Fault detection method, network device, first platform, storage medium and product

Abstract

The application discloses a fault detection method, which comprises the following steps: a network device acquires first information; wherein the first information is used for indicating the network state of a transmission link between the network device and a first terminal; based on the first information, a network fault performance evaluation index parameter is determined; the network fault performance evaluation index parameter is input into a fault detection model, and the network state of the transmission link is detected through the fault detection model to obtain fault information of a fault node; wherein the fault node comprises the network device and the first terminal; and the fault detection model is a model based on a two-layer graph convolution network. The application also discloses a network device, a first platform, a communication device, a computer readable storage medium and a computer program product.

Description

Technical Field

[0001] This application relates to, but is not limited to, the field of communications, and particularly to a fault detection method, network equipment, a first platform, a computer-readable storage medium, and a computer program product. Background Technology

[0002] Currently, common types of Large-scale Deep Coverage Field Area Networks (LDC FANs) include mobile wireless sensor networks, wireless mesh networks, and wireless ad hoc networks. Wireless sensor networks are a type of network formed by organizing and combining various heterogeneous sensor nodes through wireless communication technology; wireless mesh networks consist of gateways, mesh routers, and mesh clients; and wireless ad hoc networks are wireless multi-hop networks composed of mobile terminals. All of these networks can communicate with each other without base station support.

[0003] It should be noted that network fault detection is a key link in ensuring network service quality. Because LDC FAN includes a large number of terminals, routing nodes and other devices, and is in a complex communication environment, its failure probability is much higher than that of traditional networks.

[0004] Currently, network fault detection methods in related technologies include distributed wireless mesh network self-diagnosis schemes, decision tree-based mesh network fault diagnosis algorithms, and statistical fault detection methods. However, most of these methods are used in cellular networks, which have relatively fixed base station nodes, coverage areas, and topologies, making them unsuitable for fault management in large-scale, highly dynamic wireless networks. Furthermore, in distributed wireless mesh networks, different nodes may collect different types of data, and network latency or node failures can lead to data asynchrony. Each distributed node has different computing resources, and running complex fault detection algorithms cannot guarantee data synchronization of fault results. Therefore, the aforementioned network fault detection methods at least suffer from the problem of being unsuitable for dynamic networks or experiencing data asynchrony. Summary of the Invention

[0005] This application provides a fault detection method, a network device, a first platform, a computer-readable storage medium, and a computer program product, and provides a network fault detection method based on graph convolutional networks.

[0006] The technical solution of this application embodiment is implemented as follows:

[0007] Firstly, embodiments of this application provide a fault detection method applied to a network device, the method comprising:

[0008] Obtain first information; wherein, the first information is used to indicate the network status of the transmission link between the network device and the first terminal;

[0009] Based on the first information, determine the network fault performance evaluation index parameters;

[0010] The network fault performance evaluation index parameters are input into the fault detection model, and the network status of the transmission link is detected by the fault detection model to obtain the fault information of the fault node; wherein, the fault node includes the network device and the first terminal; the fault detection model is a model based on a two-layer graph convolutional network.

[0011] Secondly, embodiments of this application provide a fault detection method applied to a first platform, the method comprising:

[0012] Receive fault information of a faulty node sent by a network device; wherein, the fault information is obtained by the network device by inputting network fault performance evaluation index parameters determined based on the first information obtained into a fault detection model, and by detecting the network status of the transmission link between the network device and the first terminal through the fault detection model;

[0013] Based on the fault information, adjust the network resource configuration of the faulty node;

[0014] Send a resource adjustment message to the network device; wherein the resource adjustment message includes a resource adjustment strategy.

[0015] Thirdly, embodiments of this application provide a network device, which includes:

[0016] A first acquisition module is configured to acquire first information; wherein the first information is used to indicate the network status of the transmission link between the network device and the first terminal;

[0017] The first processing module is used to determine network fault performance evaluation index parameters based on the first information.

[0018] The first processing module is further configured to input the network fault performance evaluation index parameters into the fault detection model, and detect the network status of the transmission link through the fault detection model to obtain fault information of the fault node; wherein, the fault node includes the network device and the first terminal; the fault detection model is a model based on a two-layer graph convolutional network.

[0019] Fourthly, embodiments of this application provide a first platform, the first platform comprising:

[0020] The second receiving module is used to receive fault information of a faulty node sent by the network device; wherein, the fault information is obtained by the network device by inputting network fault performance evaluation index parameters determined based on the first information into a fault detection model, and by detecting the network status of the transmission link between the network device and the first terminal through the fault detection model;

[0021] The second processing module is used to adjust the network resource configuration of the faulty node based on the fault information.

[0022] The second sending module is used to send a resource adjustment message to the network device; wherein the resource adjustment message includes a resource adjustment strategy.

[0023] Fifthly, embodiments of this application provide a network device, the network device comprising:

[0024] The first memory is used to store executable instructions;

[0025] The first processor, when executing executable instructions stored in the first memory, performs the steps of the above-described fault detection method.

[0026] Sixthly, embodiments of this application provide a first platform, the first platform comprising:

[0027] The second memory is used to store executable instructions;

[0028] When the second processor executes executable instructions stored in the second memory, it performs the steps of the above-described fault detection method.

