CN119728397B - Network fault prediction method and system - Google Patents
Network fault prediction method and system Download PDFInfo
- Publication number
- CN119728397B CN119728397B CN202510238401.9A CN202510238401A CN119728397B CN 119728397 B CN119728397 B CN 119728397B CN 202510238401 A CN202510238401 A CN 202510238401A CN 119728397 B CN119728397 B CN 119728397B
- Authority
- CN
- China
- Prior art keywords
- network
- fluctuation
- data
- historical
- sub
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
本发明涉及数据管理技术领域，公开了一种网络故障预测方法及系统，该方法包括：截取预设时段内的网络数据，构建网络波动‑时刻数据链；将网络波动值与波动范围进行比对，将网络波动‑时刻数据链拆分为异常波动数据链与正常波动数据链；提取网络特征构建同类型网络数列，获得同类型网络数列的子波动影响因子；根据所有子波动影响因子获得波动影响因子集合；对波动影响因子集合进行整合获得网络数据的最终影响因子，根据最终影响因子判断是否进行网络故障预警。本申请通过深度分析历史数据、细化波动模式以及多维度信息融合，解决了现有技术中准确性差、适应性弱等问题，提供了一种高效且智能的网络故障预测解决方案。
The present invention relates to the field of data management technology, and discloses a network fault prediction method and system, the method comprising: intercepting network data within a preset time period, constructing a network fluctuation-time data chain; comparing the network fluctuation value with the fluctuation range, and splitting the network fluctuation-time data chain into an abnormal fluctuation data chain and a normal fluctuation data chain; extracting network features to construct the same type of network series, and obtaining the sub-fluctuation influence factors of the same type of network series; obtaining a set of fluctuation influence factors based on all sub-fluctuation influence factors; integrating the set of fluctuation influence factors to obtain the final influence factor of the network data, and judging whether to issue a network fault warning based on the final influence factor. This application solves the problems of poor accuracy and weak adaptability in the prior art by deeply analyzing historical data, refining fluctuation patterns, and multi-dimensional information fusion, and provides an efficient and intelligent network fault prediction solution.
Description
Technical Field
The invention relates to the technical field of data management, in particular to a network fault prediction method and system.
Background
With the rapid development of the internet and communication technology, network systems have become an integral part of modern society. From enterprise data centers to home broadband, the stability and reliability of network services directly affect the proper operation of each industry. However, as the scale of networks increases and the complexity increases, the frequency and types of network failures become increasingly diverse, which presents a significant challenge to network operation and maintenance management. Network failures mainly include types of connection interruption, packet loss, delay fluctuation, bandwidth abnormality and the like, and these failures not only affect the performance of the network, but also may cause a series of cascading reactions, such as user experience degradation, service unavailability and the like.
Most of the existing network fault prediction methods are based on single monitoring indexes (such as delay and bandwidth) to perform abnormality detection, and the method is easily influenced by environmental noise and short-term fluctuation, so that the accuracy of fault early warning is low, and normal fluctuation and potential faults cannot be effectively distinguished. With the continuous change of network architecture and traffic patterns, traditional methods based on static thresholds or fixed rules are difficult to adapt to real-time changing network conditions. Moreover, the existing method does not fully utilize the historical network data for deep analysis, in particular to analysis of a historical fluctuation mode.
Therefore, there is a need to design a network failure prediction method and system to solve the problems in the prior art.
Disclosure of Invention
In view of the above, the invention provides a network fault prediction method and a system, which aim to solve the problems of low accuracy of the current network fault early warning, poor adaptability in a dynamic environment and lack of historical data analysis.
In one aspect, the present invention provides a network failure prediction method, including:
intercepting network data in a preset period, and analyzing the network data to construct a network fluctuation-moment data chain;
comparing the network fluctuation value of each time window in the network fluctuation-moment data chain with the fluctuation range, and splitting the network fluctuation-moment data chain into an abnormal fluctuation data chain and a normal fluctuation data chain according to the comparison result;
Extracting network characteristics of each time window in the abnormal fluctuation data chain, carrying out cluster analysis on network data of each time window in the abnormal fluctuation data chain according to the network characteristics, constructing network arrays of the same type, and obtaining wavelet influence factors of the network arrays of the same type according to the network arrays of the same type and corresponding historical same type fluctuation times;
Analyzing the network sequences of the same type remained in the abnormal fluctuation data chain, determining corresponding wavelet influence factors, and obtaining a fluctuation influence factor set according to all the wavelet influence factors;
And integrating the fluctuation influence factor set to obtain a final influence factor of the network data, and judging whether to perform network fault early warning according to the final influence factor.
