Intelligent alarm dynamic suppression and collaborative traceability method and system for 5G multilayer heterogeneous network
Through the combination of deep learning and graph neural networks, intelligent alarm dynamic suppression and collaborative tracing are achieved in 5G multi-layer heterogeneous networks, solving the problems of alarm overload and difficulty in fault location, and improving operation and maintenance efficiency and network stability.
Patent Information
- Application Number
- CN202510927486.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-07
- Publication Date
- 2025-09-05
AI Technical Summary
5G multi-layer heterogeneous networks suffer from problems such as alarm overload, difficulty in fault location, and high manual dependence, resulting in low operation and maintenance efficiency.
A deep learning model is used for real-time alarm classification and priority sorting. A rule engine and reinforcement learning are combined to generate dynamic suppression strategies. A network topology map is constructed and a graph neural network is used to trace the fault propagation path, providing optimization suggestions based on historical fault data.
It realizes intelligent alarm identification, dynamic suppression and collaborative tracing, improves operation and maintenance efficiency and network stability, and reduces the probability of similar failures.
Smart Images
Figure CN120602316A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of 5G communication technology, and specifically to a method and system for intelligent alarm dynamic suppression and collaborative tracing in a 5G multi-layer heterogeneous network. Background Art
[0002] With the large-scale deployment of 5G technology, multi-layer heterogeneous network architecture has become the mainstream form, but its complexity leads to the following problems:
[0003] Alarm overload: Massive alarms generated by multi-layer network devices make it difficult for operations and maintenance personnel to handle them efficiently.
[0004] Difficulty locating faults: Cross-layer and cross-domain faults are difficult to quickly trace using traditional methods;
[0005] High dependence on manual labor: Existing methods rely on expert experience and lack automation.
[0006] The present invention aims to solve the above problems through intelligent means. Summary of the Invention
[0007] The purpose of the present invention is to provide a method and system for intelligent alarm dynamic suppression and collaborative tracing of 5G multi-layer heterogeneous networks to solve the problems raised in the above background technology.
[0008] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent alarm dynamic suppression and collaborative tracing in a 5G multi-layer heterogeneous network, comprising the following steps:
[0009] Classify and prioritize real-time alarms based on deep learning models, and link them with historical alarm libraries to identify duplicate or low-priority alarms, enabling intelligent alarm recognition.
[0010] Dynamically generate alarm filtering, merging, and suppression strategies through a rule engine and reinforcement learning. Prioritize suppressing non-critical alarms in network congestion scenarios, completing dynamic suppression strategy development.
[0011] Build a network topology map, combine business traffic and device status data, use graph neural networks to trace the fault propagation path, generate a visual traceability report, and mark the root cause node for collaborative traceability analysis;
[0012] Based on historical fault data, it recommends network configuration optimization solutions and provides preventive maintenance suggestions, enabling the generation of intelligent optimization suggestions.
[0013] Preferably, real-time alarms are classified and prioritized based on a deep learning model, wherein the deep learning model is an LSTM or Transformer model, which is used to analyze and process real-time alarm data, and the alarms are divided into different categories and their priorities are determined according to preset rules.
[0014] Preferably, alarm filtering, merging and suppression strategies are dynamically generated through rule engines and reinforcement learning. The rule engine preliminarily screens and processes alarms based on a preset rule base, and the reinforcement learning algorithm continuously optimizes the alarm processing strategy based on network status and historical alarm data. In network congestion scenarios, non-critical alarms are prioritized for filtering, merging or suppression based on their priority and criticality.
[0015] Preferably, a network topology map is constructed, and the fault propagation path is traced using a graph neural network in combination with business traffic and device status data. Specifically, the connection relationships between devices in the network are first collected to construct a network topology map, and business traffic data and device status data are collected at the same time. These data are input into the graph neural network model. Through model learning and analysis, the fault propagation path in the network is determined, and a visual tracing report containing the fault propagation path and root cause node annotations is generated.
[0016] Preferably, based on historical fault data, network configuration optimization solutions are recommended and preventive maintenance suggestions are provided. Specifically, historical fault data is mined and analyzed to identify network configuration factors that cause faults, and corresponding network configuration optimization solutions are recommended based on the analysis results, such as load balancing strategy adjustments. At the same time, combined with the patterns and trends of historical faults, preventive maintenance suggestions are provided, including equipment inspection cycle adjustments and software upgrade plans, to reduce the probability of similar faults occurring.
[0017] A system for intelligent alarm dynamic suppression and collaborative tracing method in 5G multi-layer heterogeneous networks, comprising:
[0018] Intelligent alarm recognition module, which is used to classify and prioritize real-time alarms based on deep learning models and correlate historical alarm libraries to identify duplicate or low-priority alarms;
[0019] Dynamic suppression strategy module, which uses a rule engine and reinforcement learning to dynamically generate alarm filtering, merging, and suppression strategies, and prioritizes suppressing non-critical alarms in network congestion scenarios;
[0020] The collaborative traceability analysis module is used to build a network topology map, combine business traffic and device status data, use graph neural networks to trace the fault propagation path, and generate a visual traceability report with the root cause node marked;
[0021] The intelligent optimization suggestion module is used to recommend network configuration optimization solutions and provide preventive maintenance suggestions based on historical fault data to reduce the probability of similar faults.
[0022] Preferably, in the intelligent alarm recognition module, the deep learning model is trained to learn a large number of real-time alarm data samples to establish an alarm classification and priority sorting model. The model can quickly and accurately output alarm category and priority information based on the real-time input alarm data, and at the same time compare it with the historical alarm library to identify duplicate or low-priority alarms.
[0023] Preferably, in the dynamic suppression strategy module, the rule engine has a built-in preset alarm processing rule base, which matches the real-time alarm characteristics with the rule base to perform preliminary screening and processing of the alarms; the reinforcement learning algorithm continuously adjusts and optimizes the alarm filtering, merging and suppression strategies based on the real-time status of the network and historical alarm processing data. In the network congestion scenario, non-critical alarms are prioritized for filtering, merging or suppression operations based on the priority and criticality assessment results of the alarms.
[0024] Preferably, in the collaborative traceability analysis module, the network topology map construction unit is responsible for collecting the connection relationship information between each device in the network and building a complete network topology structure; the data fusion unit integrates and processes the business traffic data and device status data; the graph neural network analysis unit uses the integrated data to learn the propagation law of the fault in the network through the graph neural network model and determine the fault propagation path; the report generation unit generates a visual traceability report based on the analysis results, and clearly marks the root cause node of the fault in the report.
[0025] Preferably, in the intelligent optimization suggestion module, the historical fault data analysis unit mines and analyzes the historical fault data to extract the network configuration factors that cause the faults; the network configuration optimization scheme recommendation unit recommends appropriate network configuration optimization schemes based on the analysis results and the actual network conditions, such as load balancing strategy adjustment and routing strategy optimization; the preventive maintenance suggestion generation unit formulates equipment inspection plans, software upgrade schedules and preventive maintenance suggestions based on the laws and trends of historical faults to reduce the probability of similar faults occurring in the future.
[0026] Compared with the prior art, the present invention has the following beneficial effects:
[0027] The proposed method and system for intelligent dynamic alarm suppression and collaborative tracing for 5G multi-layer heterogeneous networks uses a machine learning algorithm to analyze network alarm information in real time, dynamically generating alarm suppression strategies to reduce redundant alarm interference. This system also enables rapid fault root cause location based on multi-dimensional collaborative analysis of network topology, service traffic, and device status. This method can significantly improve the operational efficiency and stability of 5G multi-layer heterogeneous networks, while also providing intelligent optimization recommendations to prevent similar faults from occurring. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION
[0029] In order to clearly and completely describe the objectives and technical solutions of the present invention and make the advantages more clearly understood, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, not all of them, and are only used to explain the embodiments of the present invention, not to limit the embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0030] For example 1, please refer to Figure 1 The present invention provides a technical solution: a method for intelligent alarm dynamic suppression and collaborative tracing in a 5G multi-layer heterogeneous network, comprising the following steps:
[0031] Real-time alarms are classified and prioritized based on deep learning models, and linked to historical alarm libraries to identify duplicate or low-priority alarms and achieve intelligent alarm recognition. The deep learning model is an LSTM or Transformer model, which is used to analyze and process real-time alarm data, classify alarms into different categories and determine their priorities based on preset rules.
[0032] Through the rule engine and reinforcement learning, alarm filtering, merging and suppression strategies are dynamically generated. Non-critical alarms are suppressed first in network congestion scenarios, and dynamic suppression strategies are formulated. The rule engine preliminarily screens and processes alarms based on the preset rule library. The reinforcement learning algorithm continuously optimizes the alarm processing strategy based on the network status and historical alarm data. In network congestion scenarios, non-critical alarms are prioritized for filtering, merging or suppression based on their priority and criticality.
[0033] Build a network topology map, combine business traffic and device status data, and use graph neural networks to trace the fault propagation path: first, collect the connection relationship between each device in the network to build a network topology map, and at the same time collect business traffic data and device status data. Input these data into the graph neural network model. Through model learning and analysis, determine the fault propagation path in the network, and generate a visual traceability report containing the fault propagation path and root cause node annotations; generate a visual traceability report and mark the root cause nodes of the fault for collaborative traceability analysis.
[0034] Recommend network configuration optimization solutions and provide preventive maintenance suggestions based on historical fault data: This system mines and analyzes historical fault data to identify network configuration factors that lead to faults. Based on the analysis results, it recommends corresponding network configuration optimization solutions, such as load balancing strategy adjustments. Furthermore, based on the patterns and trends of historical faults, it provides preventive maintenance suggestions, including adjustments to equipment inspection cycles and software upgrade plans, to reduce the probability of similar faults. This enables the generation of intelligent optimization suggestions.
[0035] In the second embodiment, based on the first embodiment, a system for intelligent alarm dynamic suppression and collaborative tracing method in a 5G multi-layer heterogeneous network is proposed, including:
[0036] The intelligent alarm recognition module is used to classify and prioritize real-time alarms based on a deep learning model, and associate it with the historical alarm library to identify duplicate or low-priority alarms. The deep learning model is trained on a large number of real-time alarm data samples to establish an alarm classification and prioritization model. The model can quickly and accurately output alarm category and priority information based on real-time input alarm data, and at the same time compare it with the historical alarm library to identify duplicate or low-priority alarms.
[0037] The dynamic suppression strategy module is used to dynamically generate alarm filtering, merging, and suppression strategies through a rule engine and reinforcement learning, and prioritizes suppressing non-critical alarms in network congestion scenarios. The rule engine has a built-in preset alarm processing rule library, which matches real-time alarm characteristics with the rule library to perform preliminary screening and processing of alarms. The reinforcement learning algorithm continuously adjusts and optimizes alarm filtering, merging, and suppression strategies based on the real-time network status and historical alarm processing data. In network congestion scenarios, non-critical alarms are prioritized for filtering, merging, or suppression based on their priority and criticality assessment results.
[0038] The collaborative traceability analysis module is used to construct a network topology map, combine business traffic and equipment status data, use graph neural networks to trace the fault propagation path, and generate a visual traceability report with the root cause nodes marked. The network topology map construction unit is responsible for collecting the connection relationship information between each device in the network and building a complete network topology structure. The data fusion unit integrates and processes business traffic data and equipment status data. The graph neural network analysis unit uses the integrated data to learn the propagation law of faults in the network through the graph neural network model and determine the fault propagation path. The report generation unit generates a visual traceability report based on the analysis results and clearly marks the root cause nodes in the report.
[0039] The intelligent optimization suggestion module is used to recommend network configuration optimization solutions and provide preventive maintenance suggestions based on historical fault data to reduce the probability of similar faults. The historical fault data analysis unit mines and analyzes historical fault data to extract network configuration factors that lead to faults. The network configuration optimization solution recommendation unit recommends appropriate network configuration optimization solutions, such as load balancing strategy adjustment and routing strategy optimization, based on the analysis results and the actual network situation. The preventive maintenance suggestion generation unit formulates equipment inspection plans, software upgrade schedules, and preventive maintenance suggestions based on the patterns and trends of historical faults to reduce the probability of similar faults occurring in the future.
[0040] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A method for intelligent dynamic alarm suppression and collaborative tracing in 5G multi-layer heterogeneous networks, characterized by: The following steps are involved: Classify and prioritize real-time alarms based on deep learning models, and link them with historical alarm libraries to identify duplicate or low-priority alarms, enabling intelligent alarm recognition. Dynamically generate alarm filtering, merging, and suppression strategies through a rule engine and reinforcement learning. Prioritize suppressing non-critical alarms in network congestion scenarios, completing dynamic suppression strategy development. Build a network topology map, combine business traffic and device status data, use graph neural networks to trace the fault propagation path, generate a visual traceability report, and mark the root cause node for collaborative traceability analysis; Based on historical fault data, it recommends network configuration optimization solutions and provides preventive maintenance suggestions to achieve the generation of intelligent optimization suggestions.
2. The method for intelligent dynamic suppression and collaborative tracing of alarms in a 5G multi-layer heterogeneous network according to claim 1 is characterized by: Real-time alarms are classified and prioritized based on a deep learning model, where the deep learning model is an LSTM or Transformer model. The model is used to analyze and process real-time alarm data, and the alarms are divided into different categories and their priorities are determined according to preset rules.
3. The method for intelligent dynamic suppression and collaborative tracing of alarms in a 5G multi-layer heterogeneous network according to claim 2 is characterized in that: Alarm filtering, merging, and suppression strategies are dynamically generated through the rule engine and reinforcement learning. The rule engine preliminarily screens and processes alarms based on a preset rule base. The reinforcement learning algorithm continuously optimizes alarm processing strategies based on network status and historical alarm data. In network congestion scenarios, non-critical alarms are prioritized for filtering, merging, or suppression based on their priority and criticality.
4. The method for intelligent dynamic suppression and collaborative tracing of alarms in a 5G multi-layer heterogeneous network according to claim 3 is characterized by: Build a network topology map, combine business traffic and device status data, and use graph neural networks to trace the fault propagation path. Specifically: first collect the connection relationship between each device in the network to build a network topology map, and at the same time collect business traffic data and device status data. Input this data into the graph neural network model. Through model learning and analysis, determine the fault propagation path in the network, and generate a visual tracing report containing the fault propagation path and root cause node annotations.
5. The method for intelligent dynamic suppression and collaborative tracing of alarms in a 5G multi-layer heterogeneous network according to claim 4 is characterized in that: Based on historical fault data, we recommend network configuration optimization solutions and provide preventive maintenance suggestions. Specifically, we mine and analyze historical fault data to identify network configuration factors that lead to faults, and recommend corresponding network configuration optimization solutions based on the analysis results, such as load balancing strategy adjustments. At the same time, based on the patterns and trends of historical faults, we provide preventive maintenance suggestions, including equipment inspection cycle adjustments and software upgrade plans, to reduce the probability of similar faults.
6. A system for the intelligent alarm dynamic suppression and collaborative tracing method for 5G multi-layer heterogeneous networks according to claim 5, characterized in that: include: Intelligent alarm recognition module, which is used to classify and prioritize real-time alarms based on deep learning models and correlate historical alarm libraries to identify duplicate or low-priority alarms; Dynamic suppression strategy module, which uses a rule engine and reinforcement learning to dynamically generate alarm filtering, merging, and suppression strategies, and prioritizes suppressing non-critical alarms in network congestion scenarios; The collaborative traceability analysis module is used to build a network topology map, combine business traffic and device status data, use graph neural networks to trace the fault propagation path, and generate a visual traceability report with the root cause node marked; The intelligent optimization suggestion module is used to recommend network configuration optimization solutions and provide preventive maintenance suggestions based on historical fault data to reduce the probability of similar faults.
7. A system according to claim 6, characterized in that: In the intelligent alarm identification module, the deep learning model is trained to learn a large number of real-time alarm data samples to establish an alarm classification and priority sorting model. The model can quickly and accurately output alarm category and priority information based on real-time input alarm data, and at the same time compare it with the historical alarm library to identify duplicate or low-priority alarms.
8. A system according to claim 7, characterized in that: In the dynamic suppression strategy module, the rule engine has a built-in preset alarm processing rule library. It matches the real-time alarm characteristics with the rule library to perform preliminary screening and processing of alarms. The reinforcement learning algorithm continuously adjusts and optimizes alarm filtering, merging, and suppression strategies based on real-time network status and historical alarm processing data. In network congestion scenarios, it prioritizes filtering, merging, or suppressing non-critical alarms based on their priority and criticality assessment results.
9. A system according to claim 8, characterized in that: In the collaborative traceability analysis module, the network topology map construction unit is responsible for collecting the connection relationship information between each device in the network and building a complete network topology structure; the data fusion unit integrates and processes the business traffic data and device status data; the graph neural network analysis unit uses the integrated data to learn the propagation law of faults in the network through the graph neural network model and determine the fault propagation path; the report generation unit generates a visual traceability report based on the analysis results, and clearly marks the root cause node of the fault in the report.
10. A system according to claim 9, characterized in that: In the intelligent optimization suggestion module, the historical fault data analysis unit mines and analyzes historical fault data to extract the network configuration factors that cause faults. The network configuration optimization solution recommendation unit recommends appropriate network configuration optimization solutions based on the analysis results and the actual network situation, such as load balancing strategy adjustment and routing strategy optimization. The preventive maintenance suggestion generation unit formulates equipment inspection plans, software upgrade schedules, and preventive maintenance suggestions based on the patterns and trends of historical faults to reduce the probability of similar faults occurring in the future.