[0029] Seventhly, embodiments of this application provide a computer-readable storage medium having stored thereon one or more computer programs, which can be executed by one or more processors to implement the steps of the above-described fault detection method.

[0030] Eighthly, this application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-described fault detection method.

[0031] This application proposes a fault detection method based on a two-layer graph convolutional network (GCR) model. This GCR model improves the accuracy of fault detection model analysis, enabling it to adapt well to dynamic network environments and changes in node behavior. This allows for real-time monitoring of network faults in large-scale dynamic network environments, ensuring service quality. Furthermore, this application addresses the issue of data synchronization caused by distributed nodes running their own fault detection algorithms, where network devices process network status information into a unified data format table, which is then input into the fault detection model. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of an LDC FAN according to an embodiment of this application;

[0033] Figure 2 Flowchart of the fault detection method provided in the embodiments of this application Figure 1 ;

[0034] Figure 3 A schematic diagram of a network fault detection process provided in an embodiment of this application;

[0035] Figure 4 A flowchart of a two-layer training model of a graph convolutional network provided in this application embodiment;

[0036] Figure 5 A flowchart illustrating a fault detection method provided in this application embodiment. Figure 2 ;

[0037] Figure 6 A schematic block diagram of a network device provided in an embodiment of this application;

[0038] Figure 7 A schematic block diagram of a first platform provided for embodiments of this application;

[0039] Figure 8 This is a schematic structural diagram of a communication device provided in an embodiment of this application. Detailed Implementation

[0040] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0041] Figure 1 This is a schematic diagram of an LDC FAN according to an embodiment of this application.

[0042] like Figure 1 As shown, LDC FAN includes a cloud management platform 110, an edge platform 120, and an end platform 130.

[0043] It should be noted that large-scale deep coverage area networks (MDNs) are local heterogeneous communication networks that integrate high-speed power line carrier, low-power wireless, and low-power wireless communication methods. MDNs use wireless and power line carrier as transmission media, enabling the connection of multiple protocol devices and providing deep coverage communication capabilities.

[0044] The cloud management platform 110 includes a network management system that provides real-time monitoring of node status, network analysis, network assessment, and network management functions, and can function as a remote cloud center within the network. Simultaneously, it features a large-scale front-end data visualization interface, allowing users to view the hardware information and network topology of each node in real time, thus assisting operations and maintenance personnel in better diagnosing faults.

[0045] The edge platform 120 includes a head node, also known as a central master node, which is responsible for network construction, maintenance, and management, as well as data transmission between the head node and tail nodes and terminal nodes. It can be an edge base station. The head node can sense the edge network environment, collect and process terminal data, and perform edge network fault detection and feedback.

[0046] The terminal platform 130 consists of terminal nodes and tail nodes. Terminal nodes are terminal devices that generate source data, such as user mobile smartphones. Tail nodes, also known as routing aggregation nodes, are responsible for data collection, transmission, and forwarding from the terminal nodes; they can be access points (APs) or micro base stations. Based on the transmission medium used, tail nodes are classified as carrier nodes, wireless nodes, and dual-mode nodes. Carrier nodes use carrier transmission media, wireless nodes use wireless transmission media, and dual-mode nodes use both wired and wireless transmission channels for dual-mode data transmission. These nodes automatically select the most suitable transmission mode based on actual conditions to ensure reliable data transmission and efficient network performance.

[0047] It should be noted that, Figure 1This application merely illustrates the network to which this application applies; of course, the methods shown in the embodiments of this application can also be applied to other networks. Furthermore, the terms "system" and "network" are often used interchangeably herein. The term "and / or" in this application merely describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " in this application generally indicates that the preceding and following related objects are in an "or" relationship. It should also be understood that "instruction" mentioned in the embodiments of this application can be a direct instruction, an indirect instruction, or an indication of a relationship. For example, A instructing B can mean that A directly instructs B, for example, B can be obtained through A; it can also mean that A indirectly instructs B, for example, A instructs C, B can be obtained through C; or it can mean that there is a relationship between A and B. It should also be understood that "correspondence" mentioned in the embodiments of this application can mean a direct or indirect correspondence between two things, an association between two things, or a relationship of instruction and being instructed, configuration and being configured, etc. It should also be understood that the "predefined" or "predefined rules" mentioned in the embodiments of this application can be implemented by pre-storing corresponding codes, tables, or other means that can be used to indicate relevant information in the device (e.g., including terminal devices and network devices), and this application does not limit the specific implementation method. For example, predefined can refer to those defined in a protocol. It should also be understood that in the embodiments of this application, the "protocol" can refer to standard protocols in the field of communication, such as the Long Term Evolution (LTE) protocol, the New Radio (NR) protocol, and related protocols applied to future communication systems, and this application does not limit this.

[0048] To facilitate understanding of the technical solutions of the embodiments of this application, the relevant technologies of the embodiments of this application are described below. The following relevant technologies are optional solutions and can be combined with the technical solutions of the embodiments of this application in any way, and they all fall within the protection scope of the embodiments of this application.

