Redundancy damage detection method and device of redundancy network and electronic equipment

Abstract

The invention relates to a redundancy damage detection method and device for a redundancy network and electronic equipment, and the method comprises the steps: obtaining data synchronously collected by a plurality of network nodes, carrying out the data preprocessing and multi-source data fusion of the collected data, and creating a data set containing a plurality of network state indexes; identifying the dynamic change of the network state index, and outputting the prediction result of the future network state; carrying out lightweight calculation processing on the network edge node, and outputting a calculation processing result; and in response to the fact that the calculation processing result exceeds the normal range, instructing a network route adjustment strategy or triggering an early warning mechanism. According to the method, data preprocessing and multi-source data fusion are performed on the collected data, flow data smoothing and data set integration of different nodes are facilitated, the detection standard is dynamically adjusted by capturing the dynamic change of the network state in real time, the method can adapt to the dynamic network environment, the detection result is improved, and the detection efficiency is improved. And the stability and the reliability of the multi-redundancy network are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of redundant networks, and in particular to a method, a device and an electronic device for detecting redundancy damage of a redundant network. Background Art

[0002] In the current communication network field, the adaptive function of redundant networks has become an important technology. These networks can automatically adjust routing and transmission parameters according to real-time traffic and channel quality to optimize bandwidth allocation and signal transmission quality. However, this dynamic change brings challenges to redundancy damage detection, because the reference standard for detection needs to be continuously updated as the network status changes. Traditional detection methods are often unable to adapt to such rapid changes, resulting in inaccurate detection results and even failure to detect network damage in time, thus affecting the stability and reliability of the network. Therefore, it is necessary to improve them. Summary of the invention

[0003] Based on this, it is necessary to provide a redundancy damage detection method, device and electronic equipment for a redundant network to address the problem that traditional detection methods are often unable to adapt to such rapid changes, resulting in inaccurate detection results and even failure to promptly detect network damage, thereby affecting the stability and reliability of the network.

[0004] The present invention provides a redundancy damage detection method for a redundant network, the method comprising:

[0005] Acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of the network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength, and network utilization;

[0006] Identify dynamic changes in network status indicators and output prediction results for future network status;

[0007] Perform lightweight computing on network edge nodes and output the computing results;

[0008] In response to the result of the computing process being outside the normal range, the network routing is instructed to adjust the strategy or trigger an early warning mechanism.

[0009] In one embodiment, the preprocessing of the collected data includes:

[0010] Clean and standardize the data;

[0011] And / or, extracting features useful for damage detection from the data, wherein the features include statistical features, frequency domain features, time domain features, waveform features, and machine learning features.

[0012] In one embodiment, the cleaning and standardizing of data includes:

[0013] Delete data not relevant to damage detection;

[0014] Fill in missing values ​​in the data set or delete corresponding data rows;

[0015] Eliminate random noise in data according to filtering algorithms;

[0016] Identify outliers in the data and delete or correct them;

[0017] Standardize the data and convert it to a uniform scale.

[0018] In one embodiment, the multi-source data fusion of the collected data includes:

[0019] Output the pre-processed data from different devices to a central database and associate them through the key fields of device ID and timestamp;

[0020] Or, data frames from different devices are spliced ​​horizontally or vertically according to timestamps or device IDs;

[0021] Alternatively, additional features are added to the data set of each device, wherein the additional features include device type and location information.

[0022] In one embodiment, identifying the dynamic changes of the network status indicator and outputting the prediction result of the future network status includes:

[0023] Obtain a dataset containing various network status indicators;

[0024] The data set is input into an advanced machine learning model to output a prediction result of the future network state. The advanced machine learning model is based on reinforcement learning and adaptive neural network to monitor and predict the network state in real time.

[0025] In one of the embodiments, the adaptive neural network uses a long short-term memory network to capture long-term dependencies and short-term dynamic changes of network states.

[0026] In one embodiment, performing lightweight computing processing on the network edge node and outputting the computing processing result includes:

[0027] Obtain traffic data and environmental parameters collected by network devices in real time;

[0028] The data is preprocessed and preliminarily analyzed according to the algorithm built into the computing unit, and the results of the preliminary analysis are output.

[0029] The present invention also provides a redundancy damage detection device for a redundant network, comprising:

[0030] An acquisition module is used to acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of the network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength and network utilization;

[0031] The identification module is used to identify the dynamic changes of network status indicators and output the prediction results of future network status;

[0032] A processing module, used to perform lightweight computing processing on network edge nodes and output the computing processing results;

[0033] The response module is used to instruct the network routing to adjust the strategy or trigger the early warning mechanism in response to the result of the calculation processing exceeding the normal range.

[0034] The present invention also provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the redundancy damage detection method for a redundant network as described in any one of the above is implemented.

[0035] The present invention also provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the redundancy damage detection method for a redundant network as described in any one of the above is implemented.

[0036] The above-mentioned redundancy damage detection method, device and electronic equipment for the redundant network, by performing data preprocessing and multi-source data fusion on the collected data, facilitate the smoothing of traffic data and the integration of data sets of different nodes, dynamically adjust the detection standards by capturing the dynamic changes of the network status in real time, can adapt to the dynamic network environment, improve the detection results, and effectively improve the stability and reliability of the redundant network. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 A schematic flow chart of a method for detecting redundancy damage in a redundant network in one embodiment;

[0039] Figure 2 A schematic diagram of a data preprocessing process in an embodiment;

[0040] Figure 3 A schematic diagram of a process for cleaning and standardizing data in an embodiment;

[0041] Figure 4 A schematic diagram of a process of multi-source data fusion in one embodiment;

[0042] Figure 5 A schematic diagram of a network status prediction process in one embodiment;

[0043] Figure 6 A schematic diagram of a process of performing lightweight computing processing on a network edge node in an embodiment;

[0044] Figure 7 It is a structural schematic diagram of a redundancy damage detection device for a redundant network in one embodiment;

[0045] Figure 8 is a flow data table arranged in chronological order in one embodiment;

[0046] Fig. 9 is a smoothed flow data table in an embodiment;

[0047] Fig.10 The figure is a diagram showing the internal structure of an electronic device according to an embodiment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 creative work are within the scope of protection of the present invention.

[0049] Combine the following Figure 1-Figure 10 The present invention describes a redundancy damage detection method, device and electronic equipment for a redundant network.

[0050] like Figure 1 As shown, in one embodiment, a redundancy damage detection method for a redundant network includes the following steps:

[0051] Step S100, obtain data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength and network utilization.

[0052] By preprocessing the collected data and fusing multi-source data, it is convenient to smooth the flow data and integrate the data sets of different nodes.

[0053] Multi-source data, for example: traffic statistics of router A; traffic statistics of switch B; channel quality indicators of router A and switch B.

[0054] Among them, network delay is the time required for a data packet to travel from one node to another; packet loss rate is the ratio of data packets lost during transmission; error rate is the ratio of errors occurring during data transmission; signal strength is an indicator of the strength of the wireless signal; and network utilization is the usage of network bandwidth.

[0055] Step S200, identifying dynamic changes in network status indicators and outputting prediction results of future network status.

[0056] It is convenient to dynamically adjust the detection standards and capture dynamic changes in network status.

[0057] Step S300, performing lightweight computing processing on the network edge node and outputting the computing processing result.

[0058] It is convenient to realize real-time analysis and rapid response of data.

[0059] Step S400, in response to the result of the calculation processing exceeding the normal range, instructing the network routing to adjust the strategy or triggering the early warning mechanism.

[0060] The automated response strategy ensures that actions can be taken quickly and automatically when redundancy impairment is detected. The core of this strategy is to monitor the output of the redundancy impairment detection system in real time and automatically adjust the network routing based on this information. The specific steps are as follows:

[0061] 1. Real-time monitoring of detection results: Design a monitoring module that continuously receives the output results of the redundancy impairment detection system. These results include the health status of each path in the network, such as the packet loss rate, delay and bandwidth usage of the channel.

[0062] 2. Route adjustment decision making: When the monitoring module detects that the damage index of a path exceeds the preset threshold, it triggers the route adjustment decision process. This process involves a comprehensive evaluation of the current network status and calculating the optimal route adjustment solution.

[0063] 3. Automatically perform routing adjustments: After the decision is made, the automated response policy system interacts with the network control layer to automatically perform routing adjustments. For example, if the channel quality of a path degrades, the system will instruct the router to transfer traffic to other healthier paths.

[0064] Specific application scenario examples are as follows:

[0065] Suppose that during network operation, the detection system finds that the packet loss rate of a certain path suddenly increases and exceeds the normal operating range. The monitoring module immediately captures this change and starts the route adjustment decision process. The system evaluates the status of all optional paths and selects the path that is currently the least congested as an alternative route. Subsequently, the system automatically sends instructions to the relevant routers to transfer traffic from the damaged path to the alternative path. This fast and automatic response mechanism significantly improves the stability and reliability of the network.

[0066] The relevant mathematical formula is described as follows:

[0067] P i >T, trigger routing adjustment, R j =min(R1,R2,…,R n ), select R j The smallest path is used as an alternative path.

[0068] Among them, P i represents the packet loss rate of the ith path; T represents the preset packet loss rate threshold; when P i >T, trigger routing adjustment; set R j is the bandwidth utilization of the jth optional path, select R j The smallest path is used as an alternative route.

[0069] Through the above formulas and strategies, the automated response system can respond quickly when network damage occurs to ensure network continuity and service quality.

[0070] The redundancy damage detection method of the redundant network performs data preprocessing and multi-source data fusion on the collected data, facilitates the smoothing of traffic data and the integration of data sets of different nodes, dynamically adjusts the detection standard by capturing the dynamic changes of the network status in real time, can adapt to the dynamic network environment, improves the detection results, and effectively improves the stability and reliability of the redundant network.

[0071] In this embodiment, see Figure 2 , the collected data is preprocessed, including the following steps:

[0072] Step S110, cleaning and standardizing the data.

[0073] By cleaning and standardizing the data, it is convenient to ensure the accuracy and efficiency of subsequent analysis.

[0074] And / or, step S120, extracting features useful for damage detection in the data, the features including statistical features, frequency domain features, time domain features, waveform features and machine learning features.

[0075] The feature extraction is related to the selected fusion scheme. If the fusion scheme requires features, the preprocessing in this step may include feature extraction. Otherwise, feature extraction is not required.

[0076] Among them, statistical features include mean, median, standard deviation, minimum and maximum values, which can be directly calculated from the original data; frequency domain features can obtain features in the frequency domain, such as power spectrum density, by performing fast Fourier transform (FFT) on the data; time domain features such as autocorrelation function, cross-correlation function, etc. can be used to analyze the correlation of data at different time points; waveform features such as waveform entropy, kurtosis, skewness, etc. of the signal can describe the shape characteristics of the signal; machine learning features are extracted by using machine learning algorithms (such as principal component analysis PCA) to extract the main features in the data.

[0077] These features are very useful for damage detection because they reveal possible problems in the network, allowing for deeper analysis and repair.

[0078] In this embodiment, see Figure 3 , clean and standardize the data, including the following steps:

[0079] Step S111, deleting data irrelevant to damage detection.

[0080] For example: delete some non-critical log information.

[0081] Step S112, fill in the missing values ​​in the data set or delete the corresponding data rows.

[0082] For example: filling missing values ​​in a dataset using the mean, median, or using a predictive model.

[0083] Step S113, eliminating random noise in the data according to a filtering algorithm.

[0084] This method can effectively reduce short-term data deviations caused by network fluctuations and provide a stable data basis for subsequent analysis.

[0085] For example, to monitor the traffic data in the network in real time, a moving average filter is used to process the data. This filter smoothes the short-term data fluctuations by calculating the average value of the data within a certain time window.

[0086] The moving average filter is an algorithm commonly used in signal processing to smooth time series data, reducing the impact of short-term fluctuations and highlighting long-term trends.

[0087] Here is an example of using a moving average filter to smooth traffic data:

[0088] See also Figure 8 ,Assume there is a set of traffic data arranged in time order.

[0089] Apply a simple arithmetic moving average to smooth the data. For example, to calculate the moving average of 3 data points (i.e. a time series with a window size of 3):

[0090] For the data at time point 3, the moving average is (100+110+120) / 3=110, for the data at time point 4, the moving average is (110+120+115) / 3=116.67, for the data at time point 5, the moving average is (120+115+105) / 3=112.33, and so on.

[0091] After applying the moving average filter, the smoothed flow data is shown in Fig. 9 .

[0092] It should be added that the moving average filter will result in a loss of time resolution of the data points, since each moving average is calculated based on the previous data points. In addition, for the beginning of the series (the first two data points in this example), the moving average cannot be calculated because there are not enough preceding data points.

[0093] Step S114, identifying outliers in the data, and deleting or correcting the outliers.

[0094] By deleting or correcting outliers, the impact of outliers on analysis results can be avoided and data quality can be ensured.

[0095] Step S115, standardize the data and convert the data to a uniform scale.

[0096] By formatting the data, it is easier to compare and integrate data from different sources.

[0097] In this embodiment, see Figure 4 , multi-source data fusion is performed on the collected data, including the following steps:

[0098] Step S130, output the pre-processed data of different devices to a central database and associate them through the key fields of device ID and timestamp.

[0099] By merging databases, data management and analysis are facilitated, and data flexibility and adaptability are improved.

[0100] Or, in step S140, data frames of different devices are spliced ​​horizontally or vertically according to timestamps or device IDs.

[0101] In data processing software (such as the Pandas library), the risk of data loss during network transmission is reduced by splicing data frames, while improving network transmission efficiency and reducing latency.

[0102] Or, in step S150, additional features are added to the data set of each device, where the additional features include device type and location information.

[0103] It is convenient to distinguish data from different devices in a unified data set.

[0104] For example, network traffic data for a certain period of time is collected and preprocessed, and then combined with channel quality indicators collected from other nodes to create a data set containing multiple network status indicators, providing comprehensive data support for subsequent redundancy damage detection.

[0105] The specific network status analysis process is as follows:

[0106] 1. Traffic analysis: By comparing the traffic data of different devices, traffic hotspots and bottlenecks in the network can be identified. For example, if the ingress traffic of a switch is much larger than the egress traffic, it may indicate that a large amount of data is generated or aggregated near the device.

[0107] 2. Load balancing: Analyzing the load conditions of different network devices can ensure that network resources are effectively utilized and avoid performance degradation caused by overload.

[0108] 3. Fault detection: A sudden increase or decrease in traffic may indicate a fault in the network, such as a link failure or device failure. By monitoring these indicators in real time, the problem can be quickly located.

[0109] 4. User experience: Channel quality indicators (such as packet loss rate and latency) are directly related to the user's network experience. Unified data analysis helps to evaluate and improve the user experience.

[0110] 5. Security monitoring: Abnormal traffic patterns may be signs of network attacks, such as DDoS attacks. Multi-dimensional data analysis can detect potential security threats in a timely manner.

[0111] 6. Trend prediction: By analyzing historical data, we can predict the future trend of the network and provide data support for network planning.

[0112] For example, suppose we find that the egress traffic of a router suddenly drops during a certain period of time, while the ingress traffic of the switch does not decrease accordingly. Combined with the channel quality indicators, if we find that the packet loss rate and latency increase, we can infer that there may be a link failure or a performance problem with the router. Such multi-dimensional analysis helps network administrators take quick action to restore normal network operation.

[0113] In this embodiment, see Figure 5 , identify the dynamic changes of network status indicators and output the prediction results of future network status, including the following steps:

[0114] Step S210, obtaining a data set including multiple network status indicators.

[0115] It should be noted that these data sets are pre-processed and fused with multi-source data. By integrating these data into a unified data set, we can analyze the network status from multiple dimensions.

[0116] Step S220, input the data set into an advanced machine learning model and output a prediction result of the future network status. The advanced machine learning model is based on reinforcement learning and adaptive neural network to monitor and predict the network status in real time.

[0117] It should be noted that the adaptive neural network uses a long short-term memory network to capture the long-term dependencies and short-term dynamic changes of the network state.

[0118] In redundancy impairment detection in redundant networks, applying advanced machine learning models is a key step, which enables the detection process to adapt to the dynamic changes of the network.

[0119] The reinforcement learning implementation details are as follows:

[0120] A reinforcement learning agent is designed that can dynamically adjust the detection criteria for redundancy impairments based on real-time feedback from the network (such as latency and packet loss rate). The agent optimizes its detection strategy by exploring different combinations of detection criteria and obtaining rewards based on improvements in network performance (such as reductions in impairment rates). This self-learning mechanism enables the detection criteria to be adaptively adjusted as the network environment changes.

[0121] The RL (Reinforcement Learning) model consists of an agent and an environment. The agent selects action A (detection criterion) based on the state S (network performance indicator) of the environment, and the environment returns a reward R (the degree of improvement in network performance).

[0122] The state (S) consists of the real-time performance indicators of the network, expressed as:

[0123] S=(L,P),

[0124] Among them, L represents delay and P represents packet loss rate.

[0125] Action (A) is the combination of detection criteria chosen by the agent, expressed as:

[0126] A=(T,F),

[0127] Where T is the damage threshold and F is the detection frequency.

[0128] The reward function (R) is calculated based on the improvement of network performance, for example, the reduction of damage rate. If the network performance improves, a positive reward is given, otherwise a negative reward is given. The reward function can be expressed as:

[0129]

[0130] The Q function is used to evaluate the expected return of taking a specific action in a specific state. A deep neural network is used to approximate this function, which is expressed as follows:

[0131] Q(S,A)≈NN(S,A),

[0132] Among them, NN represents a neural network function.

[0133] The policy (π) is the rule by which the agent chooses actions according to the Q function. We usually use the ε-greedy policy, where the agent chooses the best action with probability (1-∈), where ∈ usually represents a small positive number representing the difference in error or approximation; actions are chosen randomly with probability (∈).

[0134] The core update formula in the Q-learning algorithm is used to recursively calculate and update the Q value. The formula is as follows:

[0135] Q(S t ,A t )←Q(S t ,A t )+α(R t +γmaxA′Q(S t+1 ,A′)-Q(S t ,A t )),

[0136] Among them, Q(S t ,A t ) is in state S t Take action A t The Q value at that time; α is the learning rate, which determines the degree to which the new information adjusts the old Q value; R tis the reward received at time step t; γ is the discount factor used to weigh immediate rewards against future rewards; maxA′Q(S t+1 ,A′) is in the next state S t+1 Take the action A' with the maximum Q value among all possible actions.

[0137] The specific agent learning process is to continuously update the Q function through the following steps:

[0138] 1. Observe the current status (S t ).

[0139] 2. Select an action (A) based on the strategy (π) t ).

[0140] 3. Execute action (A t ), observe the new state S t+1 and reward (R t ).

[0141] 4. Update the Q function.

[0142] The specific implementation method is:

[0143] Initialization: Set the Q network, target network, experience replay buffer, learning rate, discount factor, and ε value of the ε-greedy strategy.

[0144] Loop execution: Observe the current network status (S t ); Use the ε-greedy strategy to select the detection standard (A t ); apply detection criteria, collect network feedback, and calculate rewards (R t ); t ,A t ,R t ,S t+1 ) is stored in the experience replay buffer; a batch of data is randomly extracted from the experience replay buffer to update the Q network; and the target network is updated regularly.

[0145] In this way, the agent is able to continuously learn and optimize its detection strategy to adapt to the changing network environment.

[0146] The implementation details of the adaptive neural network are as follows:

[0147] RNN (recursive neural network), especially LSTM (Long Short-Term Memory), is used to process time series data. This network structure can effectively capture the long-term dependencies and short-term dynamic changes of network states. When the network state changes, LSTM can automatically adjust its internal parameters to predict the future network state.

[0148] The LSTM network is a special RNN that effectively solves the gradient vanishing or gradient exploding problem of the standard RNN when dealing with long-term dependencies by introducing three gate structures (input gate, forget gate, and output gate).

[0149] Among them, the forget gate determines which information is discarded from the cell state, the input gate determines which new information is stored in the cell state, the cell state is the information flowing in the network, which can be updated, read or forgotten; the output gate determines what the next hidden state should be.

[0150] The core update formula of LSTM is as follows:

[0151] f t =σ(W f ·[h t-1 ,x t ]+b f ),

[0152] i t =σ(W i ·[h t-1 ,x t ]+b i ),

[0153]

[0154] O t =σ(W O ·[h t-1 ,x t ]+b O ),

[0155] h t =O t tanh(C t ),

[0156] Among them, f t is the output of the forget gate; i t is the input gate output; is the candidate cell state; C t is the updated cell state; O t is the output gate output; h tis the output of the LSTM unit; W f , W i , W C , W O is the weight matrix; b f , b i , b C , b O is the bias term; σ is the sigmoid activation function; x t is the input of the current time step; h t-1 is the hidden state at the previous time step.

[0157] The specific implementation steps are:

[0158] 1. Initialize the LSTM unit: set the initial values ​​of the weight matrix and bias term.

[0159] 2. Input sequence processing: Take the time series data as the input sequence x t Input to the LSTM cell.

[0160] 3. Gating mechanism: At each time step, the gating mechanism calculates the forget gate, input gate, candidate cell state, cell state, output gate, and hidden state.

[0161] 4. Predict output: Use the hidden state h of the last time step t Make predictions or classifications.

[0162] In the process of implementing reinforcement learning and adaptive neural networks, an optimal detection standard adjustment strategy is gradually found through continuous trial and error and learning by the reinforcement learning agent. At the same time, the adaptive neural network uses its powerful time series processing capabilities to monitor and predict the network status in real time, so that the model can accurately capture and respond to changes in the network status.

[0163] The following are some application scenarios:

[0164] When network traffic fluctuates abnormally, the reinforcement learning agent receives information about the change and tries to adjust the detection criteria. If this adjustment leads to improved network performance, the agent receives positive feedback and strengthens the adjustment strategy.

[0165] At the same time, the LSTM network analyzes the time series characteristics of traffic data and predicts future traffic patterns, so that the detection system can respond in advance. This linkage mechanism greatly improves the accuracy and real-time performance of redundancy damage detection.

[0166] In this embodiment, see Figure 6 , performing lightweight computing processing on the network edge nodes and outputting the computing processing results, including the following steps:

[0167] Step S310, obtaining the flow data and environmental parameters collected by the network equipment in real time.

[0168] By deploying lightweight computing units at the edge nodes of the network, data transmission delay can be reduced and response speed can be improved. These computing units have sufficient processing power to perform real-time analysis of the collected data, allowing changes in network status to be quickly identified at the source of data generation.

[0169] In terms of specific operations, hardware platforms with low energy consumption, small size and moderate computing power were selected. These platforms are capable of running complex data processing algorithms while maintaining real-time requirements.

[0170] By collecting traffic data and environmental parameters directly from network devices (such as routers and switches) through computing units, data transmission latency can be reduced.

[0171] Step S320, preprocessing and preliminary analysis of the data are performed according to the algorithm built into the computing unit, and the result of the preliminary analysis is output.

[0172] The data can be directly pre-processed and preliminarily analyzed through the algorithm placed in the computing unit, without sending the data to the central server, thus improving the response speed.

[0173] The application scenario examples are as follows:

[0174] When a network edge node detects a traffic surge, the computing unit deployed at the node immediately smooths and identifies patterns in the traffic data. Once it is determined that the traffic pattern is beyond the normal range, the computing unit quickly adjusts the routing strategy to distribute traffic to other less congested paths. This real-time local response mechanism significantly improves the overall performance and reliability of the network. The following is a detailed description of this process:

[0175] Example: Suppose an edge node detects a sharp increase in data traffic during the afternoon peak hours. The computing unit immediately processes the data, identifies the change in traffic patterns, and makes a decision within milliseconds to redirect part of the traffic to an alternate path. This not only alleviates congestion on the original path, but also ensures the quality of network services.

[0176] The redundancy damage detection device for a redundancy network provided by the present invention is described below. The redundancy damage detection device for a redundancy network described below and the redundancy damage detection method for a redundancy network described above can be referred to each other.

[0177] like Figure 7 As shown, in one embodiment, a redundancy damage detection device for a redundant network includes an acquisition module 710 , an identification module 720 , a processing module 730 and a response module 740 .

[0178] The acquisition module 710 is used to acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength and network utilization.

[0179] The identification module 720 is used to identify the dynamic changes of the network status indicators and output the prediction results of the future network status.

[0180] The processing module 730 is used to perform lightweight computing processing on the network edge node and output the computing processing result.

[0181] The response module 740 is used to instruct the network routing to adjust the strategy or trigger the early warning mechanism in response to the result of the calculation processing exceeding the normal range.

[0182] In this embodiment, data preprocessing is performed on the collected data, specifically for:

[0183] Clean and standardize the data;

[0184] And / or, extracting features useful for damage detection from the data, the features including statistical features, frequency domain features, time domain features, waveform features, and machine learning features.

[0185] In this embodiment, the data is cleaned and standardized, specifically for:

[0186] Delete data not relevant to damage detection;

[0187] Fill in missing values ​​in the data set or delete corresponding data rows;

[0188] Eliminate random noise in data according to filtering algorithms;

[0189] Identify outliers in the data and delete or correct them;

[0190] Standardize the data and convert it to a uniform scale.

[0191] In this embodiment, multi-source data fusion is performed on the collected data, specifically for:

[0192] Output the pre-processed data from different devices to a central database and associate them through the key fields of device ID and timestamp;

[0193] Or, data frames from different devices are spliced ​​horizontally or vertically according to timestamps or device IDs;

[0194] Alternatively, additional features may be added to the data set for each device, where the additional features include device type and location information.

[0195] In this embodiment, the identification module 720 is specifically used for:

[0196] Obtain a dataset containing various network status indicators;

[0197] The data set is input into an advanced machine learning model, which outputs the prediction results of the future network status. The advanced machine learning model monitors and predicts the network status in real time based on reinforcement learning and adaptive neural networks.

[0198] In this embodiment, the processing module 730 is specifically used for:

[0199] Obtain traffic data and environmental parameters collected by network devices in real time;

[0200] The data is preprocessed and preliminarily analyzed according to the algorithm built into the computing unit, and the results of the preliminary analysis are output.

[0201] The redundant damage detection device of the redundant network performs data preprocessing and multi-source data fusion on the collected data, facilitates the smoothing of traffic data and integration of data sets of different nodes, dynamically adjusts the detection standard by capturing the dynamic changes of the network status in real time, can adapt to the dynamic network environment, improves the detection results, and effectively improves the stability and reliability of the redundant network.

[0202] Fig.10 An example of a physical structure diagram of an electronic device is shown. The electronic device may be a smart terminal, and its internal structure diagram may be as follows: Fig.10 As shown. The electronic device includes a processor, a memory and a network interface connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a redundancy damage detection method for a redundant network is implemented, and the method includes:

[0203] Acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength and network utilization;

[0204] Identify dynamic changes in network status indicators and output prediction results for future network status;

[0205] Perform lightweight computing on network edge nodes and output the computing results;

[0206] In response to the result of the computing process being outside the normal range, the network routing is instructed to adjust the strategy or trigger an early warning mechanism.

[0207] Those skilled in the art will understand that Fig.10 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present invention, and does not constitute a limitation on the electronic device to which the scheme of the present invention is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0208] On the other hand, the present invention also provides a computer storage medium storing a computer program, which implements a redundancy impairment detection method for a redundant network when the computer program is executed by a processor, the method comprising:

[0209] Acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength and network utilization;

[0210] Identify dynamic changes in network status indicators and output prediction results for future network status;

[0211] Perform lightweight computing on network edge nodes and output the computing results;

[0212] In response to the result of the computing process being outside the normal range, the network routing is instructed to adjust the strategy or trigger an early warning mechanism.

[0213] In another aspect, a computer program product or a computer program is provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and when the processor executes the computer instructions, a redundancy impairment detection method for a redundant network is implemented, the method comprising:

[0214] Acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength and network utilization;

[0215] Identify dynamic changes in network status indicators and output prediction results for future network status;

[0216] Perform lightweight computing on network edge nodes and output the computing results;

[0217] In response to the result of the computing process being outside the normal range, the network routing is instructed to adjust the strategy or trigger an early warning mechanism.

[0218] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0219] By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0220] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0221] The above-described embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

Claims

1. A redundancy damage detection method for a redundant network, characterized in that: The method comprises: Acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of the network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength, and network utilization; Identify dynamic changes in network status indicators and output prediction results for future network status; Perform lightweight computing on network edge nodes and output the computing results; In response to the result of the computing process being outside the normal range, the network routing is instructed to adjust the strategy or trigger an early warning mechanism.

2. The redundancy damage detection method of a redundant network according to claim 1, characterized in that: The data preprocessing of the collected data includes: Clean and standardize the data; And / or, extracting features useful for damage detection from the data, wherein the features include statistical features, frequency domain features, time domain features, waveform features, and machine learning features.

3. The redundancy damage detection method of a redundant network according to claim 2, characterized in that: The data is cleaned and standardized, including: Delete data not relevant to damage detection; Fill in missing values ​​in the data set or delete corresponding data rows; Eliminate random noise in data according to filtering algorithms; Identify outliers in the data and delete or correct them; Standardize the data and convert it to a uniform scale.

4. The redundancy damage detection method of a redundant network according to claim 3, characterized in that: The multi-source data fusion of the collected data includes: Output the pre-processed data from different devices to a central database and associate them through the key fields of device ID and timestamp; Or, data frames from different devices are spliced ​​horizontally or vertically according to timestamps or device IDs; Alternatively, additional features are added to the data set of each device, wherein the additional features include device type and location information.

5. The redundancy damage detection method of a redundant network according to claim 4, characterized in that: The identifying dynamic changes of network status indicators and outputting prediction results of future network status include: Obtain a dataset containing various network status indicators; The data set is input into an advanced machine learning model to output a prediction result of the future network state. The advanced machine learning model is based on reinforcement learning and adaptive neural network to monitor and predict the network state in real time.

6. The redundancy damage detection method of a redundant network according to claim 5, characterized in that: The adaptive neural network adopts a long short-term memory network to capture the long-term dependencies and short-term dynamic changes of network states.

7. The redundancy damage detection method for a redundant network according to any one of claims 1 to 6, characterized in that: The performing lightweight computing processing on the network edge node and outputting the computing processing result includes: Obtain traffic data and environmental parameters collected by network devices in real time; The data is preprocessed and preliminarily analyzed according to the algorithm built into the computing unit, and the results of the preliminary analysis are output.

8. A redundancy damage detection device for a redundant network, characterized in that: include: An acquisition module is used to acquire data collected synchronously by multiple network nodes, perform data preprocessing and multi-source data fusion on the collected data, and create a data set containing multiple network status indicators. The data of the network nodes include traffic statistics, channel quality indicators, network delay, packet loss rate, error rate, signal strength and network utilization; The recognition module is used to identify the dynamic changes of network status indicators and output the prediction results of future network status; A processing module, used to perform lightweight computing processing on network edge nodes and output the computing processing results; The response module is used to instruct the network routing to adjust the strategy or trigger the early warning mechanism in response to the result of the calculation processing exceeding the normal range.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the redundancy impairment detection method for a redundant network according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the redundancy impairment detection method for a redundant network according to any one of claims 1 to 7 are implemented.

Images