Network self-healing device based on node state data

Abstract

The invention relates to the technical field of network communication, provides a network self-healing device based on node state data, and aims to improve the stability and self-healing capability of network communication. Through real-time monitoring of the network state, state estimation and filtering are carried out in combination with an extended Kalman filtering algorithm, and fault prediction, path selection and resource allocation are carried out through an ant colony algorithm and a simulated annealing algorithm, so that prediction and rapid recovery of potential faults are realized. The device comprises a state monitoring module, a state estimation and filtering module, a fault prediction and prevention module, a fault recovery module and a resource redistribution module, forms a closed-loop optimization mechanism, and is particularly suitable for a complex large-scale network environment, such as a data center network.

Description

Technical Field

[0001] The present invention relates to the field of network communication technology, and specifically to a network self-healing device based on node status data, which aims to improve the stability and self-healing ability of network communication by real-time monitoring of network status, prediction and detection of potential faults by a state observer, and taking preventive measures for intelligent decision-making path selection and resource allocation. Background Art

[0002] With the continuous advancement of network technology, the stability and reliability of network communication have become the core requirements of various application scenarios. In actual applications, the network faces various sudden failures, such as equipment failure and link interruption. First, it relies on simple fault detection and recovery mechanisms, such as manual experience to check each link and command operation, which are often slow to respond and difficult to quickly restore network communication, seriously affecting the communication quality and user experience; second, it lacks the ability to observe and intelligently analyze the network status. The network status includes multiple dimensions such as traffic load, link quality, and equipment health, which are crucial for predicting and detecting potential failures. The existing technology does not comprehensively observe and utilize these status information, resulting in the inability to take effective preventive measures before network failures occur; third, path selection and resource allocation are key factors affecting the efficiency and stability of network communication. Traditional path selection algorithms often find it difficult to find the optimal path in the global scope and are prone to fall into local optimal solutions, resulting in uneven allocation of network resources and reduced communication efficiency. Summary of the invention

[0003] This invention aims to solve the key problems existing in the current data center network self-healing technology: high communication delay, low monitoring status data accuracy, path selection limitations, and uneven resource allocation. By designing a network self-healing device, introducing advanced algorithm models and status observation mechanisms, real-time and efficient communication guarantees are achieved, while improving the network's ability to predict and respond to potential failures.

[0004] The technical solution adopted by the present invention is:

[0005] A network self-healing device based on node status data, comprising:

[0006] The status monitoring module is used to capture the status information of the network and each device node in real time, including the link quality, traffic dynamics and device load information in the network, and pass the collected device status data to the state estimation and filtering module, and pass the resource utilization information and network status information transmitted by the resource reallocation module to the fault prediction and prevention module;

[0007] The state estimation and filtering module is used to receive the equipment state data, use the extended Kalman filter algorithm to perform fine processing on the equipment state data, obtain the filtered equipment state data, and pass it to the fault prediction and prevention module and the fault recovery module;

[0008] The fault prediction and prevention module is used to receive the filtered device status data transmitted by the state estimation and filtering module and the network status data transmitted by the state monitoring module; for the device status data, if it is judged that the current device has a fault, the relevant device control strategy is generated; for the network status data, if it is judged to be fault data, the fault data is combined with the resource utilization information transmitted by the state monitoring module and the historical fault data of the network node for analysis and diagnosis, and a network time series data set is constructed, and a control strategy is generated based on the local control strategy library. If it is judged to be normal status data, the existing ARIMA prediction model is used to predict the nodes that may fail in the future set time period, and the corresponding control strategy is generated and passed to the fault recovery module;

[0009] The fault recovery module is used to receive the control strategy generated by the fault prediction and prevention module. For network fluctuation faults, different algorithms are selected for path planning according to the complexity of the current network. For equipment failures, the relevant equipment is temporarily isolated and an equipment alarm signal is generated. The isolation state is removed after the relevant equipment is normal. After taking corresponding measures, it is determined whether the measures are effective. If effective, the fault recovery result is passed to the resource reallocation module. If it fails, the relevant alarm information is generated.

[0010] The resource reallocation module is used to reallocate network resources according to the fault recovery status and the current network node resource usage, and to feed back the resource utilization information to the status monitoring module in real time to form a closed-loop optimization mechanism.

[0011] Furthermore, the state estimation and filtering module uses a two-stage Kalman filter model to perform fine processing on the received device state data, including:

[0012] In the first stage, the extended Kalman filter (EKF) is used to perform a preliminary assessment of the equipment status and obtain the original equipment status data;

[0013] In the second stage, based on the original equipment status data in the first stage, the standard Kalman filter is used for further fine-tuning to achieve data preprocessing, wave control denoising and normalization to obtain effective features.

