Network equipment abnormity monitoring and early warning method and system based on AI
Through the AI-based multi-source data collection and analysis module, combined with anomaly quantitative scoring and graded response, the problems of delayed response and high false alarm rate in traditional network equipment monitoring are solved, and real-time anomaly detection and potential risk warning of network equipment are achieved, ensuring the stable operation of network equipment.
Patent Information
- Application Number
- CN202510723750.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-31
- Publication Date
- 2025-09-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional network equipment monitoring methods have delayed responses, high false alarm rates, difficulty adapting to complex network environments, and are unable to predict potential anomalies, leading to serious consequences such as network outages and data leaks.
It adopts AI-based multi-source data collection, processing and analysis modules, combined with AI knowledge graph and parameter analysis, and uses abnormal quantitative scoring and graded response mechanisms to monitor and warn of abnormal conditions of network equipment in real time.
It realizes real-time anomaly detection and potential risk prediction of network equipment, improves the accuracy and timeliness of monitoring, and ensures the stable operation of network equipment.
Smart Images

Figure CN120639654A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of network equipment monitoring, and specifically relates to an AI-based network equipment anomaly monitoring and early warning method and system. Background Art
[0002] With the rapid development of information technology and the continuous advancement of enterprise digital transformation, network systems have become critical infrastructure supporting the operation of various businesses. As core components of network communications, the stability and reliability of network equipment directly impact the quality and security of the entire network service. However, factors such as equipment aging, configuration errors, network attacks, and traffic spikes can lead to serious consequences such as network outages, performance degradation, and even data leaks. Therefore, network equipment monitoring and early warning are essential.
[0003] Traditional network device monitoring relies primarily on manual inspections and rule-based alerting mechanisms, such as SNMP monitoring, log analysis, and threshold-based alarms. While these methods can detect some issues to a certain extent, they suffer from delayed response times, high false alarm rates, and difficulty adapting to complex network environments. Furthermore, these methods can only generate alerts for obvious anomalies in network devices and cannot predict potential anomalies. Summary of the Invention
[0004] The purpose of the present invention is to provide an AI-based network equipment abnormality monitoring and early warning method and system to solve the problems faced in the above-mentioned background technology.
[0005] The purpose of the present invention can be achieved through the following technical solutions:
[0006] An AI-based network equipment anomaly monitoring and early warning system, the early warning system comprising:
[0007] A multi-source data acquisition module, wherein the multi-source data acquisition module is used to obtain various characteristic parameters of network devices;
[0008] A data transmission and processing module, which is used to transmit the acquired characteristic parameters to the processing section for pre-processing, thereby obtaining the processed characteristic parameters;
[0009] An intelligent analysis module, which includes an AI knowledge graph and a parameter analysis module. The AI knowledge graph quickly determines whether an abnormality has occurred in a feature item corresponding to the acquired feature parameters based on the acquired feature parameters. The parameter analysis module analyzes multiple feature parameters of the device to determine whether an abnormality has occurred in the device.
[0010] An early warning response module includes an intelligent response module and a hierarchical response module. The intelligent response module quickly determines the abnormal location based on the AI knowledge graph and automatically generates an abnormal diagnosis report, while listing the possible causes of the abnormality; the hierarchical response module performs a hierarchical response based on the results of the parameter analysis module.
[0011] Furthermore, the data transmission and processing module preprocessing includes: processing missing values of each characteristic parameter, detecting and correcting abnormal values, deduplicating duplicate data, time series synchronization and normalization of each characteristic parameter.
[0012] Furthermore, the parameter analysis module analysis method is:
[0013] Based on the historical normal operation data of network equipment, set the optimal characteristic parameter value DE for each characteristic item sdi ,DE sdi is the optimal characteristic parameter value of the i-th characteristic item, and the characteristic parameters De of each characteristic item are obtained. i Finally, the data fusion technology is used to obtain the abnormal quantitative score Qr of the network equipment. The mathematical expression is:
[0014]
[0015] When Qr>Qr th When the error message is " ," it is determined that the network device is abnormal;
[0016] Where n is the number of acquired feature items, and i∈[1,n], A i is the influence weight of the i-th characteristic item, Qr th It is the preset abnormal quantitative scoring threshold.