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0050] Before explaining this application, the following describes network fault detection methods in related technologies:

[0051] Distributed wireless mesh network self-diagnostic schemes achieve this by running fault detection algorithms on multiple nodes in the network. These algorithms can monitor the status of nodes, the quality of communication links, and the overall performance of the network, thereby promptly detecting and reporting faults.

[0052] The decision tree-based Mesh network fault diagnosis algorithm identifies normal and abnormal behaviors in the network by training a machine learning model, thereby achieving fault detection.

[0053] Statistical fault detection methods detect faults by analyzing network statistics such as latency, packet loss rate, and throughput. When these statistics deviate from normal ranges, the system considers a fault to have occurred. Statistical methods require setting reasonable thresholds to distinguish between normal and abnormal behavior.

[0054] Figure 2 This is a flowchart illustrating a fault detection method provided in an embodiment of this application, as shown below. Figure 2 As shown, this method is applied to Figure 1 The method includes: (Image of edge platform 120 in the network shown)

[0055] Step 201: The network device obtains the first information.

[0056] The first information is used to indicate the network status of the transmission link between the network device and the first terminal.

[0057] In this embodiment of the application, the network device, i.e. Figure 1 The edge platform 120, the monitoring terminal equipment is... Figure 1 The process of the tail node sending data to the network device to obtain the first information.

[0058] In this embodiment of the application, the parameters corresponding to the network status include, but are not limited to, bandwidth, latency, packet loss rate, and throughput.

[0059] It should be noted that the first terminal receives data from at least one second terminal, i.e. Figure 1 The terminal node in the network sends at least one source data generated by at least one second terminal; the first terminal aggregates at least one source data to obtain aggregated data; the first terminal sends the aggregated data to the network device based on a transmission mode that matches its own network status; wherein the transmission mode includes one of the following: carrier transmission, wireless transmission and dual-mode transmission, and during the transmission of aggregated data, the network device obtains first information.

[0060] It should be noted that source data refers to unprocessed data, that is, the actual collected data. Aggregation includes data normalization, preprocessing and other processes to process at least one source data.

[0061] In this embodiment of the application, the number of first terminals can be one or more.

[0062] In some embodiments, the first terminal determines the transmission mode based on the transmission medium used between the network device and the first terminal, and then transmits aggregated data according to the transmission mode.

[0063] For example, if a carrier transmission medium is used, the transmission mode is carrier transmission; if a wireless transmission medium is used, the transmission mode is wireless transmission; if a wired and wireless transmission medium is used, the transmission mode is dual-mode data transmission.

[0064] Step 202: Based on the first information, the network device determines the network fault performance evaluation index parameters.

[0065] In some embodiments, the network device also needs to determine node demand data, wireless / wired channel status, and network resource status from the end platform based on the first information.

[0066] In some embodiments, during network fault detection, it is necessary to unify the input data format of the neural network. Therefore, network fault performance evaluation index parameters are designed. The network fault performance evaluation index parameters can be a network fault performance evaluation index table, which is a unified data format (table) that processes the first information into a format that can be input into the neural network.

[0067] Step 203: The network device inputs the network fault performance evaluation index parameters into the fault detection model, and uses the fault detection model to detect the network status of the transmission link, thereby obtaining the fault information of the fault node output by the fault detection model.

[0068] The faulty nodes include network devices and the first terminal; the fault detection model is based on a two-layer graph convolutional network.

[0069] The fault information includes one or more of the following: the type of fault; the location of the fault node.

[0070] In this embodiment, after network fault detection is completed and the predicted network fault is obtained, it is necessary to unify the output data format of the neural network and then forward it to the corresponding faulty node or the first platform. Figure 1 Therefore, this application designs a network fault prediction label, i.e., fault information, based on the cloud management platform 110 in the cloud.

[0071] In some embodiments, node demand data, wireless / wired channel status, network resource status, and network fault performance evaluation parameters from the end platform are input into a fault detection model based on a two-layer graph convolutional network to obtain a predicted label of the network fault output by the fault detection model; the predicted label includes the type of network fault and the location of the fault node.

[0072] This application discloses a fault detection method, which includes: a network device acquiring first information; wherein the first information indicates the network status of the transmission link between the network device and a first terminal; determining network fault performance evaluation index parameters based on the first information; inputting the network fault performance evaluation index parameters into a fault detection model, and detecting the network status of the transmission link through the fault detection model to obtain fault information of a faulty node; wherein the faulty node includes the network device and the first terminal; and the fault detection model is based on a two-layer graph convolutional network. In other words, this application proposes a fault detection method based on a two-layer graph convolutional network fault detection model. This two-layer graph convolutional network can improve the accuracy of the fault detection model in analyzing fault information, enabling the fault detection model to adapt well to dynamic network environments and changes in node behavior, achieving real-time monitoring of network faults in large-scale dynamic network environments and ensuring service quality. Furthermore, this application solves the problem of data asynchrony caused by distributed nodes running their own fault detection algorithms by having the network device process the network status information into a unified data format table and then input it into the fault detection model to obtain fault information.