[0014] Furthermore, the specific process of the fault prediction and prevention module is as follows:

[0015] The fault prediction and prevention module first determines whether the data is device status data or network status data; if it is determined to be device status data and fault data, then the relevant device control strategy is generated for device isolation and recovery; if it is determined to be network status data, then it is processed according to whether a fault has occurred. If it is fault data, the key fault information is extracted and combined with the resource utilization information transmitted by the status monitoring module, it is passed into the fault judgment vector machine to diagnose the cause of the fault and query the local preset control strategy library to generate a control strategy; the resource utilization information is used to help diagnose environmental abnormalities; if it is determined to be normal status data, a time series data set is constructed based on the data packet transmission status and network delay in the status data, and the model is predicted as the input of the ARIMA prediction model. According to the prediction results, the trend of network fluctuations and potential fault points in the future set time period are judged, and the corresponding control strategy is generated.

[0016] Furthermore, when performing path planning, the fault recovery module adopts different path planning strategies according to the complexity of different networks: for networks with less than a set number of nodes, a fixed network topology, and a simple and predictable traffic pattern, the ant colony algorithm is used to dynamically plan the optimal or suboptimal communication path; for networks with more than a set number of nodes, a dynamically changing network topology, and a complex and unpredictable traffic pattern, the simulated annealing algorithm is enabled to accept non-optimal paths by probability.

[0017] Compared with the prior art, the present invention achieves the following effective gains:

[0018] (1) The original data is denoised through the two-stage EKF algorithm, which greatly improves the accuracy and reliability of the state data.

[0019] (2) Based on the network status data continuously collected by the monitoring module, a relevant time series dataset is constructed as the input of the ARIMA model for modeling and predicting the network status, thereby realizing the prediction of potential faults.

[0020] (3) Combining the ant colony algorithm with the simulated annealing algorithm, global and local optimal path selection and resource allocation are achieved, which enables rapid recovery from network fluctuation failures and improves network communication efficiency.

[0021] (4) By combining the state observation mechanism with the intelligent algorithm model, it is possible to predict network state changes in real time, detect and prevent potential failures in a timely manner, and significantly improve the network's self-healing capabilities and service continuity. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a diagram of a network self-healing system device according to an embodiment of the present invention.

[0023] Figure 2This is a structural diagram of the status monitoring module according to an embodiment of the present invention.

[0024] Figure 3 This is a processing diagram of the fault prediction and prevention module according to an embodiment of the present invention.

[0025] Figure 4 This is a diagram of the fault recovery module processing process of an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following is combined with Figure 1-4 The present invention is further explained.

[0027] The present invention adopts a network self-healing mechanism based on state detection. The main purpose is to provide a new network control system automation solution, which integrates the extended Kalman filter (EKF) algorithm, the ant colony algorithm and the simulated annealing algorithm, aiming to achieve efficient denoising of the original data of the network state, potential fault prediction, intelligent path selection and global optimization resource allocation. The solution can monitor the network state in real time, design a network self-healing device with the ability to predict and detect potential faults, perform intelligent decision-making path selection and resource allocation, and achieve the goal of improving the stability and self-healing ability of network communication. It not only significantly improves the stability and self-healing ability of network communication, but is also particularly suitable for complex large-scale network environments, such as data center networks.

[0028] like Figure 1 As shown, a network self-healing device based on node status data includes:

[0029] In the network control system described in the present invention, multiple key components are deployed for status monitoring, such as Figure 2 As shown, the condition monitoring module includes sensors (or data collectors), controllers, actuators, and the communication network between them. The specific process is:

[0030] Sensors capture the status information of the network and each device node (controlled device) in real time, such as link quality, traffic dynamics, device load, and physical parameters such as temperature and pressure.

[0031] The state monitoring module deployed in the sensor network collects data from sensors regularly by using controllers to construct state vector data. The collected original equipment status data D0 is passed to the state estimation and filtering module through the communication network, and the resource utilization information and network status data D1 transmitted by the resource reallocation module are passed to the fault prediction and prevention module.

[0032] The state estimation and filtering module is the core processing unit in the device. It uses a two-stage Kalman filter model to finely process the received device state data. It combines the state transfer function and observation function defined by the system dynamic model, and the Kalman gain calculated based on the observation noise and process noise to make a preliminary estimate of the current state and obtain the filtered state information. Passed to the fault prediction and prevention module and the fault recovery module.