Further, when analyzing the network data to construct a network fluctuation-time data chain, the method includes:
acquiring a first blank data chain and a second blank data chain, wherein each blank data chain is provided with a plurality of data chain nodes and connecting joints;
Acquiring network fluctuation values of each time window according to the network data, and generating a chain change mark by using the network data with all the network fluctuation values not being zero;
And transferring all network data carrying the chain change mark to the first blank data chain to obtain the network fluctuation-moment data chain.
Further, when the network fluctuation value of each time window in the network fluctuation-time data chain is compared with a fluctuation range, the fluctuation range is obtained by:
Collecting a historical operation data set, and extracting a non-fault network fluctuation set from the historical operation data set;
Normalizing the non-fault network fluctuation centralized data, wherein the normalization process comprises Min-Max normalization;
Dividing the normalized non-fault network fluctuation set into a training set and a testing set according to a ratio of 4:1, training an LSTM (Long Short-Term Memory network) model, and obtaining a predicted value of each historical time window;
obtaining an actual value of each historical time window according to the non-fault network fluctuation set, and obtaining a predicted deviation multiplying power of each historical time window according to the predicted value of each historical time window and the actual value of each historical time window, wherein the predicted deviation multiplying power is a ratio of the predicted value to the actual value;
Dividing the data in the network fluctuation-moment data chain according to time windows, taking each time window as an input of an LSTM model, obtaining an output result of the LSTM model, combining the output result with a predicted deviation multiplying power of a corresponding historical time window, and determining a fluctuation range of each time window.
Further, when the network data of each time window in the abnormal fluctuation data chain is subjected to cluster analysis according to the network characteristics, the method comprises the following steps:
The network characteristics comprise throughput, packet loss rate and network traffic load;
determining an initialization neighborhood radius through a k-distance graph, and determining MinPts as 6;
taking each network characteristic as a point, scanning all points, and finding out points with the number of points being greater than or equal to MinPts in the neighborhood as core points;
starting from each core point, checking the points in the neighborhood, if the points in the neighborhood are core points, continuing to expand the cluster, if the points in the neighborhood are boundary points, adding the points into the current cluster, and if one point is not a point in the neighborhood of any core point and cannot form clusters with other points, marking the point as a noise point.
Further, when obtaining the wavelet influence factor of the network number series of the same type according to the network number series of the same type and the corresponding historical same type fluctuation times, the method comprises the following steps:
wherein Gz represents a wavelet influence factor, N represents the number of network fluctuation values in the same type network sequence, bi represents an i-th network fluctuation value in the same type network sequence, C represents the number of historical fluctuation times, and B0 represents a right boundary value.
Further, when obtaining the fluctuation influence factor combination according to all the wavelet influence factors, the method comprises the following steps:
determining the median and variance of all of the wavelet impact factors;
Extracting wavelet influence factors greater than the median from all the wavelet influence factors to construct a first data set;
Extracting wavelet influence factors larger than the variance from all the wavelet influence factors to construct a second data set;
Determining whether an intersection exists between the first data set and the second data set;
if yes, taking the intersection value as the fluctuation influence factor set;
if not, carrying out non-repeated fusion on the first data set and the second data set to construct the fluctuation influence factor set, wherein the non-repeated fusion is to reserve the non-repeated sub-fluctuation influence factors in the first data set and the second data set, reserve one repeated sub-fluctuation influence factor in the first data set and the second data set, and delete the rest repeated sub-fluctuation influence factors.