[0073] In some embodiments, considering the application scenarios of LDC FAN networks, network faults include one or more of the following: node performance faults; device faults; application software faults; node hardware faults; network-level faults; network fault performance evaluation parameters include one or more of the following: inclusiveness parameters; sustainability parameters; credibility parameters; wherein, inclusiveness parameters include one or more of the following: network throughput; reference signal received strength; average signal-to-noise ratio; packet loss rate; sustainability parameters include one or more of the following: transmission delay; transmission redundancy; credibility parameters include one or more of the following: node spacing variance.

[0074] Node performance failures, device failures, and application software failures are soft failures. Although the node is not in a normal operating state related to the upper-layer application, it can still continue to operate and communicate with other nodes. These failures can be detected by software. Node hardware failures and network-level failures are hard failures. The node can only send and receive control signals but cannot communicate with other nodes.

[0075] Here, node hardware failures can be caused by factors such as wireless network card malfunction, node crash, or node damage. When a terminal node experiences a hardware failure, the edge base stations in the cluster will not periodically respond to management requests. After a specified period, the edge base stations send a node loss report to the cloud management platform. The cloud management platform queries whether the node has moved to another cluster based on the latest topology information. Then, it returns a hardware failure message to the edge base stations and marks and records the node's hardware failure in the logs.

[0076] Here, network-level failures are considered the highest level of failure and require detection using specially trained fault diagnosis models. Other types of failures can be identified through cloud management platforms or monitoring software to determine if a failure has occurred.

[0077] Network-level failures are defined in two categories: abnormal node behavior and abnormal network state.

[0078] Abnormal node behavior includes the following situations:

[0079] Node offline: This occurs when a large portion of the nodes in the network are unable to connect or are disconnected.

[0080] Node signaling issues: This problem occurs when more than a certain percentage of nodes in the network receive weak signals, which can lead to degraded communication quality or unstable connections.

[0081] System failures include incorrect configuration parameters and resource issues such as insufficient memory or too many processes.

[0082] Network status anomalies include the following situations:

[0083] Link failure: A failure occurs when more than a certain percentage of network links fail, which may interrupt communication or cause data transmission failure.

[0084] Network congestion occurs when more than a certain proportion of nodes continue to send large amounts of data, leading to a decline in network performance or an increase in data transmission delay.

[0085] In some embodiments, faults can be reflected by corresponding Key Performance Indicators (KPIs) parameters, considering three dimensions: inclusivity, sustainability, and trustworthiness. These three types of network performance can cover the main performance of the LDC FAN network; that is, by measuring the relevant indicators of these three dimensions—inclusivity, sustainability, and trustworthiness—it can be reflected whether the network has experienced faults.

[0086] Inclusivity parameters: network throughput, reference signal received strength, average signal-to-noise ratio, and packet loss rate. Sustainability parameters: transmission delay and redundancy. Reliability parameter: node spacing variance.

[0087] Therefore, to facilitate network fault diagnosis, Table 1 presents the network fault performance evaluation index table designed in this application. The reference values ​​in the table are the ranges set by maintenance personnel based on experience and network capabilities, while the actual values ​​are the actual measured network performance. By comparing the reference values ​​and actual values ​​of the evaluation indexes, it can be determined whether a network fault has occurred.

[0088] Evaluation indicators symbol Reference value actual value Reference signal received strength RSSI Average network packet loss rate PD Average number of jumps AH Average signal-to-noise ratio SINR Network throughput NT Number of network nodes Num_node Number of network links Num_link Network redundancy NR Node Spacing Variance Node_m_var Average network latency ND

[0089] Table 1 Network Fault Performance Evaluation Indicators

[0090] Received Signal Strength Indication (RSSI): This refers to the strength of the signal received in the downlink. In large-scale deep coverage networks, the RSSI range for wireless nodes is between -120dBm and -10dBm, while the RSSI range for carrier nodes is between 20dBm and 120dBm. RSSI-1 represents the received signal strength of the wireless node, and RSSI-2 represents the received signal strength of the carrier node.

[0091] Average packet loss rate (PD): refers to the ratio of the number of lost data packets to the total number of data packets sent.

[0092] Average hop count (AH): Defined as the ratio of the total number of hops to the total number of paths from all tail nodes to the end of the node.

[0093] Average signal-to-noise ratio (SINR): This is an important metric for measuring network performance. A higher SINR indicates better network performance, meaning less noise in the signal. Conversely, a lower SINR indicates worse network performance.

[0094] Network throughput (NT): It is defined as the amount of data downloaded or uploaded per unit of time. Average throughput can be calculated by dividing the total throughput by the duration of the statistical period.

[0095] Network node count (Num_node): The sum of the number of tail nodes and head nodes in the network.

[0096] Number of network links (Num_link): The sum of the number of power line carrier links and wireless links in the network.

[0097] Network redundancy (NR): It is defined as the average number of independent paths at each tail end.

[0098] Node margin variance (Node_m_var): It is defined as the ratio of the number of paths through a node to the total number of paths to all tails.

[0099] Network average latency (ND): The time interval between sending and receiving.

[0100] After determining whether a network fault has occurred, the head node identifies the type of network fault and the location of the faulty node. Based on the reference and actual values ​​in the indicator table, it formulates the node's demand for the four key resources, further generating a network fault label, which is then sent to the cloud management platform to assist in network management and resource adjustment. The network fault labels are shown in Table 2 below.