[0017] Furthermore, the parameter analysis module is also used to judge the potential abnormal risk of the network device when there is no abnormal risk of the network device.
[0018] Furthermore, the parameter analysis module determines the potential abnormal risk of the network device by:
[0019] The risk exposure value KE is obtained by the following formula;
[0020]
[0021] Among them, Ep is the environmental impact value, Pen is the first deviation value, Pem is the second deviation value, α1 and α2 are their respective weight coefficients, T is the average ambient temperature, T Z is the standard temperature of the environment, H is the average humidity of the environment, H Zis the standard humidity of the environment, U is the voltage fluctuation value, EMI is the electromagnetic interference intensity of the environment, m is the number of associated devices, Qr j Qr is the abnormal quantitative score of the j-th associated device, 0j is the standard abnormal quantitative score of the jth associated device determined based on historical operation data, γ j is the weight coefficient of the jth associated device, t1 is the detection start time point, t2 is the current detection time point, Qr j (t) is the abnormal quantitative score of the jth associated device at time t, and j∈[1,m];
[0022] After the risk exposure value KE is obtained, it is compared with the preset risk exposure judgment threshold KE th Compare, when KE>KE th When the error message is displayed, it is determined that there is a potential abnormal risk in the network device.
[0023] Furthermore, the method for the early warning response module to perform a graded response is:
[0024] When it is determined that there is a potential abnormal risk in the network equipment, a yellow alarm is generated;
[0025] When it is determined that there is an abnormality in the network equipment and Qr th <Qr<Qr th 1, an orange alarm is generated; when Qr>Qr th 1, a red alert is generated;
[0026] where Qr th 1 is another preset abnormal quantitative scoring threshold, and the yellow alarm level is lower than the orange alarm level, and the orange alarm level is lower than the red alarm level.
[0027] An AI-based network equipment anomaly monitoring and early warning method, the early warning method is controlled and implemented by the above-mentioned AI-based network equipment anomaly monitoring and early warning system, the early warning method comprising:
[0028] Step 1: Acquire characteristic parameters of characteristic items related to the network device from multiple data terminals of the network device, and preprocess each characteristic parameter;
[0029] Step 2: Perform fusion analysis based on the acquired feature parameters to obtain anomaly quantitative scores and determine the abnormality of network devices;
[0030] Step 3: When the network equipment is operating normally, a detection cycle is planned. Based on the environmental impact of the network equipment and the deviation of related equipment during the detection cycle, a fusion analysis is performed to determine whether there is a potential abnormal risk in the network equipment.
[0031] Step 4: Based on the judgment results, make different levels of graded responses.
[0032] Beneficial effects of the present invention:
[0033] The present invention can perform a simple analysis of the parameters of a single characteristic item of a network device through the AI knowledge graph, thereby determining whether the network device has an abnormality. At the same time, it can perform a comprehensive analysis based on multiple parameter data of the network device. This can more accurately discover the more hidden abnormalities of the network device, and thus handle them in a timely manner to ensure the normal operation of the network device.
[0034] The present invention can judge the potential abnormal risks of network devices according to the environmental impact of the network devices and the relative deviations of various network devices under the normal operation of the network devices, so as to facilitate timely processing of potential abnormalities, thereby reducing the occurrence of subsequent abnormalities and ensuring the normal and stable operation of the network devices.
[0035] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0037] Figure 1 It is a system module block diagram of the present invention;
[0038] Figure 2 This is a flow chart of the monitoring and early warning method of the present invention. DETAILED DESCRIPTION
[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.
[0040] In one embodiment, an AI-based network equipment abnormality monitoring and early warning system is disclosed, such as Figure 1 As shown in the figure, the monitoring and early warning system mainly includes the following modules:
[0041] Multi-source data acquisition module, which allows users to obtain various characteristic parameters of network devices;
[0042] The data transmission and processing module is used to transmit the acquired characteristic parameters to the processing section for preprocessing, thereby obtaining the processed characteristic parameters. The preprocessing includes: processing missing values in each characteristic parameter, detecting and correcting outliers, deduplicating duplicate data, time series synchronization, and normalization of each characteristic parameter.
[0043] Intelligent analysis module, which includes an AI knowledge graph and a parameter analysis module. The AI knowledge graph quickly determines whether the feature items corresponding to the feature parameters are abnormal based on the acquired feature parameters. The parameter analysis module analyzes multiple feature parameters of the device to determine whether the device is abnormal.
[0044] The early warning response module includes an intelligent response module and a hierarchical response module. The intelligent response module quickly determines the abnormal location based on the AI knowledge graph and automatically generates an abnormal diagnosis report, while listing the possible causes of the abnormality; the hierarchical response module performs a hierarchical response based on the results of the parameter analysis module.
[0045] Through the above technical solution, the present application first obtains various characteristic parameters of network equipment through a multi-source data acquisition module, such as the device temperature, response time, CPU utilization, operating speed, packet loss rate and other characteristic parameters of the network equipment during operation, and then performs data processing and transmission. The data processing includes but is not limited to processing missing values of each characteristic parameter, detecting and correcting outliers, deduplicating duplicate data, timing synchronization and normalization of each characteristic parameter, etc., so that the original characteristic parameters obtained can be more accurate and the dimensions are consistent, which is convenient for subsequent analysis and calculation, intelligent analysis module; the intelligent analysis module is equipped with an AI knowledge graph and a parameter analysis module. The AI knowledge graph can quickly determine whether the characteristic item corresponding to the characteristic parameter is abnormal based on the acquired characteristic parameters. The AI knowledge graph can perform a simple analysis of the parameters of a single characteristic item of the network device to determine whether an abnormality occurs, and the parameter analysis module is based on the analysis of multiple characteristic parameters of the device, and can perform comprehensive analysis from multiple data ends of the network device. The system uses a comprehensive analysis to analyze the overall situation of the network equipment, so as to determine whether there are real-time abnormalities in the network equipment. In addition, the parameter analysis module is also used to judge the potential abnormal risks of the network equipment in combination with the environmental impact of the network equipment and the deviation of each network device when there is no abnormal risk in the network equipment, so as to deal with them in time to ensure the normal and stable operation of the network equipment. The early warning response module includes an intelligent response module and a hierarchical response module. The intelligent response module quickly determines the abnormal location and automatically generates an abnormal diagnosis report based on the AI knowledge graph, and lists the possible causes of the abnormality. In this way, the causes of some simple and common problems can be quickly determined, which is convenient for maintenance personnel to handle. The hierarchical response module performs hierarchical responses based on the results of the parameter analysis module. Different response levels are specified for deeper abnormalities or potential abnormal situations, so as to distinguish the severity of the abnormality and the urgency of the processing, which is convenient for operation and maintenance personnel to reasonably allocate resources, give priority to high-risk abnormalities, and ensure the stable operation of network equipment.
[0046] The parameter analysis module analysis method is: based on the historical normal operation data of the network equipment, set the optimal characteristic parameter value DE for each characteristic item sdi ,DE sdi is the optimal characteristic parameter value of the i-th characteristic item, and the characteristic parameters De of each characteristic item are obtained. i Finally, the data fusion technology is used to obtain the abnormal quantitative score Qr of the network equipment. The mathematical expression is:
[0047]
[0048] When Qr>Qr th When the error message is " ," it is determined that the network device is abnormal;
[0049] Among them, n is the number of acquired feature items, and i∈[1,n], A i is the influence weight of the i-th characteristic item, Qr th It is the preset abnormal quantitative scoring threshold.
[0050] The above scheme provides a method for the parameter analysis module to determine whether there is a real-time abnormality in the network equipment. First, based on the historical normal operation data of the network equipment, the optimal characteristic parameter value DE is set for each characteristic item. sdi For example, according to the data information of the network equipment during normal operation, an optimal characteristic parameter value is set for the device temperature, response time, CPU utilization, operating speed, and packet loss rate, and then the characteristic parameter De of each characteristic item is obtained in real time. i Finally, the data fusion technology is used to obtain the abnormal quantitative score Qr of the network equipment. The mathematical expression is:
[0051] A i is the influence weight of the i-th characteristic item, which is determined independently by the operation and maintenance management personnel based on experience. It can be seen from the formula that the greater the difference between the characteristic parameters of each characteristic item and the set optimal characteristic parameters, the greater the difference between the network equipment and the normal operation, and the greater the possibility of abnormality. Therefore, after obtaining the abnormal quantitative score of the network equipment, it is compared with the abnormal quantitative score threshold Qr preset based on experience and historical data. th Compare and when Qr>Qr th When the network device is abnormal, it is judged that there is a real-time abnormality in the network device. In this way, a comprehensive analysis can be performed by combining multiple parameter data of the network device, which can more accurately discover the relatively hidden abnormalities of the network device and deal with them in time to ensure the normal operation of the network device.
[0052] The parameter analysis module determines the potential abnormal risk of network equipment by: deriving the risk exposure value KE through the following formula;
[0053]
[0054] Among them, Ep is the environmental impact value, Pen is the first deviation value, Pem is the second deviation value, α1 and α2 are their respective weight coefficients, T is the average ambient temperature, T Z is the standard temperature of the environment, H is the average humidity of the environment, H Z is the standard humidity of the environment, U is the voltage fluctuation value, EMI is the electromagnetic interference intensity of the environment, m is the number of associated devices, Qr j Qr is the abnormal quantitative score of the j-th associated device, 0j is the standard abnormal quantitative score of the jth associated device determined based on historical operation data, γ jis the weight coefficient of the jth associated device, t1 is the detection start time point, t2 is the current detection time point, Qr j (t) is the abnormal quantitative score of the jth associated device at time t, and j∈[1,m];
[0055] After the risk exposure value KE is obtained, it is compared with the preset risk exposure judgment threshold KE th Compare, when KE>KE th When the error message is displayed, it is determined that there is a potential abnormal risk in the network device.
[0056] The above scheme provides a specific method for the parameter analysis module to determine whether there is a potential abnormal risk in the network device. Since the existing abnormal detection of network devices is mostly done after the device has an abnormal problem, it is impossible to analyze the potential abnormal risk, thus missing the best processing time. Therefore, this embodiment combines the risk exposure value to determine the potential abnormal risk of the network device. Specifically: because most network devices are interrelated, when a certain device has a problem, the corresponding devices will also have problems. Therefore, the formula is first used to determine the potential abnormal risk of the network device. The first deviation value is obtained, m is the number of associated equipment, Qr j Qr is the abnormal quantitative score of the j-th associated device, 0j is the standard abnormal quantitative score of the jth associated device determined based on historical operation data, γ j is the weight coefficient of the jth associated device. The first deviation value represents the difference between the abnormal quantitative score of each associated device and the respective standard abnormal quantitative score. Obviously, the larger the difference, the greater the difference between the associated device during operation and normal operation, and the greater the possibility of potential abnormal risk. Then, through the formula
[0057]
[0058] The second deviation value is obtained. In the second deviation value characterization formula,
[0059] It shows the difference between the abnormal quantitative score of a single associated device and all associated devices within the historical set detection time. It shows the difference between a single associated device and all associated devices at the current time. Because most network devices are interrelated, the larger the deviation value, the more it indicates that there are security risks in the network device. Similarly, the environment will also have an impact on the network device. For example, if the difference between the ambient temperature and the standard operating temperature during the detection time is too large, the ambient humidity is too large compared to the standard humidity, the voltage fluctuation or electromagnetic interference during the detection time is large, the impact on the network device will be greater. Therefore, the average ambient temperature, average ambient humidity, voltage fluctuation value, and electromagnetic interference intensity of the environment are obtained during the detection time. The electromagnetic interference intensity and temperature and humidity can be obtained through the corresponding detection equipment, and the voltage fluctuation value can be determined by the standard deviation, so through the formula
[0060] Ep=ln(1+|TT Z |+|HH Z |+EMI+U) to calculate the environmental impact value. It can be seen that when the value is larger, the impact of the environment on the network equipment is greater, and correspondingly, the possibility of potential abnormality of the network equipment is also greater. Therefore, based on the formula KE=Ep*(α1*Pen+α2*Pem), the risk exposure value KE is obtained, and then compared with the preset risk exposure judgment threshold KE th Compare, when KE>KE th When a network device experiences a potential abnormality risk, it is determined that the network device has a potential abnormality risk. This approach allows the network device to be judged for potential abnormality risks based on the environmental impact of the network device and the relative deviations of each network device while the network device is operating normally. This allows for timely handling of potential abnormalities, thereby reducing the occurrence of subsequent abnormalities and ensuring the normal and stable operation of the network device.
[0061] The early warning response module performs a graded response method as follows: when it is determined that a network device has a potential abnormal risk, a yellow alarm is generated;
[0062] When it is determined that there is an abnormality in the network equipment and Qr th <Qr<Qr th 1, an orange alarm is generated; when Qr>Qr th 1, a red alert is generated;
[0063] where Qr th 1 is another preset abnormal quantitative scoring threshold, and the yellow alarm level is lower than the orange alarm level, and the orange alarm level is lower than the red alarm level.
[0064] Through the above scheme, the corresponding response is made according to the judgment result. When it is judged that the network equipment has potential abnormal risk, it means that the equipment has potential risk at this time, but the performance will not affect the normal operation of the business in the short term, and it can be repaired within the set period, then a yellow alarm is generated; when it is judged that the network equipment has abnormality, and Qr th<Qr<Qr th 1, indicating that the equipment may affect the normal operation of local business and may even trigger a chain reaction, which needs to be resolved in a shorter time, then an orange alarm is generated; and when Qr>Qr th When the value is 1, it indicates a serious anomaly that could lead to core business interruption, network paralysis, or a major security incident. Immediate action is required, resulting in a red alert. This designates different response levels for deeper or potential anomalies, distinguishing their severity and the urgency of their response. This allows operations personnel to allocate resources appropriately, prioritize high-risk anomalies, and ensure stable network equipment operation.
[0065] In one embodiment, an AI-based network equipment abnormality monitoring and early warning method is also disclosed, such as Figure 2 As shown in FIG, the monitoring and early warning method mainly includes:
[0066] Step 1: Acquire characteristic parameters of characteristic items related to the network device from multiple data terminals of the network device, and preprocess each characteristic parameter;
[0067] Step 2: Perform fusion analysis based on the acquired feature parameters to obtain anomaly quantitative scores and determine the abnormality of network devices;
[0068] Step 3: When the network equipment is operating normally, a detection cycle is planned. Based on the environmental impact of the network equipment and the deviation of related equipment during the detection cycle, a fusion analysis is performed to determine whether there is a potential abnormal risk in the network equipment.
[0069] Step 4: Based on the judgment results, make different levels of graded responses.
[0070] The above content is merely an example and explanation of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they should all fall within the scope of protection of the present invention.
Claims
1. An AI-based network equipment anomaly monitoring and early warning system, characterized by: The early warning system includes: A multi-source data acquisition module, wherein the multi-source data acquisition module is used to obtain various characteristic parameters of network devices; A data transmission and processing module, which is used to transmit the acquired characteristic parameters to the processing section for pre-processing, thereby obtaining the processed characteristic parameters; An intelligent analysis module, which includes an AI knowledge graph and a parameter analysis module. The AI knowledge graph quickly determines whether an abnormality has occurred in a feature item corresponding to the acquired feature parameters based on the acquired feature parameters. The parameter analysis module analyzes multiple feature parameters of the device to determine whether an abnormality has occurred in the device. An early warning response module includes an intelligent response module and a hierarchical response module. The intelligent response module quickly determines the abnormal location based on the AI knowledge graph and automatically generates an abnormal diagnosis report, while listing the possible causes of the abnormality; the hierarchical response module performs a hierarchical response based on the results of the parameter analysis module.
2. The AI-based network equipment abnormality monitoring and early warning system according to claim 1 is characterized in that: The data transmission and processing module preprocessing includes: processing missing values of each characteristic parameter, detecting and correcting abnormal values, deduplicating duplicate data, time series synchronization and normalization of each characteristic parameter.
3. The AI-based network equipment abnormality monitoring and early warning system according to claim 2 is characterized in that: The parameter analysis module analysis method is: Based on the historical normal operation data of network equipment, set the optimal characteristic parameter value DE for each characteristic item sdi ,DE sdi is the optimal characteristic parameter value of the i-th characteristic item, and the characteristic parameters De of each characteristic item are obtained. i Finally, the data fusion technology is used to obtain the abnormal quantitative score Qr of the network equipment. The mathematical expression is: When Qr>Qr th When the error message is " ," it is determined that the network device is abnormal; Where n is the number of acquired feature items, and i∈[1,n], A i is the influence weight of the i-th characteristic item, Qr th It is the preset abnormal quantitative scoring threshold.
4. The AI-based network equipment abnormality monitoring and early warning system according to claim 3 is characterized in that: The parameter analysis module is also used to judge the potential abnormal risk of the network device when there is no abnormal risk of the network device.
5. The AI-based network equipment abnormality monitoring and early warning system according to claim 4 is characterized in that: The method by which the parameter analysis module judges the potential abnormal risk of network equipment is: The risk exposure value KE is obtained by the following formula; Among them, Ep is the environmental impact value, Pen is the first deviation value, Pem is the second deviation value, α1 and α2 are their respective weight coefficients, T is the average ambient temperature, T Z is the standard temperature of the environment, H is the average humidity of the environment, H Z is the standard humidity of the environment, U is the voltage fluctuation value, EMI is the electromagnetic interference intensity of the environment, m is the number of associated devices, Qr j is the abnormal quantitative score of the j-th associated device, is the standard abnormal quantitative score of the jth associated device determined based on historical operation data, γ j is the weight coefficient of the jth associated device, t1 is the detection start time point, t2 is the current detection time point, Qr j (t) is the abnormal quantitative score of the jth associated device at time t, and j∈[1,m]; After the risk exposure value KE is obtained, it is compared with the preset risk exposure judgment threshold KE th Compare, when KE>KE th When the error message is displayed, it is determined that there is a potential abnormal risk in the network device.
6. The AI-based network equipment abnormality monitoring and early warning system according to claim 5 is characterized in that: The method for the early warning response module to perform a hierarchical response is: When it is determined that there is a potential abnormal risk in the network equipment, a yellow alarm is generated; When it is determined that there is an abnormality in the network equipment and Qr th <Qr<Qr th 1, an orange alarm is generated; when Qr>Qr th 1, a red alert is generated; where Qr th 1 is another preset abnormal quantitative scoring threshold, and the yellow alarm level is lower than the orange alarm level, and the orange alarm level is lower than the red alarm level.
7. An AI-based network equipment anomaly monitoring and early warning method, wherein the early warning method is controlled and implemented by the AI-based network equipment anomaly monitoring and early warning system according to any one of claims 1 to 6, characterized in that: The early warning method includes: Step 1: Acquire characteristic parameters of characteristic items related to the network device from multiple data terminals of the network device, and preprocess each characteristic parameter; Step 2: Perform fusion analysis based on the acquired feature parameters to obtain anomaly quantitative scores and determine the abnormality of network devices; Step 3: When the network equipment is operating normally, a detection cycle is planned. Based on the environmental impact of the network equipment and the deviation of related equipment during the detection cycle, a fusion analysis is performed to determine whether there is a potential abnormal risk in the network equipment. Step 4: Based on the judgment results, make different levels of graded responses.
Citation Information
Cited By
Performance monitoring method and device of network equipment and electronic equipment
CN120856548A
A method, apparatus, and electronic device for monitoring the performance of network devices.
CN120856548B
Network equipment online management system based on security protocol and AI intelligent agent
CN121690559A
Network operation and maintenance management system based on big data analysis
CN121690972A