Further, when the fluctuation influence factor set is integrated to obtain a final influence factor of the network data, the method includes:
Comparing the fluctuation influence factor set with history judgment data, wherein the history judgment data comprises a history fluctuation influence factor set and a history final influence factor;
When the historical judgment data contains data with the similarity larger than a similarity threshold value with the fluctuation influence factor set, taking a historical final influence factor corresponding to the historical fluctuation influence factor set as the final influence factor;
and when the similarity between the historical fluctuation influence factor set and the fluctuation influence factor set in the historical judgment data is smaller than or equal to a similarity threshold value, determining the final influence factor according to the fluctuation influence factor set.
Further, when determining the final influence factor according to the fluctuation influence factor set, the method includes:
wherein Y represents the final influence factor, M represents the number of the fluctuation influence factors in the fluctuation influence factor set, Representing the weight corresponding to the i < th > sub-fluctuation influencing factor,Representing the i-th sub-fluctuation influencing factor,Representing the smallest sub-fluctuation influencing factor,For the maximum wavelet impact factor,For all ofIs the maximum value of (a).
Further, when judging whether to perform network fault early warning according to the final influence factor, the method includes:
Comparing the final influence factor with an influence factor threshold value, and judging whether to perform network fault early warning according to the comparison result;
and when the final influence factor is larger than the influence factor threshold, judging to perform network fault early warning, acquiring a difference value between the final influence factor and the influence factor threshold, and determining a network fault early warning level according to the difference value, wherein the network fault early warning level and the difference value are in a proportional relation.
Compared with the prior art, the method has the beneficial effects that the dynamic change of the network can be captured by intercepting the network data in the preset period and constructing the fluctuation-moment data chain, so that the problem that the traditional method only depends on a single network index is avoided. By comparing the fluctuation range, the data chain is split into an abnormal fluctuation data chain and a normal fluctuation data chain, and normal fluctuation and potential faults are effectively distinguished, so that the fault early warning precision is improved. Network characteristics in the abnormal fluctuation data chain are extracted and cluster analysis is carried out, so that different types of fluctuation modes are further refined, and the fluctuation influence factors of the operators are calculated by combining the times of the historical fluctuation of the same type, so that the fault prediction is more accurate. By integrating the fluctuation influence factor sets, final influence factors of network data are generated, whether the network has fault risks or not can be accurately judged, early warning can be timely sent out, and the efficiency and response speed of network operation and maintenance are improved.
On the other hand, the application also provides a network fault prediction system, which is used for applying the network fault prediction method and comprises the following steps:
The acquisition unit is configured to intercept network data in a preset period, analyze the network data and construct a network fluctuation-moment data chain;
The processing unit is configured to compare the network fluctuation value of each time window in the network fluctuation-moment data chain with the fluctuation range, and split the network fluctuation-moment data chain into an abnormal fluctuation data chain and a normal fluctuation data chain according to the comparison result;
the execution unit is configured to extract the network characteristics of each time window in the abnormal fluctuation data chain, perform cluster analysis on the network data of each time window in the abnormal fluctuation data chain according to the network characteristics, construct the same type of network array, and obtain the wavelet influence factors of the same type of network array according to the same type of network array and the corresponding historical same type fluctuation times;
The analysis unit is configured to analyze the network sequences of the same type remained in the abnormal fluctuation data chain, determine corresponding wavelet influence factors and obtain a fluctuation influence factor set according to all the wavelet influence factors;
And the early warning unit is configured to integrate the fluctuation influence factor set to obtain a final influence factor of the network data, and judge whether to perform network fault early warning according to the final influence factor.
It can be appreciated that the network fault prediction method and the system have the same beneficial effects, and are not described herein.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:
fig. 1 is a flowchart of a network failure prediction method according to an embodiment of the present invention;
fig. 2 is a functional block diagram of a network failure prediction system according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other. The invention will be described in detail below with reference to the drawings in connection with embodiments.
In some embodiments of the present application, referring to fig. 1, a network failure prediction method includes:
And S100, intercepting network data in a preset period, and analyzing the network data to construct a network fluctuation-moment data chain.
And S200, comparing the network fluctuation value of each time window in the network fluctuation-moment data chain with the fluctuation range, and splitting the network fluctuation-moment data chain into an abnormal fluctuation data chain and a normal fluctuation data chain according to the comparison result.