[0101] Network Fault Types Fault node location Demand for the four elements of resources Node offline (Data, connectivity, computing power, algorithms) Node signal problem (Data, connectivity, computing power, algorithms) System failure (Data, connectivity, computing power, algorithms) Link failure (Data, connectivity, computing power, algorithms) Network congestion (Data, connectivity, computing power, algorithms)

[0102] Table 2 Network Fault Labels

[0103] In some embodiments, the method provided in this application includes the following:

[0104] Step A1: The network device sends fault information to the first platform.

[0105] Step A2: The first platform receives the fault information and adjusts the network resource configuration of the faulty node based on the fault information.

[0106] In this embodiment of the application, adjusting network resource configuration includes adjusting the parameter configuration of network devices and adjusting network planning configuration.

[0107] Here, the parameter configuration of network devices mainly refers to the queue configuration of the network device's output ports. This configuration depends on parameters such as the service load carried by the network device and the network device's own Service Level Agreement (SLA) estimate.

[0108] Here, network planning and configuration mainly refers to the traffic configuration of each network device in the entire network. This configuration mainly depends on the topology of the entire network, the SLA estimation of each network device, and other parameters.

[0109] Step A3: The first platform sends a resource adjustment message to the network device.

[0110] The resource adjustment message includes the resource adjustment strategy.

[0111] In this embodiment, the resource adjustment strategy is transmitted via various methods, including in-band, out-of-band, media, signaling, data, message, control plane, and user plane. Among these, a relevant media channel is used to transmit the resource adjustment strategy, improving compatibility with existing systems and reducing system upgrade costs. Furthermore, when there are multiple network devices, the established media plane communication channel is a one-to-many multicast / broadcast communication channel. This ensures that the resource adjustment strategy is transmitted only once through the established multicast / broadcast communication channel, and all other network devices can receive it, effectively reducing the number of messages sent.

[0112] Step A4: The network device receives the resource adjustment message.

[0113] It should be noted that network devices perform network fault detection and, after the fault detection is completed, feed the results back to the corresponding faulty node so that the faulty node can first perform self-feedback adjustment; the results should also be fed back to the first platform so that the first platform can complete network analysis, evaluation and management on a larger time scale.

[0114] In some embodiments, the method provided in this application includes the following: obtaining a fault detection model based on a two-layer graph convolutional network; further, obtaining the fault detection model based on a two-layer graph convolutional network includes the following:

[0115] Step B1: The network device preprocesses the data collected under different network conditions into a small sample dataset.

[0116] Step B2: The network device divides the small sample dataset into a training set and a validation set.

[0117] Step B3: The network device sequentially inputs the selected training set into the first-layer graph convolutional network and the second-layer graph convolutional network to obtain the parameters of the fault detection model output by the convolutional network; wherein, each network node in each layer of the graph convolutional network sends its own transformed feature information to its neighboring nodes, and each network node aggregates the transformed feature information sent by all neighboring nodes and transforms the aggregated feature information.

[0118] Step B4: The network device trains the parameters of the fault detection model based on the selected training set.

[0119] In some embodiments, the selected training set is sequentially input into the first layer graph convolutional network and the second layer graph convolutional network to output the relevant parameters of the fault detection model; wherein, each network node in each layer graph convolutional network sends its own transformed feature information to its neighboring nodes, and each network node aggregates the transformed feature information sent by all neighboring nodes and transforms the aggregated feature information.

[0120] Here, in the first-layer graph convolutional network, all network nodes collect feature information from their neighboring nodes; in the second-layer graph convolutional network, each network node performs a feature concatenation operation on its own feature information and the feature information of its neighboring nodes collected from the first-layer graph convolutional network, and then inputs the concatenated feature information into the first-layer graph convolutional network and the second-layer graph convolutional network in turn; the activation functions set for the first-layer graph convolutional network and the second-layer graph convolutional network are different.

[0121] Step B5: The network device verifies the fault detection model based on the selected validation set and adjusts the meta-parameters.

[0122] It should be noted that a network fault detection neural network based on graph convolutional networks was trained and deployed at the edge base station. Figure 3 This application provides a network fault detection process; such as Figure 3 As shown:

[0123] Step 1: Collect raw network data from the LDC FAN network, i.e., actual business KPI data under different network conditions. To facilitate analysis, the collected sample data is normalized, and the analysis is based on the normalized data.

[0124] It should be noted that the process in step one occurs between the AP and the mobile terminal device.

[0125] Step 2: The edge platform receives the normalized data and generates a network performance metric table; the preprocessed small sample data is divided into training and validation datasets. These sets are fed into a feature selector to select the optimal feature combination. For feature selection, algorithms such as Extreme Gradient Boosting (XGBoost) and ReliefF can be used. The feature selector is deployed in the base station of the edge platform to process the dataset from the edge platform.

[0126] Step 3: After matching the dataset with the feature selector, the training dataset is used to train the model parameters of the network fault detection model. The network fault detection model is deployed in the base station of the edge platform.

[0127] Step 4: After training is complete, use the validation dataset to validate the network fault detection model and adjust the meta-parameters of the learning process (such as learning rate, number of iterations, etc.).

[0128] Repeat the above process until the model's performance reaches the preset target.

[0129] It should be noted that the input to the network fault detection model is the network-labeled KPI data, and the output is the predicted label of the network fault. The model input and output are transmitted between network nodes as important intermediate data.

[0130] This application designs a two-layer graph convolutional network model to implement a network fault detection algorithm. Each graph convolutional layer undergoes three processes: "transmit-receive-transform".

[0131] Transmission involves each node sending its own transformed feature information to its neighboring nodes. This step involves extracting and transforming the node's feature information.

[0132] Reception involves each node aggregating the feature information of its neighbors and fusing the local structural information of the nodes.

[0133] Transformation involves a nonlinear transformation of the information gathered earlier, thereby increasing the expressive power of the model.

[0134] It should be noted that the number of layers is the maximum number of jumps that each node can perceive.

[0135] Figure 4 This is a flowchart of a two-layer training model of a graph convolutional network provided in an embodiment of this application; as shown... Figure 4 As shown: The input of the network fault detection model is the network-labeled KPI data, and the output is the predicted fault label. In the first layer (layer 1) of the Graph Convolutional Network (GCN), each node obtains information from its neighbors. The information collection process of each node is independent, and all nodes collect information simultaneously. In the second layer (layer 2) of the GCN, neighboring nodes already have information about their neighbors (from the previous layer). This information is merged according to weight rules, and then the feature information of the current node is combined with the feature information of the neighboring nodes, and then a feature concatenation operation is performed before passing through another layer of the neural network.

[0136] The two graph convolutional network layers each use different activation functions; for example, the first layer uses ReLU, and the second layer uses Softmax. The model output is an M×N dimensional node feature matrix, where c and M are the predefined number of network fault categories, and N is the number of samples in the dataset. During GCN training, labeled samples from the training set are used to compute the cross-entropy loss function. This loss function serves as the model's optimization objective, and gradient descent is used to backpropagate errors through weight updates in each graph convolutional layer.

[0137] Figure 5 This is a flowchart illustrating a fault detection method provided in an embodiment of this application, as shown below. Figure 5 As shown, this method is applied to Figure 1 In the network shown, the method includes:

[0138] Step 501: The terminal node generates source data and sends the source data to the tail node.

[0139] Step 502: The tail node performs data aggregation and selects an appropriate transmission mode (carrier transmission, wireless transmission, dual-mode transmission) based on the actual network conditions.

[0140] Step 503: The tail node transmits the aggregated data to the head node.

[0141] Step 504: The head node receives network status information and forms a network fault performance evaluation index table, and inputs the network fault performance evaluation index table into the fault detection model based on graph convolutional network.

[0142] It should be noted that the parameters need to be adjusted and iterative optimization needs to be performed.

[0143] Step 505: Output fault labels from the head node.

[0144] Step 506: The head node uploads the fault label to the cloud management platform, which then adjusts the network resource configuration (i.e., the resource adjustment strategy) and sends the instructions back to the head node.

[0145] Embodiments of this application provide a network device that can be used to implement... Figure 2 The corresponding embodiment provides a fault detection method, referring to... Figure 6 As shown, network device 600 includes:

[0146] The first acquisition module 601 is used to acquire first information; wherein, the first information is used to indicate the network status of the transmission link between the network device and the first terminal;

[0147] The first processing module 602 is used to determine network fault performance evaluation index parameters based on the first information;

[0148] The first processing module 602 is used to input network fault performance evaluation index parameters into the fault detection model, and to detect the network status of the transmission link through the fault detection model to obtain fault information of the fault node; wherein, the fault node includes network devices and the first terminal; the fault detection model is a model based on a two-layer graph convolutional network.

[0149] In other embodiments of this application, the first sending module 603 is used to send fault information to the first platform so that the first platform can adjust the network resource configuration of the faulty node based on the fault information;

[0150] The first receiving module 604 is used to receive a resource adjustment message sent by the first platform; wherein the resource adjustment message includes a resource adjustment strategy.

[0151] In other embodiments of this application, the first processing module 602 is used to preprocess the data collected under different network states into a small sample dataset; divide the small sample dataset into a training set and a validation set; perform feature selection processing on the training set and the validation set to obtain the selected training set and the selected validation set; input the selected training set into the first layer graph convolutional network and the second layer graph convolutional network in sequence to obtain the parameters of the fault detection model output by the convolutional network; wherein, each network node in each layer graph convolutional network sends its own transformed feature information to its neighboring nodes, and each network node aggregates the transformed feature information sent by all neighboring nodes and transforms the aggregated feature information; and verify the trained fault detection model based on the selected validation set to adjust the meta-parameters.

[0152] In other embodiments of this application, in the first-layer graph convolutional network, all network nodes collect feature information of neighboring nodes; in the second-layer graph convolutional network, each network node performs feature concatenation operation on its own feature information and the feature information of neighboring nodes collected from the first-layer graph convolutional network, and then inputs the concatenated feature information into the first-layer graph convolutional network and the second-layer graph convolutional network in sequence; the activation functions set for the first-layer graph convolutional network and the second-layer graph convolutional network are different.

[0153] In other embodiments of this application, network faults include one or more of the following: node performance faults; device faults; application software faults; node hardware faults; network-level faults; network fault performance evaluation parameters include one or more of the following: inclusiveness parameters; sustainability parameters; credibility parameters; wherein, inclusiveness parameters include one or more of the following: network throughput; reference signal received strength; average signal-to-noise ratio; packet loss rate; sustainability parameters include one or more of the following: transmission delay; transmission redundancy; credibility parameters include one or more of the following: node spacing variance.

[0154] In other embodiments of this application, the first receiving module 602 is used to receive aggregated data sent by the first terminal; wherein the first terminal sends the aggregated data to the network device according to a transmission mode that matches the network state; the aggregated data is obtained by the first terminal aggregating at least one source data received; the at least one source data is generated by at least one second terminal; the transmission mode includes one of the following: carrier transmission, wireless transmission, and dual-mode transmission;

[0155] The first processing module 602 is used to collect first information during the process of receiving aggregated data.

[0156] The descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0157] It should be noted that, in the embodiments of this application, if the above-described fault detection method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a terminal device to execute all or part of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0158] The embodiments of this application provide a first platform, which can be used to implement the fault detection method provided in the embodiments of this application, with reference to... Figure 7 As shown, the first platform 700 includes:

[0159] The second receiving module 701 is used to receive fault information of the faulty node sent by the network device; wherein, the fault information is obtained by the network device inputting the network fault performance evaluation index parameters determined based on the first information obtained into the fault detection model, and detecting the network status of the transmission link between the network device and the first terminal through the fault detection model.

[0160] The second processing module 702 is used to adjust the network resource configuration of the faulty node based on the fault information;

[0161] The second sending module 703 is used to send a resource adjustment message to the network device; wherein the resource adjustment message includes a resource adjustment strategy.

[0162] The descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0163] It should be noted that, in the embodiments of this application, if the above-mentioned fault detection method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a terminal device to execute all or part of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, ROMs, magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0164] Figure 8 This is a schematic structural diagram of a communication device 800 provided in an embodiment of this application. The communication device may be a first device or a first platform. Figure 8 The communication device 800 shown includes a processor 810, which can call and run computer programs from memory to implement the methods in the embodiments of this application.

[0165] Optionally, such as Figure 8 As shown, the communication device 800 may further include a memory 820. The processor 810 can retrieve and run computer programs from the memory 820 to implement the methods described in this embodiment.

[0166] The memory 820 can be a separate device independent of the processor 810, or it can be integrated into the processor 810.

[0167] Optionally, such as Figure 8 As shown, the communication device 800 may also include a transceiver 830, and the processor 810 may control the transceiver 830 to communicate with other devices. Specifically, it may send information or data to other devices or receive information or data sent by other devices.

[0168] The transceiver 830 may include a transmitter and a receiver. The transceiver 830 may further include an antenna, and the number of antennas may be one or more.

[0169] Optionally, the communication device 800 may specifically be a network device in the embodiments of this application, and the communication device 800 may implement the corresponding processes implemented by the network device in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0170] Optionally, the communication device 800 may specifically be the first platform in the embodiments of this application, and the communication device 800 may implement the corresponding processes implemented by the first platform in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0171] In some embodiments, this application also provides a computer program product, including a computer program that can be executed by the processor 810 of the communication device 800 to perform the steps described in any of the foregoing methods.

[0172] It should be understood that the processor in the embodiments of this application may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor described above can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0173] As one embodiment, the processor may include one or more general-purpose central processing units (CPUs). Each of these processors may be a single-core processor or a multi-core processor. Here, "processor" may refer to one or more devices, circuits, and / or processing cores used for processing data (e.g., executing instructions).

[0174] It is understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be ROM, Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), or flash memory. The volatile memory can be Random Access Memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0175] This application also provides a computer-readable storage medium for storing computer programs.

[0176] Optionally, the computer-readable storage medium can be applied to the network device in the embodiments of this application, and the computer program causes the computer to execute the corresponding processes implemented by the network device in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0177] Optionally, the computer-readable storage medium can be applied to the first platform in the embodiments of this application, and the computer program causes the computer to execute the corresponding processes implemented by the first platform in the various methods of the embodiments of this application. For the sake of brevity, it will not be described in detail here.

[0178] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0179] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

[0180] The fault detection method, network device, first platform, computer-readable storage medium, and computer program product provided in the embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0181] It should be understood that the phrases "an embodiment," "an embodiment," "an embodiment of this application," "the foregoing embodiment," "some implementations," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, the phrases "an embodiment," "an embodiment," "an embodiment of this application," "the foregoing embodiment," "some implementations," or "some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments of this application are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0182] Unless otherwise specified, any step in the embodiments of this application performed by the network device / first platform may be performed by the processor of the network device / first platform. Unless otherwise specified, the embodiments of this application do not limit the order in which the network device / first platform performs the following steps. Furthermore, the methods used to process data in different embodiments may be the same or different methods.

[0183] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0184] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0185] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0186] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined to obtain new method embodiments without conflict. The features disclosed in the several product embodiments provided in this application can be arbitrarily combined to obtain new product embodiments without conflict. The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined to obtain new method embodiments or device embodiments without conflict.

[0187] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0188] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0189] The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0190] It should be noted that in the various embodiments involved in this application, all steps or some steps may be performed, as long as a complete technical solution can be formed.

[0191] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A fault detection method, characterized in that, Applied to network devices, the method includes: Obtain first information; wherein, the first information is used to indicate the network status of the transmission link between the network device and the first terminal; Based on the first information, determine the network fault performance evaluation index parameters; The network fault performance evaluation index parameters are input into the fault detection model, and the network status of the transmission link is detected by the fault detection model to obtain the fault information of the fault node; wherein, the fault node includes the network device and the first terminal; the fault detection model is a model based on a two-layer graph convolutional network.

2. The method according to claim 1, characterized in that, The method further includes: The fault information is sent to the first platform so that the first platform can adjust the network resource configuration of the faulty node based on the fault information. Receive a resource adjustment message sent by the first platform; wherein the resource adjustment message includes a resource adjustment strategy.

3. The method according to claim 1, characterized in that, Obtain a fault detection model based on a two-layer graph convolutional network, including: Data collected under different network conditions were preprocessed into small sample datasets; The small sample dataset is divided into a training set and a validation set; Feature selection processing is performed on the training set and validation set to obtain the selected training set and the selected validation set; The selected training set is sequentially input into the first layer graph convolutional network and the second layer graph convolutional network to obtain the parameters of the fault detection model output by the convolutional network; wherein, each network node in each layer graph convolutional network sends its own transformed feature information to the neighboring nodes, and each network node aggregates the transformed feature information sent by all neighboring nodes and transforms the aggregated feature information. The trained fault detection model is validated based on the selected validation set in order to adjust the meta-parameters.

4. The method according to claim 3, characterized in that, In the first layer of the graph convolutional network, all network nodes collect feature information from their neighboring nodes. In the second layer of the graph convolutional network, each network node performs a feature concatenation operation on its own feature information and the feature information of its neighboring nodes collected from the first layer of the graph convolutional network. The concatenated feature information is then input into the first layer of the graph convolutional network and the second layer of the graph convolutional network in sequence. The activation functions set for the first layer of the graph convolutional network and the second layer of the graph convolutional network are different.

5. The method according to claim 1, characterized in that, Network failures include one or more of the following: node performance failure; device failure; application software failure; node hardware failure; network-level failure. The network fault performance evaluation parameters include one or more of the following: inclusiveness parameters; sustainability parameters; credibility parameters; The inclusiveness parameters include one or more of the following: network throughput; reference signal received strength; average signal-to-noise ratio; packet loss rate; Sustainability parameters include one or more of the following: transmission delay; transmission redundancy; The confidence parameters include one or more of the following: node interval variance.

6. The method according to claim 1, characterized in that, The acquisition of the first information includes: The system receives aggregated data sent by a first terminal; wherein the first terminal sends the aggregated data to the network device according to a transmission mode that matches the network state; the aggregated data is obtained by the first terminal aggregating at least one source data received; the at least one source data is generated by at least one second terminal; the transmission mode includes one of the following: carrier transmission, wireless transmission, and dual-mode transmission; During the process of receiving the aggregated data, the first information is collected.

7. A fault detection method, characterized in that, Applied to a first platform, the method includes: Receive fault information of a faulty node sent by a network device; wherein, the fault information is obtained by the network device by inputting network fault performance evaluation index parameters determined based on the first information obtained into a fault detection model, and by detecting the network status of the transmission link between the network device and the first terminal through the fault detection model; Based on the fault information, adjust the network resource configuration of the faulty node; Send a resource adjustment message to the network device; wherein the resource adjustment message includes a resource adjustment strategy.

8. A network device, characterized in that, Network equipment includes: A first acquisition module is configured to acquire first information; wherein the first information is used to indicate the network status of the transmission link between the network device and the first terminal; The first processing module is used to determine network fault performance evaluation index parameters based on the first information. The first processing module is further configured to input the network fault performance evaluation index parameters into the fault detection model, and detect the network status of the transmission link through the fault detection model to obtain fault information of the fault node; wherein, the fault node includes the network device and the first terminal; the fault detection model is a model based on a two-layer graph convolutional network.

9. A first platform, characterized in that, The first platform includes: The second receiving module is used to receive fault information of a faulty node sent by the network device; wherein, the fault information is obtained by the network device by inputting network fault performance evaluation index parameters determined based on the first information into a fault detection model, and by detecting the network status of the transmission link between the network device and the first terminal through the fault detection model; The second processing module is used to adjust the network resource configuration of the faulty node based on the fault information. The second sending module is used to send a resource adjustment message to the network device; wherein the resource adjustment message includes a resource adjustment strategy.

10. A network device, characterized in that, The network device includes: The first memory is used to store executable instructions; The first processor, when executing executable instructions stored in the first memory, implements the fault detection method of any one of claims 1 to 6.

11. A first platform, characterized in that, The first platform includes: The second memory is used to store executable instructions; The second processor, when executing executable instructions stored in the second memory, implements the fault detection method of claim 7.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, which can be executed by one or more processors to implement the fault detection method of any one of claims 1 to 6, or the fault detection method of claim 7.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the fault detection method according to any one of claims 1 to 6, or the fault detection method according to claim 7.

Images