[0033] The embodiment of the present invention adopts a two-stage Kalman filter model to perform fine processing on the received data, specifically:

[0034] Construct a two-stage model: In the first stage, the EKF is used to make a preliminary assessment of the system state, and the formula is expressed as:

[0035]

[0036] In the formula, is the current state estimate, f(x) is the state transfer function, K k is the Kalman gain, z k is the observed value (from the sensor), h(x) is the observation function (mapping the system state to the observation space);

[0037] In the second stage, based on the estimation in the first stage, the standard Kalman filter is used for further fine-tuning to implement preprocessing steps such as denoising and normalization of the data to obtain effective features.

[0038] like Figure 3 As shown in the figure, the fault prediction and prevention module receives the fault information from the state estimation and filtering module. After receiving the network status data D1 from the status monitoring module, it will determine whether it is fault data. If it is fault data, it will analyze the status data, extract the key fault information and pass it into the fault judgment vector machine to diagnose the cause of the fault, and query the local preset control strategy library (dictionary) StD to generate the control strategy S; if it is normal status data, it will build a time series data set TSDS according to the data packet transmission status and network delay in the status data. This data set will be used as the input of the ARIMA model for model prediction. According to the prediction results, the trend of network fluctuations and potential fault points in the future period T will be judged, and the control strategy S will be generated, such as dynamically adjusting the data packet sending rate, optimizing routing selection, caching key data, backing up communication paths, etc. The generated control strategy S is sent to the fault recovery module.

[0039] Among them, the ARIMA model is used for time series forecasting. The model mainly consists of three parts: autoregression (AR), difference (I) and moving average (MA). By integrating different components in the time series set TSDS (such as trend, residual, etc.), the obtained non-stationary time series is converted into a stationary time series, and then the autoregression and moving average models are used for forecasting. The autoregression and moving average models are used to obtain the linear relationship between the current value and the historical record value and the historical error term.

[0040] like Figure 4 As shown in the figure, after the fault recovery module obtains the control strategy, it will first determine the type of control strategy S. If it is a device failure, the device will be temporarily isolated and an equipment alarm signal will be generated. After the device is normal, the isolation state will be removed. After taking corresponding solutions, the fault recovery module will determine whether the measures are effective. If effective, the fault recovery result will be sent to the resource reallocation module. If it fails, the relevant alarm information W will be generated. If it is a network fluctuation failure, different algorithms are flexibly selected for path planning according to the complexity of the current network. In simple networks, the ant colony algorithm is used to dynamically plan the optimal or suboptimal communication path; in complex large networks, the simulated annealing algorithm is enabled to enhance the global search capability by probabilistically accepting non-optimal paths.

[0041] When performing path planning, the fault recovery module can adopt different path planning strategies according to the complexity of different networks. For simple networks, that is, networks with a small number of nodes (such as dozens to hundreds of nodes), relatively fixed network topology, and relatively simple and predictable traffic patterns, the ant colony algorithm is used to dynamically plan the optimal or suboptimal communication path; for complex and large networks, that is, networks with a large number of nodes (such as tens of thousands of nodes), dynamically changing network topology, and complex and unpredictable traffic patterns, the simulated annealing algorithm is enabled to enhance the global search capability by probabilistically accepting non-optimal paths to avoid falling into local optimal solutions. In addition, in order to improve the response speed, a parallel construction method is adopted in the network path query process.

[0042] Among them, the ant colony algorithm is used to dynamically plan the optimal or suboptimal communication path. The specific process is as follows:

[0043] First, initialize the pheromone τ on each edge of the network ij (0) = C, where C is a constant, τ ij (0) represents the initial amount of pheromone; assuming that ant k is on edge e ij After moving, the pheromone is updated to Where ρ is the volatility coefficient of pheromone, which indicates the volatility ratio of pheromone over time, m is the number of ants, Δτ ij is the number of ants on edge e during the time period t~t+1.ij the amount of pheromone left; and the probability that the ant chooses to move on edge e ij is where is the α-th power of the pheromone concentration on edge e ij and is the β-th power of the heuristic information on edge e ij (the reciprocal of the edge length), ij and is the sum over all edges allowed for the k-th ant to choose, where allowed k represents the set of nodes that the ant can move to in the current state, is the α-th power of the pheromone concentration on edge e is and is the β-th power of the heuristic information on edge e is where α and β are algorithm parameters used to control the relative importance of pheromone and heuristic information; then the path record table Ta is used to record the nodes passed by ant k currently. After t moments, all ants complete one cycle; finally, the path length traveled by each ant is calculated, the shortest path length is saved, and it jumps to the pheromone initialization stage at the beginning of the model to prepare for the next cycle;

[0044] However, the ant colony algorithm will fall into the situation of local optimal solution when dealing with complex large-scale networks. Therefore, in order to improve the performance under complex networks, a scheme of ant colony algorithm + simulated annealing algorithm is designed. For the current solution S and the new solution S′, if w(S') < w(S), then S' is accepted as the current solution. Otherwise, with probability S' is accepted. Where w(S') - w(S) is the weight difference between the new solution and the current solution. That is, the way of probabilistically accepting non-optimal paths is used to enhance the global search ability, and a parallel construction method is adopted to reduce the time consumption in the path query process.