And S300, extracting network characteristics of each time window in the abnormal fluctuation data chain, carrying out cluster analysis on the network data of each time window in the abnormal fluctuation data chain according to the network characteristics, constructing the same type network array, and obtaining wavelet influence factors of the same type network array according to the same type network array and the corresponding historical same type fluctuation times.
S400, analyzing the network sequences of the same type remained in the abnormal fluctuation data chain, determining corresponding wavelet influence factors, and obtaining a fluctuation influence factor set according to all the wavelet influence factors.
And S500, integrating the fluctuation influence factor set to obtain a final influence factor of the network data, and judging whether to perform network fault early warning according to the final influence factor.
Specifically, in S100, network data within a certain period of time is intercepted from the latest network data, and a plurality of network characteristics (such as bandwidth, delay, packet loss rate, etc.) are covered. By analyzing the data, the network data of each time window are organized in time sequence to form a 'network fluctuation-time data chain'. And S200, identifying which fluctuation belongs to abnormal fluctuation by comparing the network fluctuation values of each time window. The comparison process determines whether the network fluctuation is within a normal range by comparing with the fluctuation range. And splitting the data chain into an abnormal fluctuation data chain and a normal fluctuation data chain according to the comparison result so as to analyze potential faults in the network. And S300, extracting key network characteristics from an abnormal fluctuation data chain, classifying the data of each time window by a cluster analysis method, and classifying the moments with similar fluctuation modes into the same class. And constructing 'network number columns of the same type', and calculating 'wavelet influence factors' of each network number column by combining the historical same type fluctuation times. These factors reflect the potential impact of each type of surge pattern on network failure prediction. And S400, further analyzing the remaining network sequences of the same type, calculating the corresponding wavelet influence factors, and integrating all the wavelet influence factors into a fluctuation influence factor set. The set contains the influence factors of various fluctuation modes, and reflects the complexity and diversity of network fluctuation. And S500, integrating all the sub-wavelets to obtain a comprehensive final influence factor. The factor reflects the current network condition, and it is determined whether the network has a fault risk by the factor. If the final influence factor reaches a certain threshold, triggering network fault early warning and timely notifying network operation and maintenance personnel to process.
It can be understood that the normal network fluctuation and the potential faults are distinguished through the embodiment, and the defect that the traditional method is easy to misjudge is avoided. By carrying out fine splitting and analysis on the network fluctuation data, not only the instantaneous characteristics of the network fluctuation are considered, but also the historical data and the fluctuation mode are combined. Compared with the prior art, the method and the device can deal with complex and changeable network environments, monitor network conditions in real time, accurately predict potential risks before faults occur, and therefore achieve more effective network fault early warning. By adopting the method, preventive measures can be taken before faults occur, the reliability and stability of the network are improved, and the risk of service interruption is reduced.
In some embodiments of the present application, when analyzing network data to construct a network fluctuation-time data chain, the method includes acquiring a first blank data chain and a second blank data chain, wherein each blank data chain is provided with a plurality of data chain nodes and connection nodes. And acquiring the network fluctuation value of each time window according to the network data, and generating a chain-changing mark for the network data with all the network fluctuation values not being zero. And transferring all network data carrying the chain change mark to a first blank data chain to obtain a network fluctuation-moment data chain.
It will be appreciated that two "blank data chains" are first created, a first blank data chain and a second blank data chain, respectively. These data chains do not have any initial data themselves, mainly as containers for subsequent storage of network fluctuation data. Each data chain is provided with a plurality of data chain nodes and connecting nodes, and the nodes provide structural support for subsequent data storage and fluctuation analysis. Network data (such as bandwidth, delay, packet loss rate and the like) are acquired in real time, and a network fluctuation value of each time window is calculated. The fluctuation value reflects the performance state of the network at each time window. The moment when the network fluctuation value is not zero means that the network state changes, and fluctuation or abnormality occurs. By introducing blank data chains and chain change marks, network fluctuation data can be dynamically transferred between two independent data chains, normal fluctuation data and abnormal fluctuation data are prevented from being mixed in one data chain, and the accuracy of subsequent analysis is ensured. By generating the chain change mark, abnormal fluctuation data can be automatically marked and processed, and the efficiency and the accuracy of abnormal fluctuation identification are improved. Valuable information can be automatically screened out from network data, and high-quality data support is provided for subsequent fluctuation mode analysis, cluster analysis and fault prediction.
