoising using methods such as filters, principal component analysis (PCA), dictionary-based learning methods, non-local means (NLM) algorithms, wavelet transforms, or deep learning-based algorithms.
[0050] The normalization sub-module 115 performs normalization using standardization (Z-score standardization) or Min-Max scaling.
[0051] In addition, the data visualization module 110 is implanted in the maintenance terminal 107.
[0052] The data visualization module 110 is used to display network performance data in an intuitive chart form.
[0053] When using a network performance monitoring system based on artificial intelligence in this embodiment, during specific use, the eBPF data collection module 101 utilizes eBPF (Extended Berkeley Packet Filter) technology to collect raw network performance data such as data packets, system calls, and network traffic in real time at the network level. The eBPF technology ensures the real-time and accuracy of data collection with its characteristics of high efficiency and low overhead. Subsequently, the microservice processing unit 102 preprocesses the data captured by the eBPF data collection module 101 and extracts useful network performance metrics and features. The extracted feature data is sent to the predictive maintenance module 103, which applies machine learning algorithms (such as time series analysis, anomaly detection, etc.) to perform predictive analysis on network performance. By analyzing historical data and real-time data, the predictive maintenance module 103 can identify potential problems in network performance and provide a basis for subsequent alarm generation and response. When the alarm generation module 105 combines with the predictive maintenance module 103 to detect network performance anomalies or potential problems, it will generate alarm information according to the preset rule library 104. The alarm information is pushed to relevant personnel or systems in a timely manner through the GSM push module 106 (Global System for Mobile Communications) or other communication methods so as to take corresponding measures. And the automated maintenance module 108 is used to automatically execute predefined maintenance operations according to the alarm information to quickly restore network performance. The alarm analysis module 109 is used to deeply analyze the alarm information to identify duplicate alarms, false alarms, and missed alarms, thereby solving the technical problems of incomplete data collection, low processing efficiency, and insufficient predictability existing in the network performance monitoring methods used in the prior art.
[0054] The second embodiment of this application is as follows:
[0055] Based on the first embodiment, please refer to Figure 2 , Figure 2 which is the principle block diagram of the second embodiment of the present invention.
[0056] The present invention provides a network performance monitoring system based on artificial intelligence, further including a security monitoring module 201, a configuration management module 202, a log management module 203, and a login module 204.
[0057] For this specific embodiment, the security monitoring module 201 is connected to the microservice processing unit 102. The security monitoring module 201 monitors abnormal traffic in the network using anomaly detection algorithms (such as Isolation Forest, LOF); and uses deep learning algorithms (such as CNN, RNN) to identify and classify malicious behaviors.
[0058] Among them, the configuration management module 202 is connected to the rule library 104, and the log management module 203 is connected to the microservice processing unit 102. By introducing the configuration management module 202, system administrators can flexibly configure monitoring parameters, policies, etc. according to actual needs, making the system more adaptable to different network environments and business requirements. This flexibility improves the applicability and scalability of the system. The log management module 203 details various log information during the system operation, including operation logs, error logs, performance logs, etc. These log information provide rich historical data for system operation and maintenance personnel, helping to quickly locate problems, troubleshoot faults, and enhancing the traceability and maintainability of the system.
[0059] Secondly, the login module 204 is connected to the configuration management module 202, and the login module 204 is used for management personnel to log in to the configuration management module 202 to manage the rule library 104.
[0060] Using an artificial intelligence-based network performance monitoring system according to this embodiment, by introducing the configuration management module 202, system administrators can flexibly configure monitoring parameters, policies, etc. according to actual needs, making the system more adaptable to different network environments and business requirements. This flexibility improves the applicability and scalability of the system. The log management module 203 details various log information during the system operation, including operation logs, error logs, performance logs, etc. These log information provide rich historical data for system operation and maintenance personnel, helping to quickly locate problems, troubleshoot faults, and enhancing the traceability and maintainability of the system.
[0061] The third embodiment of this application is as follows:
[0062] Based on the second embodiment, please refer to Figure 3 , Figure 3 which is the principle block diagram of the third embodiment of the present invention.
[0063] The present invention provides an artificial intelligence-based network performance monitoring system, further including a permission management module 301, a data storage module 302, and a data compression module 303.
[0064] For this specific embodiment, the permission management module 301 is connected to the login module 204, and the permission management module 301 realizes strict control over user identities and permissions. Only authenticated users can log in to the system, and according to the user's permission level, their access scope to system functions and data is restricted. This mechanism effectively prevents unauthorized access and potential security threats, enhancing the security of the system.
[0065] Among them, the data storage module 302 is connected to the log management module 203, the data compression module 303 is connected to the data storage module 302, and the data storage module 302 adopts a distributed storage system, which can easily cope with the rapid growth of data volume. With the expansion of the network scale and the accumulation of monitoring data, the system can flexibly add storage nodes to expand the storage capacity and performance without large-scale transformation of the existing system. The compression module is used to compress the data in the data storage module 302 to have more storage space.
[0066] Using a network performance monitoring system based on artificial intelligence in this embodiment, the data storage module 302 adopts a distributed storage system, which can easily cope with the rapid growth of data volume. With the expansion of the network scale and the accumulation of monitoring data, the system can flexibly add storage nodes to expand the storage capacity and performance without large-scale transformation of the existing system. The compression module is used to compress the data in the data storage module 302 to have more storage space.
[0067] The present invention uses the eBPF technology to collect network data in real time. The eBPF technology can capture various types of data at the kernel level, including but not limited to network data packets, system calls, process information, etc. This means that the system can obtain more comprehensive and fine-grained network performance data, providing a solid foundation for subsequent analysis and prediction.
[0068] The present invention processes the collected data through the data preprocessing module 111 for denoising, cleaning, etc., improving the data quality and reducing the computational load of subsequent analysis. At the same time, the feature extraction module 112 extracts key features using efficient algorithms, further simplifying the data processing flow. In addition, the predictive maintenance module 103 applies machine learning algorithms to automatically analyze and predict network performance, greatly improving the processing efficiency.
[0069] The predictive maintenance module 103 applies machine learning algorithms to perform predictive analysis on network performance. Machine learning algorithms can automatically learn the patterns of historical data and identify potential problems in network performance. When it is predicted that the network performance may decline or fail, the system will generate warning information in a timely manner and push it to relevant personnel or systems so as to take measures in advance to avoid problems.
[0070] The above-disclosed is only a preferred embodiment of the present invention. Of course, the scope of rights of the present invention cannot be limited thereby. Those of ordinary skill in the art can understand all or part of the processes of implementing the above embodiments, and the equivalent changes made according to the claims of the present invention still fall within the scope covered by the present invention.
Claims
1. An artificial intelligence-based network performance monitoring system, characterized in that it includes an eBPF data collection module, a microservices processing unit, a predictive maintenance module, a rule library, an alarm generation module, a GSM push module, a maintenance terminal, an automated maintenance module, and an alarm analysis module; the microservices processing unit is connected to the eBPF data collection module, the predictive maintenance module is connected to the microservices processing unit, the alarm generation module is connected to both the predictive maintenance module and the rule library, the GSM push module is connected to the alarm generation module, and the GSM module is also connected to the maintenance terminal and the automated maintenance module, and the alarm analysis module is connected to both the alarm generation module and the rule library; the eBPF data collection module is used to capture network traffic, system call information, and other relevant network performance data in real time and transmit them to the microservices processing unit; the microservices processing unit is used to preprocess the data captured by the eBPF data collection module and extract useful network performance indicators and features; the predictive maintenance module performs predictive analysis on the data transmitted by the microservices processing unit based on machine learning algorithms; the alarm generation module generates alarm information according to the prediction results of the predictive maintenance module and the rules and thresholds in the rule library, and sends it to the GSM push module, the automated maintenance module, and the alarm analysis module; the GSM push module is used to push alarm information to the maintenance terminal through the GSM network or other communication methods; the automated maintenance module is used to automatically execute predefined maintenance operations according to the alarm information to quickly restore network performance, and the alarm analysis module is used to deeply analyze the alarm information to identify repeated alarms, false alarms, and missed alarms.
2. The artificial intelligence-based network performance monitoring system according to claim 1, characterized in that the microservices processing unit includes a data preprocessing module, a feature extraction module, and a dynamic policy adjustment module. The data preprocessing module is connected to the eBPF data collection module, the feature extraction module is connected to the data preprocessing module, and the dynamic policy adjustment module is connected to the feature extraction module.
3. The artificial intelligence-based network performance monitoring system according to claim 2, characterized in that the data preprocessing module includes a denoising sub-module and a normalization sub-module. The denoising sub-module is connected to the eBPF data collection module, and the normalization sub-module is connected to both the denoising sub-module and the feature extraction module; the denoising sub-module uses methods such as filters, principal component analysis (PCA), dictionary-based learning methods, non-local means (NLM) algorithms, wavelet transforms, or deep learning-based algorithms to achieve denoising; the normalization sub-module uses standardization (Z-score standardization) or Min-Max scaling methods for normalization.
4. The artificial intelligence-based network performance monitoring system according to claim 3, characterized in that The artificial intelligence-based network performance monitoring system further includes a data visualization module, which is implanted in the maintenance terminal; The data visualization module is used to display network performance data in an intuitive chart form.
5. The artificial intelligence-based network performance monitoring system according to claim 4, wherein The artificial intelligence-based network performance monitoring system further includes a security monitoring module, which is connected to the microservice processing unit; The security monitoring module uses anomaly detection algorithms (such as Isolation Forest, LOF) to monitor abnormal traffic in the network; uses deep learning algorithms (such as CNN, RNN) to identify and classify malicious behaviors.
6. The artificial intelligence-based network performance monitoring system according to claim 5, wherein The artificial intelligence-based network performance monitoring system further includes a configuration management module and a log management module. The configuration management module is connected to the rule library, and the log management module is connected to the microservice processing unit.
7. The artificial intelligence-based network performance monitoring system according to claim 6, wherein The artificial intelligence-based network performance monitoring system further includes a login module, which is connected to the configuration management module.
8. An artificial intelligence-based network performance monitoring method, applied to the artificial intelligence-based network performance monitoring system according to claim 7, wherein It includes the following steps: First, use the eBPF data acquisition module to collect network data in real time, and preprocess and extract features from the data through the microservice processing unit: After that, the predictive maintenance module applies machine learning algorithms to perform network performance prediction analysis, and the alarm generation module generates alarms according to the analysis results and the rule library, and pushes them to the maintenance terminal and the automated maintenance module through the GSM module; The automated maintenance module automatically executes predefined maintenance operations according to the alarm information to quickly restore network performance.