[0045] When performing path planning, the simulated annealing algorithm (SA) is enabled, and the global search ability is enhanced by probabilistically accepting non-optimal paths to solve the local overfitting problem in the dynamic programming of the ant colony algorithm. Its formula is expressed as:

[0046]

[0047] where ΔE is the path quality difference and T is the current temperature. In addition, in order to reduce the time consumption of path construction, this device adopts a parallel construction method to find paths.

[0048] The resource reallocation module reallocates network resources (such as bandwidth, processor time, etc.) according to the fault recovery status and the current network node resource usage, and feeds back the resource utilization information to the status monitoring module in real time to form a closed-loop optimization mechanism.

Claims

1. A network self-healing device based on node status data, characterized in that: include: The status monitoring module is used to capture the status information of the network and each device node in real time, including the link quality, traffic dynamics and device load information in the network, and pass the collected device status data to the state estimation and filtering module, and pass the resource utilization information and network status information transmitted by the resource reallocation module to the fault prediction and prevention module; The state estimation and filtering module is used to receive the equipment state data, use the extended Kalman filter algorithm to perform fine processing on the equipment state data, obtain the filtered equipment state data, and pass it to the fault prediction and prevention module and the fault recovery module; The fault prediction and prevention module is used to receive the filtered device status data transmitted by the state estimation and filtering module and the network status data transmitted by the state monitoring module; for the device status data, if it is determined that the current device has a fault, a relevant device control strategy is generated; For network status data, if it is judged to be fault data, the fault data is analyzed and diagnosed in combination with the resource utilization information transmitted by the status monitoring module and the historical fault data of the network nodes, a network time series data set is constructed, and a control strategy is generated based on the local control strategy library. If it is judged to be normal status data, the existing ARIMA prediction model is used to predict the nodes that may fail within the set time period in the future, and the corresponding control strategy is generated and passed to the fault recovery module; The fault recovery module is used to receive the control strategy generated by the fault prediction and prevention module, and select different algorithms for path planning according to the complexity of the current network in response to network fluctuation failures; In case of equipment failure, the relevant equipment will be temporarily isolated and an equipment alarm signal will be generated. The isolation status will be removed after the relevant equipment is back to normal. After taking corresponding measures, it is determined whether the measures are effective. If effective, the fault recovery result is passed to the resource reallocation module. If it fails, relevant alarm information is generated; The resource reallocation module is used to reallocate network resources according to the fault recovery status and the current network node resource usage, and to feed back the resource utilization information to the status monitoring module in real time to form a closed-loop optimization mechanism.

2. A network self-healing device based on node status data according to claim 1, characterized in that: The state estimation and filtering module uses a two-stage Kalman filter model to perform fine processing on the received device state data, including: In the first stage, the extended Kalman filter (EKF) is used to perform a preliminary assessment of the equipment status and obtain the original equipment status data; In the second stage, based on the original equipment status data in the first stage, the standard Kalman filter is used for further fine-tuning to achieve data preprocessing, wave control denoising and normalization to obtain effective features.

3. A network self-healing device based on node status data according to claim 1, characterized in that: The specific process of the fault prediction and prevention module is as follows: The fault prediction and prevention module first determines whether the data is device status data or network status data; if it is determined to be device status data and fault data, then the relevant device control strategy is generated for device isolation and recovery; if it is determined to be network status data, then it is processed according to whether a fault has occurred. If it is fault data, the key fault information is extracted and combined with the resource utilization information transmitted by the status monitoring module, it is passed into the fault judgment vector machine to diagnose the cause of the fault and query the local preset control strategy library to generate a control strategy; the resource utilization information is used to help diagnose environmental abnormalities; if it is determined to be normal status data, a time series data set is constructed based on the data packet transmission status and network delay in the status data, and the model is predicted as the input of the ARIMA prediction model. According to the prediction results, the trend of network fluctuations and potential fault points in the future set time period are judged, and the corresponding control strategy is generated.

4. A network self-healing device based on node status data according to claim 1, characterized in that: When performing path planning, the fault recovery module adopts different path planning strategies according to the complexity of different networks: for networks with less than the set number of nodes, fixed network topology, and simple and predictable traffic patterns, the ant colony algorithm is used to dynamically plan the optimal or suboptimal communication path; for networks with more than the set number of nodes, dynamically changing network topology, and complex and unpredictable traffic patterns, the simulated annealing algorithm is enabled to accept non-optimal paths by probability.

Images