In some embodiments of the present application, when the network fluctuation value of each time window in the network fluctuation-time data chain is compared with the fluctuation range, the fluctuation range is obtained by:
Collecting a historical operation data set, and extracting a non-fault network fluctuation set from the historical operation data set;
carrying out standardization processing on the non-fault network fluctuation centralized data, wherein the standardization processing comprises Min-Max standardization;
Dividing the normalized non-fault network fluctuation set into a training set and a testing set according to the ratio of 4:1, training an LSTM model, and obtaining the predicted value of each historical time window;
Obtaining an actual value of each historical time window according to a non-fault network fluctuation set, and obtaining a predicted deviation multiplying power of each historical time window according to a predicted value of each historical time window and the actual value of each historical time window, wherein the predicted deviation multiplying power is a ratio of the predicted value to the actual value;
Dividing data in a network fluctuation-moment data chain according to time windows, taking each time window as an input of an LSTM model, obtaining an output result of the LSTM model, combining the output result with a prediction deviation multiplying power of a corresponding historical time window, and determining a fluctuation range of each time window.
Specifically, prior to using the LSTM model, a historical operating dataset is collected and a non-faulty network fluctuation set is extracted from the historical operating dataset, the network fluctuation-related data including delay, bandwidth, packet loss rate, throughput, etc. And (3) carrying out standardization processing on the extracted non-fault network fluctuation concentrated data, and scaling the data to the range of [0,1] by adopting Min-Max standardization. The dataset is split into a training set and a testing set. The training set is used for model training, and the input layer inputs the original time series data into the LSTM model. One or more LSTM layers are used to learn the time dependence in the data. Multiple LSTM cells and stacked LSTM layers are used to capture the time dependence of different hierarchies. The output of the LSTM layer is connected to a fully connected Dense layer, and finally the prediction result is output. At the output layer, a linear activation function may be used. The LSTM model is trained using the training set data. During training, an optimization algorithm (such as Adam or SGD) is selected, and the super parameters (such as learning rate, LSTM layer unit number, time window length, etc.) of the model are adjusted through cross validation. The trained models are evaluated using a test set, such as mean square error and root mean square error. After training the model, obtaining a predicted value of each historical time window, obtaining a predicted deviation multiplying power according to the actual value, dividing the current network fluctuation-moment data chain data according to the time window, taking each time window as the input of an LSTM model, obtaining a predicted result, and taking the product of the predicted result and the predicted deviation multiplying power of the corresponding historical time window as a fluctuation range.
It can be understood that accuracy and instantaneity of network fault prediction are improved through an LSTM model, dynamic fluctuation range adjustment, normalization processing and prediction deviation multiplying power. The interference of fault data on training is reduced, and the accuracy of model prediction is improved. Dynamically adjusting the fluctuation range so that the prediction is more in line with the actual network fluctuation trend. By introducing the predicted deviation multiplying power, the calculation of the fluctuation range is optimized, and the flexibility and the adaptability of the system are enhanced. The network fault early warning method and the network fault early warning device achieve more accurate network fault early warning, and improve stability and usability of the network.
In some embodiments of the present application, when the network data of each time window in the abnormal fluctuation data chain is subjected to cluster analysis according to the network characteristics, the network characteristics include throughput, packet loss rate and network traffic load.
Specifically, an initialization neighborhood radius is determined by a k-distance map, and MinPts is determined to be 6. And taking each network characteristic as one point, scanning all points, and finding out the point with the point number greater than or equal to MinPts in the neighborhood as a core point. Starting from each core point, points within its neighborhood are checked. If the point in the neighborhood is the core point, the cluster continues to be expanded. If the point in the neighborhood is a boundary point, it is added to the current cluster. If one point is not a point within the neighborhood of any core point and cannot form clusters with other points, it is marked as a noise point.
It can be appreciated that by applying the density clustering algorithm, classification of abnormal fluctuation data is achieved in combination with network characteristics (throughput, packet loss rate, network traffic load). Compared with the traditional clustering method, the DBSCAN can automatically identify different fluctuation modes and effectively process noise and outliers in network data. The number of clusters is not required to be preset, and the problem of excessive clustering or insufficient clustering possibly occurring in the traditional method is avoided. By such dynamic cluster analysis, abnormal fluctuations can be more accurately identified and distinguished.
In some embodiments of the present application, when obtaining the wavelet impact factor of the same type of network array according to the same type of network array and the corresponding historical same type of fluctuation times, the method includes:
wherein Gz represents a wavelet influence factor, N represents the number of network fluctuation values in the same type network sequence, bi represents an i-th network fluctuation value in the same type network sequence, C represents the number of historical fluctuation times, and B0 represents a right boundary value.
It can be understood that the sub-fluctuation influence factor is calculated by combining various information such as fluctuation data, historical fluctuation times and right boundary values in the same type network array, so that more quantitative and accurate basis is provided for network fault prediction. Compared with the traditional fault prediction method, the method can deeply analyze the historical rule of network fluctuation and the abnormality degree of current fluctuation, thereby providing an effective early warning mechanism. By introducing the influence of the historical fluctuation times, which fluctuation types have higher fault risks can be identified, so that more accurate network fault pre-judgment is assisted, and the loss caused by network faults is reduced.
In some embodiments of the present application, deriving the set of fluctuation influencing factors from all wavelet influence factors includes determining a median and variance of all wavelet influence factors. Extracting wavelet influence factors greater than the median from all the wavelet influence factors to construct a first data set. Extracting wavelet influence factors larger than the variance from all wavelet influence factors to construct a second data set.
Specifically, it is determined whether an intersection exists between the first data set and the second data set. If yes, the intersection value is used as a fluctuation influence factor set. If not, carrying out non-repeated fusion on the first data set and the second data set to construct a fluctuation influence factor set, wherein the non-repeated fusion is to reserve non-repeated wavelet influence factors in the first data set and the second data set, reserve one repeated sub-fluctuation influence factor in the first data set and the second data set, and delete the rest repeated sub-fluctuation influence factors.
In some embodiments of the present application, integrating the set of fluctuation influencing factors to obtain a final influencing factor of the network data includes comparing the set of fluctuation influencing factors with historical judgment data, the historical judgment data including the set of historical fluctuation influencing factors and the historical final influencing factor.
Specifically, when data with the similarity of the fluctuation influence factor set being larger than a similarity threshold exists in the history judgment data, a history fluctuation influence factor set corresponding to the maximum value of the similarity is selected, and a history final influence factor corresponding to the history fluctuation influence factor set is used as a final influence factor. And when the similarity between the historical fluctuation influence factor set and the fluctuation influence factor set in the historical judgment data is smaller than or equal to a similarity threshold value, determining a final influence factor according to the fluctuation influence factor set.
In some embodiments of the application, determining the final influence factor from the set of fluctuating influence factors comprises:
wherein Y represents the final influence factor, M represents the number of the fluctuation influence factors in the fluctuation influence factor set, Representing the weight corresponding to the i < th > sub-fluctuation influencing factor,Representing the i-th sub-fluctuation influencing factor,Representing the smallest sub-fluctuation influencing factor,For the maximum wavelet impact factor,For all ofIs the maximum value of (a).
It can be understood that the weight corresponding to the wavelet impact factor has a weight range of (0, 1), and the weight is in direct proportion to the value of the wavelet impact factor. By comprehensively analyzing the fluctuation influence factor set and combining the historical data with the current fluctuation mode, a more accurate fault prediction basis can be provided. Through data extraction and fusion, similarity comparison and weighted calculation methods, the limitation that historical data or current data are simply relied on is avoided, and final fault influence factor calculation is more comprehensive and accurate. The data fusion and deduplication part can effectively reduce the interference of redundant data, and ensures that the fluctuation factors with the most representativeness and influence are contained in the final influence factor set. By comparing the network fault early warning method with the historical data, the reliability of prediction is further improved, and the network fault early warning method can be better adapted to the dynamic change of the network, so that the network fault early warning is more efficient and accurate.
In some embodiments of the application, judging whether to perform network fault early warning according to the final influence factor comprises comparing the final influence factor with an influence factor threshold value, and judging whether to perform network fault early warning according to the comparison result.
Specifically, when the final influence factor is larger than the influence factor threshold, network fault early warning is judged, the difference value between the final influence factor and the influence factor threshold is obtained, the network fault early warning level is determined according to the difference value, and the network fault early warning level and the difference value are in a proportional relation.
It can be understood that by comparing the final influence factor with a preset influence factor threshold, whether the network needs to trigger fault early warning or not is judged, the dynamic change of the network can be responded in real time, and the early warning level can be flexibly determined according to the severity of fluctuation. The proportional relation between the difference value and the early warning level provides a refined early warning system. And when the difference value is smaller, providing low-level early warning to help an administrator know the potential risk in time. The method is favorable for taking measures in advance, preventing network faults, and further guaranteeing the continuity and stability of the service.
In the embodiment, the dynamic change of the network can be captured by intercepting the network data in the preset period and constructing the fluctuation-time data chain, so that the problem that the traditional method only depends on a single network index is avoided. By comparing the fluctuation range, the data chain is split into an abnormal fluctuation data chain and a normal fluctuation data chain, and normal fluctuation and potential faults are effectively distinguished, so that the fault early warning precision is improved. Network characteristics in the abnormal fluctuation data chain are extracted and cluster analysis is carried out, so that different types of fluctuation modes are further refined, and the fluctuation influence factors of the operators are calculated by combining the times of the historical fluctuation of the same type, so that the fault prediction is more accurate. By integrating the fluctuation influence factor sets, final influence factors of network data are generated, whether the network has fault risks or not can be accurately judged, early warning can be timely sent out, and the efficiency and response speed of network operation and maintenance are improved.
In another preferred manner based on the foregoing embodiment, referring to fig. 2, the present embodiment provides a network failure prediction system, configured to apply the foregoing network failure prediction method, including:
The acquisition unit is configured to intercept the network data in a preset period, analyze the network data and construct a network fluctuation-moment data chain;
The processing unit is configured to compare the network fluctuation value of each time window in the network fluctuation-time data chain with the fluctuation range, and split the network fluctuation-time data chain into an abnormal fluctuation data chain and a normal fluctuation data chain according to the comparison result;
The execution unit is configured to extract the network characteristics of each time window in the abnormal fluctuation data chain, perform cluster analysis on the network data of each time window in the abnormal fluctuation data chain according to the network characteristics, construct the same type of network array, and obtain the sub-wave influence factors of the same type of network array according to the same type of network array and the corresponding historical same type fluctuation times;
the analysis unit is configured to analyze the network sequences of the same type remained in the abnormal fluctuation data chain, determine the corresponding wavelet influence factors and obtain a fluctuation influence factor set according to all the wavelet influence factors;
and the early warning unit is configured to integrate the fluctuation influence factor set to obtain a final influence factor of the network data, and judge whether to perform network fault early warning according to the final influence factor.
It can be understood that by intercepting network data within a preset period and constructing a fluctuation-time data chain, dynamic changes of the network can be captured, and the problem that the traditional method only depends on a single network index is avoided. By comparing the fluctuation range, the data chain is split into an abnormal fluctuation data chain and a normal fluctuation data chain, and normal fluctuation and potential faults are effectively distinguished, so that the fault early warning precision is improved. Network characteristics in the abnormal fluctuation data chain are extracted and cluster analysis is carried out, so that different types of fluctuation modes are further refined, and the fluctuation influence factors of the operators are calculated by combining the times of the historical fluctuation of the same type, so that the fault prediction is more accurate. By integrating the fluctuation influence factor sets, final influence factors of network data are generated, whether the network has fault risks or not can be accurately judged, early warning can be timely sent out, and the efficiency and response speed of network operation and maintenance are improved.
It will be appreciated by those skilled in the art that embodiments of the application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flowchart and/or block of the flowchart illustrations and/or block diagrams, and combinations of flowcharts and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the specific embodiments of the present invention without departing from the spirit and scope of the present invention, and any modifications and equivalents are intended to be included in the scope of the claims of the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510238401.9A CN119728397B (en) | 2025-03-03 | 2025-03-03 | Network fault prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510238401.9A CN119728397B (en) | 2025-03-03 | 2025-03-03 | Network fault prediction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119728397A CN119728397A (en) | 2025-03-28 |
| CN119728397B true CN119728397B (en) | 2025-05-02 |
Family
ID=95091227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510238401.9A Active CN119728397B (en) | 2025-03-03 | 2025-03-03 | Network fault prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119728397B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119882839A (en) * | 2025-03-24 | 2025-04-25 | 中家院（北京）检测认证有限公司 | High-performance display horizontal visual angle test and adjustment system |
| CN120492204B (en) * | 2025-05-07 | 2025-10-31 | 准检河北检测技术服务有限公司 | An automatic evaluation method for the accuracy of defect identification software |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113590370A (en) * | 2021-08-06 | 2021-11-02 | 北京百度网讯科技有限公司 | Fault processing method, device, equipment and storage medium |
| CN118075791A (en) * | 2024-04-17 | 2024-05-24 | 深圳市兴意腾科技电子有限公司 | Wireless communication network fault prediction method and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20240163217A (en) * | 2023-05-09 | 2024-11-19 | 포항공과대학교 산학협력단 | Method and apparatus of virtual machine failure prediction using machine learning |
| CN118861683A (en) * | 2024-07-05 | 2024-10-29 | 广东电网有限责任公司 | Power grid historical message processing method, device and power grid monitoring system |
-
2025
- 2025-03-03 CN CN202510238401.9A patent/CN119728397B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113590370A (en) * | 2021-08-06 | 2021-11-02 | 北京百度网讯科技有限公司 | Fault processing method, device, equipment and storage medium |
| CN118075791A (en) * | 2024-04-17 | 2024-05-24 | 深圳市兴意腾科技电子有限公司 | Wireless communication network fault prediction method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119728397A (en) | 2025-03-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111459700B (en) | Equipment fault diagnosis method, diagnosis device, diagnosis equipment and storage medium | |
| CN119728397B (en) | Network fault prediction method and system | |
| CN117743909B (en) | A method and device for analyzing heating system failure based on artificial intelligence | |
| CN109800127A (en) | A kind of system fault diagnosis intelligence O&M method and system based on machine learning | |
| CN118313812A (en) | A method for collecting and processing power big data based on machine learning | |
| CN119004022A (en) | Fault state analysis method and system based on panoramic modeling of wind farm | |
| CN112540905A (en) | System risk assessment method, device, equipment and medium under micro-service architecture | |
| CN118713991A (en) | A fault prediction method and system based on observable data | |
| CN113395182B (en) | Intelligent network equipment management system and method with fault prediction | |
| CN118779798B (en) | Power edge optimization control method and system based on IoT cloud-edge collaboration | |
| CN120179435A (en) | A cluster reliability testing method and system based on fault simulation | |
| CN119557134A (en) | Fault handling method, device, electronic device and storage medium for cloud computing platform | |
| CN117200449A (en) | A power grid monitoring and management method and system based on multi-dimensional algorithm analysis | |
| CN120336190A (en) | An integrated test scenario prediction method and system based on multi-dimensional data | |
| CN119618332A (en) | Intelligent water meter remote monitoring method and equipment integrating NB-IoT | |
| CN119988240A (en) | Test risk identification method and system based on artificial intelligence | |
| CN120216237A (en) | Fault detection method, device and electronic equipment | |
| CN117574303B (en) | Construction condition monitoring and early warning method, device, equipment and storage medium | |
| CN119248617A (en) | Elasticsearch cluster fault intelligent detection system | |
| CN119884540A (en) | Dynamic truck scale metering method with automatic parameter fitting function | |
| CN118075090A (en) | Network fault prediction method based on machine learning | |
| CN116820884A (en) | A method and device for monitoring abnormal status of IT systems using intelligent operation and maintenance | |
| CN114637793A (en) | Equipment fault frequent region positioning method based on big data analysis | |
| CN109474445B (en) | Distributed system root fault positioning method and device | |
| CN118944286B (en) | Full-flow data monitoring method for power distribution network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |