Equipment management method, electronic equipment, medium and product

Abstract

The embodiment of the invention provides a device management method, an electronic device, a medium and a product, and the method comprises the steps: extracting each node device in a target communication network, extracting the node attribute of each node device, and carrying out the similarity calculation between the node devices; and then clustering is performed according to the similarity of each node device, classification is performed according to a clustering analysis result, and the node devices belonging to the same class are divided into corresponding prediction hierarchies, so that preset hierarchies corresponding to each node device are determined. Through the equipment management method, a manager of the communication network can clearly know the hierarchical structure among the node equipment of the target communication network, and can more easily understand the topological structure of the target communication network by combining with the network topological relation of the target communication network, so that the communication efficiency is improved. Potential relations and laws in the network topology of the ToB communication network are better displayed, the understanding difficulty of equipment management is simplified, and network management and maintenance are more efficient and convenient.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of communication technologies, and in particular, to a device management method, an electronic device, a medium, and a product. Background Art

[0002] A communication network for enterprises (To Business, ToB) refers to a network used for business communication and information transfer within an enterprise. Network topology management is a common management method for ToB communication networks. Network topology refers to the connection relationships between various devices in the network and is an important basis for managing and maintaining the network.

[0003] In related technologies, the device management of the ToB communication network topology adopts the network topology management method of the communication professional network. This method involves the topology connection relationships between network devices, but only manages devices through the topology connection relationships between network devices and cannot understand the hierarchical relationships of various professional network devices and application devices. Therefore, there are problems of being unintuitive and difficult to understand in device management. How to simplify the understanding difficulty of device management is an urgent problem to be discussed and solved currently. Summary of the Invention

[0004] The embodiments of the present application provide a device management method, an electronic device, a medium, and a product, aiming to simplify the understanding difficulty of device management.

[0005] In a first aspect, the embodiments of the present application provide a device management method, and the method includes: extracting a plurality of node devices from a target communication network; extracting the node attributes of each of the node devices, and determining the similarity between each of the node devices; clustering each of the node devices according to the similarity to obtain a plurality of node clusters; and classifying the plurality of node clusters according to a preset classification model to determine the preset levels corresponding to each of the node devices.

[0006] In a second aspect, the embodiments of the present application provide an electronic device, including: at least one processor; at least one memory for storing at least one program; when at least one of the programs is executed by at least one of the processors, the device management method described in the first aspect is implemented.

[0007] In a third aspect, the embodiments of the present application provide a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to execute the device management method described in the first aspect.

[0008] Fourthly, an embodiment of the present application provides a computer program product, including a computer program or computer instructions, characterized in that the computer program or the computer instructions are stored in a computer-readable storage medium, a processor of a computer device reads the computer program or the computer instructions from the computer-readable storage medium, and the processor executes the computer program or the computer instructions, so that the computer device executes the device management method as described in the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 Schematic diagram of a communication network provided by an embodiment of the present application;

[0010] Figure 2 Flowchart of a device management method provided by an example of the present application;

[0011] Figure 3 Flowchart of constructing an attribute matrix provided by an embodiment of the present application;

[0012] Figure 4 Flowchart of constructing an attribute matrix provided by another embodiment of the present application;

[0013] Figure 5 Network topology connection relationship diagram of a steel plant communication network provided by an example of the present application;

[0014] Figure 6 Schematic diagram of the structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0016] It should be noted that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order from the module division in the device or the order in the flowchart. Terms such as "first" and "second" in the specification, claims and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence.

[0017] In the description of the embodiments of the present application, unless otherwise clearly defined, words such as "set", "installed", and "connected" should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above words in the embodiments of the present application in combination with the specific content of the technical solution.

[0018] In the embodiments of the present application, words such as "furthermore", "exemplarily", or "optionally" are used to represent examples, illustrations, or explanations, and should not be construed as being more preferred or having more advantages than other embodiments or design solutions. The use of words such as "furthermore", "exemplarily", or "optionally" aims to present relevant concepts in a specific manner.

[0019] The technical solutions of the embodiments of the present application can be applied to various communication systems, such as: Wideband Code Division Multiple Access (WCDMA) mobile communication systems, Evolved Universal Terrestrial Radio Access Network (E-UTRAN) systems, Next Generation Radio Access Network (NG-RAN) systems, Long Term Evolution (LTE) systems, Worldwide Interoperability For Microwave Access (WiMAX) communication systems, 5th Generation (5G) systems, such as New Radio Access Technology (NR), and future communication systems, such as 6G systems, etc.

[0020] The technical solutions of the embodiments of the present application can be applied to various communication technologies, such as microwave communication, optical wave communication, millimeter wave communication, etc. The embodiments of the present application do not limit the specific technologies and specific device forms adopted.

[0021] In the related art, a communication network for enterprises (To Business, ToB) refers to a network used for business communication and information transmission within an enterprise. Network topology management is a commonly used management method for ToB communication networks. Network topology refers to the connection relationships between various devices in a network and is an important basis for managing and maintaining the network.

[0022] The ToB communication network is applied to multiple industries such as manufacturing, ports, mines, and power. The network scope covers the management domain of the park / factory and the production domain of the workshop / site. There are many types of devices involved in the ToB communication network, such as wireless access devices, wired access devices, bearer network devices, core network devices, office devices, robotic arms, cameras, Programmable Logic Controllers (PLCs), Automated Guided Vehicles (AGVs), business processing devices, etc.; the number of devices involved in the ToB communication network is relatively large. For example, in scenarios such as steel / ports, a large number of cameras are often deployed. It can be seen that in various scenarios of the ToB communication network, various communication professional network devices and various operation terminal devices are usually involved, and their diversity and complexity exceed those of general communication professional networks.

[0023] In the related art, the device management of the ToB communication network topology adopts the network topology management method of the communication professional network. This method involves the topological connection relationship between network devices. However, only through the topological connection relationship between network devices for device management, it is impossible to understand the hierarchical relationship between various professional network devices and application devices. Therefore, when managing devices, there are problems of being not intuitive and difficult to understand. How to simplify the understanding difficulty of device management is an urgent problem to be discussed and solved currently.

[0024] Based on this, the embodiments of the present application provide a device management method, an electronic device, a medium, and a product. By extracting each node device in the target communication network, extracting the node attributes of each node device, and calculating the similarity between node devices; then clustering according to the similarity of each node device, and then classifying according to the clustering analysis result, and dividing the node devices belonging to the same class into the corresponding prediction levels, so as to determine the preset levels corresponding to each node device. Through the above device management method, the managers of the communication network can clearly know the hierarchical structure between each node device of the target communication network. Combining with the network topology relationship of the target communication network, it is easier to understand the topology structure of the target communication network, better display the potential relationships and rules in the network topology of the ToB communication network, simplify the understanding difficulty of device management, and make network management and maintenance more efficient and convenient. The device management method of the embodiments of the present application improves the managers' understanding of the network topology and enhances the device management efficiency by considering both network devices and application devices and analyzing the association and hierarchical structure of each node device in the overall network topology of the target communication network.

[0025] The following further elaborates on the embodiments of the present application with reference to the accompanying drawings.

[0026] Figure 1 is a schematic diagram of a communication network provided by an embodiment of the present application, asFigure 1 As shown, in an embodiment, by way of example, a communication network includes a plurality of network devices and application devices. Among them, a network device is a dedicated hardware device for connecting various nodes such as servers, personal computers (PCs), application terminals, etc. to form an information communication network. By way of example, the network device can be a router, a switch, an indoor baseband processing unit (BBU), a remote radio unit (RRU), a customer premise equipment (CPE), etc.; an application device is a terminal device for providing application services. By way of example, in an industrial scenario, the application device can be an operating device such as a camera, a PLC controller, an AGV, a robotic arm, etc., which is not specifically limited herein.

[0027] It can be understood that the number, type of devices and the communication relationship between devices in the communication system of this embodiment can be expanded and changed according to actual needs, which is not specifically limited herein.

[0028] Next, through an example, the device management method of the embodiment of the present application in Figure 1 the shown communication network will be specifically described:

[0029] The network topology of the target communication network is as Figure 1 shown. The network devices include Network Device 1, Network Device 2, Network Device 3, and Network Device 4, and the application devices include Application Device 1, Application Device 2, and Application Device 3.

[0030] Extract Network Device 1, Network Device 2, Network Device 3, Network Device 4, Application Device 1, Application Device 2, and Application Device 3 from the target communication network as node devices, and extract the node attributes corresponding to each node device such as Network Device 1, Network Device 2, Network Device 3, Network Device 4, Application Device 1, Application Device 2, and Application Device 3 based on the network topology of the target communication network. Among them, the node attributes can include, but are not limited to, any one or a combination of device name, device type, device location, device status, and device identifier.

[0031] According to the node attributes of each node device, determine the similarity between Network Device 1, Network Device 2, Network Device 3, Network Device 4, Application Device 1, Application Device 2, and Application Device 3 to obtain multiple node clusters. By way of example, perform clustering according to the similarity. Network Device 1 is Node Cluster 1, and Network Device 2, 3, and 4 are Node Cluster 2, and Application Device 1, 2, and 3 are Node Cluster 3.

[0032] Classify the three node clusters obtained above according to a preset classification model to determine the preset level corresponding to each node cluster, so as to determine the preset level corresponding to the node devices in each node cluster. Exemplarily, the preset classification model is a supervised classification algorithm; the preset levels are network levels, such as the core layer, the aggregation layer, the access layer, and the application layer; among them, through the supervised classification algorithm, it is determined that node cluster 1 belongs to the core layer, node cluster 2 belongs to the access layer, and node cluster 3 belongs to the application layer. Correspondingly, the level to which network device 1 belongs is the core layer, and the levels to which network devices 2, 3, and 4 belong are the access layer, and the levels to which application devices 1, 2, and 3 belong are the application layer.

[0033] In the above example, the network devices and application devices in the target communication network are classified by the device management method of this example, and the levels to which each device belongs are clarified, so that the managers of the communication network can clearly know the hierarchical structure among the node devices in the target communication network. Combining with the network topology relationship of the target communication network, it is easier to understand the topology of the target communication network, better display the potential relationships and rules in the network topology of the ToB communication network, simplify the understanding difficulty of device management, and make network management and maintenance more efficient and convenient.

[0034] Figure 2 This is a flowchart of the device management method provided by an example of this application. As Figure 2 shown, the device management method can be applied, but not limited to, in the management device of the communication system, or in the Figure 1 communication network shown. In the Figure 2 embodiment, the device management method can include, but is not limited to, steps S110, S120, S130, and S140.

[0035] Step S110: Extract multiple node devices from the target communication network.

[0036] Step S120: Extract the node attributes of each node device and determine the similarity between each node device.

[0037] Step S130: Cluster each node device according to the similarity to obtain multiple node clusters.

[0038] Step S140: Classify the multiple node clusters according to a preset classification model to determine the preset level corresponding to each node device.

[0039] Among them, in step S110: the target communication network refers to the communication network that the manager needs to manage the devices; the node device refers to the device corresponding to the node extracted based on the network topology of the target communication network, including network devices and application devices.

[0040] In step S120: The node attribute refers to the attribute corresponding to the node device for characterizing the situation of the node device. Exemplarily, the node attribute may include, but is not limited to, any one or a combination of device name, device type, device location, device status, and device identifier. The similarity refers to the index based on which the node devices are classified. Exemplarily, the similarity may be node attribute similarity, Jaccard similarity, etc.

[0041] In step S130: Clustering refers to dividing similar node devices together according to the similarity of each node device; a node cluster refers to a cluster composed of several similar node devices.

[0042] In step S140: The preset classification model refers to an algorithm model or rule strategy for classifying multiple node clusters according to a preset hierarchy. The preset hierarchy refers to the hierarchy set by the management personnel according to the scenario where the target communication network is located or the management requirements; Exemplarily, the preset hierarchy may be a network hierarchy including a core layer, an aggregation layer, an access layer, and an application layer, or may also be to divide the device levels according to the area where the device is located. For example, in a factory scenario, the preset hierarchy is workshop level and factory level; no specific limitation is made here.

[0043] The device management method in the above embodiments of the present application enables the management personnel of the communication network to clearly know the hierarchical structure among the node devices of the target communication network. Combining with the network topology relationship of the target communication network, it is easier to understand the topology structure of the target communication network, better display the potential relationships and rules in the network topology of the ToB communication network, simplify the understanding difficulty of device management, and make network management and maintenance more efficient and convenient. By considering both network devices and application devices, analyzing the associations and hierarchical structures of each node device in the overall network topology of the target communication network, the understanding degree of the management personnel on the network topology is improved, and the device management efficiency is enhanced.

[0044] In one embodiment, step S120 includes: extracting the node attributes of each node device to establish an attribute matrix; and determining the similarity between each node device according to the attribute matrix. Among them, the attribute matrix refers to a matrix constructed by the node attributes of multiple node devices. Each row of the attribute matrix is the node attributes of the same node device; if there are n network devices and application devices extracted from the target communication network, and each device extracts m node attributes, then an n*m attribute matrix can be established. Exemplarily, assume that a target communication network includes three node devices, namely node A, node B, and node C, and each node device includes two node attributes, then a 3*2 attribute matrix R can be obtained 3*2 This embodiment can facilitate subsequent data processing by extracting the node attributes of each node device to construct an attribute matrix.

[0045] In one embodiment, determining the similarity between each node device according to the attribute matrix includes: converting the node attributes in the attribute matrix into feature vectors through a pre-trained word vector model to obtain a node vector matrix; determining the similarity between each node device according to the node vector matrix. Exemplarily, the pre-trained word vector model can be Word2Vec (word to vector), Global Vectors for Word Representation (GloVe), FastText, etc., which is not specifically limited herein.

[0046] Exemplarily, it is also assumed that a target communication network includes three node devices, namely node A, node B, and node C, and each node device includes two node attributes, namely attribute 1 and attribute 2, then a 3*2 attribute matrix R can be obtained. 3*2 。

[0047] After converting each node attribute in the attribute matrix into a feature vector through Word2Vec, a node vector matrix can be obtained, and the node vector matrix can be expressed as:

[0048]

[0049] After obtaining the node vector matrix, use a preset similarity metric standard, such as cosine similarity, Euclidean distance, etc., to calculate the similarity between the node feature vectors. In this example, cosine similarity is used for similarity calculation. For every two node devices, by dividing the dot product of the vectors of the two node devices by the product of the vector lengths of the two node devices, the cosine value of the included angle between them can be obtained, so as to judge their similarity. The closer the cosine value is to 1, the closer the included angle is to 0 degrees, that is, the more similar the two vectors are. When the included angle is equal to 0, that is, the two vectors are equal. The formula is as follows:

[0050]

[0051] where A and B are the feature vectors of two different node devices respectively.

[0052] By converting each node attribute in the attribute matrix into a feature vector, similarity calculation can be performed more simply and quickly, thereby improving the overall efficiency of device management.

[0053] In one embodiment, before establishing the attribute matrix, the method further includes: obtaining a preset hierarchy; determining the node attribute type according to the preset hierarchy;

[0054] Correspondingly, an attribute matrix is established, including: selecting the node attributes corresponding to each node device and node attribute type from the extracted node attributes, and constructing an attribute matrix.

[0055] Figure 3 is a flowchart for constructing an attribute matrix provided by an embodiment of the present application. As Figure 3 shown, in this embodiment, first, all types of node attributes are extracted from each node device. Exemplarily, the device name, device type, device location, device status, and device identifier of each node device are extracted, and a data set is constructed according to all the node attributes of each extracted node device. Then, according to the preset hierarchy, the node attribute types required for clustering and classification are determined. Exemplarily, assuming that the preset hierarchy is a network hierarchy including a core layer, an aggregation layer, an access layer, and an application layer, at this time, the node attribute types based on which the node devices are clustered and classified are the device name, device identifier, and device type.

[0056] After determining the required node attribute types, then, from the data set containing all types of node attributes, the device name, device identifier, and device type of each node device are selected to construct an attribute matrix.

[0057] When it is necessary to manage node devices from multiple perspectives, for example, in addition to knowing the distribution of node devices at the network hierarchy level, it is also necessary to know the distribution of node devices in the region. At this time, the preset hierarchy is the hierarchical classification of the affiliated region, such as the workshop level and the factory level. At this time, from the data set, the device name, device number, and device location of each node device are selected to construct an attribute matrix.

[0058] Through the method of the embodiment of the present application, first, all the node attributes of each node device are extracted to construct a data set, and then the node attributes corresponding to the required types are extracted from the data set according to different preset hierarchies; only one extraction of node attributes is required to support the selection of node attributes in multiple different dimensions (preset hierarchies), and there is no need to re-extract the required node attributes from the node devices for each dimension, which simplifies the processing flow and improves the efficiency.

[0059] In another embodiment, before extracting the node attributes of each node device, the method further includes: obtaining the preset hierarchy; determining the node attribute types according to the preset hierarchy;

[0060] Correspondingly, when extracting the node attributes of each node device and establishing an attribute matrix, it includes: extracting the node attributes corresponding to each node device and the node attribute type according to the node attribute type, and constructing an attribute matrix.

[0061] Figure 4 is a flowchart for constructing an attribute matrix provided by another embodiment of the present application. As Figure 4As shown, in this embodiment, first determine the node attribute types required for clustering classification according to a preset hierarchy. Exemplarily, assume that the preset hierarchy is a network hierarchy including a core layer, an aggregation layer, an access layer, and an application layer. At this time, the node attribute types based on which the node devices are clustered and classified are device name, device identifier, and device type. Then extract the device name, device identifier, and device type from each node device to construct an attribute matrix.

[0062] Through the method of the embodiment of the present application, first determine the required node attribute types according to the preset hierarchy, and directly extract the required node attributes from each node device, which can reduce data caching and save space.

[0063] In another embodiment, step S120 includes: extracting the node attributes of each node device and the connection relationships between each node device; constructing an edge set according to the node attributes and connection relationships; and determining the similarity between each node device according to the edge set through the Jaccard coefficient.

[0064] Exemplarily, extract devices from the target communication network as node devices, such as routers, switches, BBU, RRU, video heads, CPEs, etc., to construct a node device set N = {n1, n2,...}. Then extract the node attributes of each node device. The node attributes may include device name, device type, device location, device status, and device identifier. For example, the node attributes of the base station PRRU1 (skin base station) and BBU in the coal mine roadway are shown in the following table:

[0065] Equipment Name Equipment Number Equipment Type Equipment Location Equipment Status PRRU1 1-2-3 Access Network PRRU Metallurgical Workshop Normal BBU 1-2-6 Access Network BBUU Metallurgical Workshop Normal

[0066] Among them, the device number is the device identifier. It should be noted that in addition to the device number, the device identifier can also be represented in other ways as long as it can uniquely correspond to the device.

[0067] Then extract the connection relationships between each node device as edges. For example, the connection between BBU and PRRU1, the connection between BBU and the switch, and the connection between switches, to construct an edge set E = {e1, e2,...}. The attributes of the edges may include source device, destination device, display name, connection type, etc. For example, the connection relationships between the base station BBU and PRRU1 in the coal mine roadway are shown in the following table:

[0068] Edge Name Source Equipment Destination Equipment Connection Type Link Status Link1 1-2-3 1-2-6 Optical Fiber Normal

[0069] Construct a network topology graph G = {N, E} corresponding to the target communication network according to the node device set N and the edge set E; based on the network topology graph G, calculate and determine the similarity between each node device through the Jaccard coefficient.

[0070] By extracting the node attributes of each node device in the target communication network, a set of node devices is obtained. The connection relationships between the node devices are extracted as edges to obtain an edge set. A network topology graph is constructed through the node device set and the edge set, and the similarity of each node device is calculated through the Jaccard coefficient. It can be applied to scenarios with a large number of device types and a large number of devices. In a scenario with a large number of node devices, the similarity of each node device can be quickly determined for node device classification, improving the overall efficiency.

[0071] In one embodiment, step S130 includes: constructing a similarity matrix according to the similarity, where each row of the similarity matrix represents the similarity between a node device and other node devices; determining a degree matrix according to the similarity matrix; and performing clustering through an unsupervised learning model based on the spectral clustering algorithm according to the similarity matrix and the degree matrix to obtain multiple node clusters.

[0072] Exemplarily, it is also assumed that a target communication network includes three node devices, namely node A, node B, and node C, and each node device includes two node attributes. The node vector matrix can be expressed as:

[0073]

[0074] Among them, the first row corresponds to node A, the second row corresponds to node B, and the third row corresponds to node C.

[0075] Through cosine similarity calculation, the similarity between each node device is obtained, and the similarity results between nodes A, B, and C are constructed into a similarity matrix W as follows:

[0076]

[0077] Among them, the first row corresponds to the similarity between node A and each node device, the second row corresponds to the similarity between node B and each node device, and the third row corresponds to the similarity between node C and each node device. The diagonal elements of the similarity matrix are always 1 because a node device is most similar to itself. The other elements of the similarity matrix are between 0 and 1, and the closer the value is to 1, the more similar the two node devices are in terms of this node attribute.

[0078] In this example, an unsupervised learning model based on the spectral clustering algorithm is used for clustering. By clustering the eigenvectors of the Laplacian matrix of the sample data, the purpose of clustering the sample data is achieved. The spectral clustering algorithm can map the data in the high-dimensional space to the low-dimensional space, and then use other clustering algorithms in the low-dimensional space, such as the k-means clustering algorithm (K-Means), for clustering, thereby improving the speed and the clustering effect. Specifically, the clustering process in this example is as follows:

[0079] Calculate the degree matrix D using the similarity matrix. D is a diagonal matrix of d i composed of n*n; where that is, d i is the sum of the elements in the i-th row of the similarity matrix W; the degree matrix D is as follows:

[0080]

[0081] Among them, in this example, the similarity matrix W is a 3*3 matrix. Correspondingly, n is 3, and D is a diagonal matrix of d i (i = 1, 2, 3) composed of 3*3.

[0082] Calculate the Laplacian matrix L = D - W, and calculate the eigenvectors (k column vectors) of the first k eigenvalues of the matrix L. Combine the k column vectors into a matrix U (U n*k ). Use the K-means or other classical clustering algorithms to cluster the n rows of data in the matrix U to obtain the node clusters C1, C2,..., C k .

[0083] In one embodiment, step S140 includes: determining category labels according to a preset hierarchy; assigning a category label to each node cluster to generate node cluster samples; training a preset classifier according to the node cluster samples to obtain a trained classifier; reclassifying multiple node clusters through the trained classifier to determine the preset hierarchy corresponding to each node device. Among them, the preset classifier can be a supervised classification algorithm.

[0084] Exemplarily, based on the clustering results obtained by clustering the node devices using the device management method provided in the above embodiments, a supervised classification algorithm is applied to reclassify the node clusters to determine the hierarchical relationship between the node devices.

[0085] First, assign a class label to each node cluster in the clustering result according to the preset hierarchy. For example, for the network topology of a ToB communication network, the preset hierarchy can be a network hierarchy including a core layer, an aggregation layer, an access layer, and an application layer. The node devices are divided into 4 layers according to their functions, namely the core layer, the aggregation layer, the access layer, and the application layer. In a ToB communication network, the 5th Generation Core (5GC) devices of the fifth-generation communication system belong to the core layer, the bearer network devices belong to the aggregation layer, the radio base stations belong to the access layer, and the application devices such as cameras and robotic arms belong to the application layer. Assign labels to each node cluster according to the known device classification, and divide the node clusters with assigned labels into a training set and a test set.

[0086] Select a suitable supervised classification algorithm according to the characteristics of the classification and the data type, such as decision tree, naive Bayes, support vector machine, etc.

[0087] Use the training set to train the selected supervised classification algorithm to obtain a trained classifier. Specifically, divide the clustering result set, that is, the multiple node clusters obtained by clustering, into a training set and a test set. Set labels for the data in the training set according to the core layer, aggregation layer, access layer, and application layer, and then input the training set into the classification algorithm for training.

[0088] After training, evaluate the performance of the trained classifier through the test set. The evaluation metrics can be accuracy, recall rate, F1 value, etc. Among them, the F1 value refers to the harmonic mean of the precision rate and the recall rate. According to the evaluation results, adjust the parameters of the classifier or select other classification algorithms.

[0089] Use the trained classifier to classify each node cluster in the clustering result, and use the classification result as the new class label. Determine the hierarchy corresponding to the node cluster through the classification label corresponding to the node cluster, so as to determine the hierarchy corresponding to the node devices in the node cluster.

[0090] By setting class labels, training a classifier with node cluster sample data, and then using the trained classifier to classify the node clusters; through supervised learning classification, the accuracy of classification can be improved, and in scenarios with a large amount of data, the reclustering of the clustering results can be completed more quickly and accurately, so as to determine the hierarchy to which each node device belongs.

[0091] In another embodiment, when the number of node clusters is less than a preset value, step S140 includes: determining the node attribute type according to a preset level; extracting the corresponding node attributes in each node cluster according to the node attribute type; matching the preset levels corresponding to each node cluster based on the corresponding node attributes in each node cluster; and determining the preset level corresponding to the node cluster to which each node device belongs as the preset level corresponding to each node device. Among them, matching the preset levels corresponding to each node cluster means matching the extracted node attributes and the preset levels through a regular expression.

[0092] Exemplarily, assume that the preset level is a network level including a core layer, an aggregation layer, an access layer, and an application layer. At this time, according to the predicted level, the node attribute type is determined to be the device type. Match the corresponding level according to the device type of the node devices in the node cluster; for example, the device type in a certain node cluster is ['Video1', 'Video2', 'Video3', 'Video4', 'Video5']. It is determined through matching that the devices in this node cluster belong to monitoring devices and belong to the application layer.

[0093] In the case of a small amount of data, re - classification is achieved by matching between node clusters and preset levels through fixed rules, which can simplify the calculation process and save computing resources.

[0094] The device management method of the present application will be specifically described as a whole through examples below. It can be understood that the following embodiments are all for better exemplarily explaining the device management method of the present application and are not specifically limited.

[0095] Example 1:

[0096] Assume that the communication network of a steel plant includes 5 monitoring cameras (Video1 - 5) on the production line, a 5G gateway, a New Radio (NR) base station, a User Port Function (UPF), and a monitoring server (Server).

[0097] The data acquisition module collects the following information:

[0098] Equipment Name Equipment Number Equipment Type Equipment Location Equipment Status PRRU1 1-2-3 Access Network PRRU Steel Rolling Workshop Normal PRRU2 1-2-4 Access Network PRRU Steel Rolling Workshop Normal BBU1 1-2-5 Access Network BBU Communication Machine Room Normal UPF1 3-1 Core Network Equipment Communication Machine Room Normal Server1 6-1 Monitoring Equipment Communication Machine Room Normal CPE1 1-2-6 Access Network Gateway Steel Rolling Workshop Normal CPE2 1-2-7 Access Network Gateway Steel Rolling Workshop Normal Video1 5-1 Monitoring Equipment Steel Rolling Workshop Normal Video2 5-2 Monitoring Equipment Steel Rolling Workshop Normal Video3 5-3 Monitoring Equipment Steel Rolling Workshop Normal Video4 5-4 Monitoring Equipment Steel Rolling Workshop Normal Video5 5-5 Monitoring Equipment Steel Rolling Workshop Normal

[0099] Figure 5 For the network topology connection relationship diagram of the steel plant communication network provided by an example of the present application, as Figure 5 shown, an attribute matrix is established according to the device name, device number, and device type attribute. The attribute matrix is as follows:

[0100] [['PRRU1', '1 - 2 - 3', 'access network PRRU'],

[0101] ['PRRU2', '1-2-4', 'Access network PRRU'],

[0102] ['BBU1', '1-2-5', 'Access network BBU'],

[0103] ['UPF1', '3-1', 'Core network device'],

[0104] ['Server1', '6-1','monitoring device'],

[0105] ['CPE1', '1-2-6', 'Access network gateway'],

[0106] ['CPE2', '1-2-7', 'Access network gateway'],

[0107] ['Video1', '5-1','monitoring device'],

[0108] ['Video2', '5-2','monitoring device'],

[0109] ['Video3', '5-3','monitoring device'],

[0110] ['Video4', '5-4','monitoring device'],

[0111] ['Video5', '5-5','monitoring device']。

[0112] Use Word2Vec to convert the attribute matrix into a node vector matrix, then use cosine similarity for similarity measurement to generate a similarity matrix, and use spectral clustering for node clustering to generate the following node clusters:

[0113] [['Video1', 'Video2', 'Video3', 'Video4', 'Video5'], ['CPE1', 'CPE1', 'PRRU1', 'PRRU2', 'BBU1'], ['UPF1'], ['Server1']]。

[0114] Since the data volume is not large, the preset regular expressions can be used to match the device types to determine the device hierarchy. For example, node cluster 1 ['Video1', 'Video2', 'Video3', 'Video4', 'Video5'] belongs to the application layer of monitoring devices, node cluster 2 ['CPE1', 'CPE1', 'PRRU1', 'PRRU2', 'BBU1'] belongs to the access layer of access devices, node cluster 3 ['UPF1'] belongs to the core layer of core network devices, and cluster 4 ['Server1'] belongs to the application layer of monitoring devices.

[0115] Example 2:

[0116] Taking the communication network of the steel plant provided in Example 1 as an example, based on the node devices and node attributes of the communication network of the steel plant in Example 1, three node attribute types of device name, device number, and device location are selected to establish an attribute matrix, and the attribute matrix is as follows:

[0117] [['PRRU1','1-2-3','Rolling workshop'],

[0118] ['PRRU2','1-2-4','Rolling workshop'],

[0119] ['BBU1','1-2-5','Communication machine room'],

[0120] ['UPF1','3-1','Communication machine room'],

[0121] ['Server1','6-1','Communication machine room'],

[0122] ['CPE1','1-2-6','Rolling workshop'],

[0123] ['CPE2','1-2-7','Rolling workshop'],

[0124] ['Video1','5-1','Rolling workshop'],

[0125] ['Video2','5-2','Rolling workshop'],

[0126] ['Video3','5-3','Rolling workshop'],

[0127] ['Video4','5-4','Rolling workshop'],

[0128] ['Video5','5-5','Rolling workshop']].

[0129] Use Word2Vec to convert the attribute matrix into a node vector matrix, then use cosine similarity for similarity measurement to generate a similarity matrix, and use spectral clustering for node clustering to generate the following node clusters:

[0130] [['Video1','Video2','Video3','Video4','Video5', 'CPE1','CPE1','PRRU1','PRRU2'],

[0131] ['BBU1','UPF1','Server1']].

[0132] Since the amount of data is not large, a preset regular expression can be used to match the device location to determine the area where the device is located. For example, node cluster 1 ['Video1', 'Video2', 'Video3', 'Video4', 'Video5', 'CPE1', 'CPE1', 'PRRU1', 'PRRU2'] is located in the rolling mill workshop and belongs to workshop-level equipment, and node cluster 2 ['BBU1', 'UPF1', 'Server1'] is located in the communication computer room and belongs to factory-level equipment.

[0133] Example 3:

[0134] Typical operations of smart ports include remote control of quay cranes, intelligent tallying, driverless container trucks, drone inspections, and sea-anchorage communications, etc., involving equipment such as PLC controllers, high-definition PTZ cameras, artificial intelligence (AI) recognition systems, monitoring and management platforms, quay cranes, driverless container trucks, drones, etc. There are many types and quantities of equipment. Assuming there are 880 application devices and 80 communication devices, that is, a total of 960 node devices.

[0135] Since the data scale is large, the Jaccard coefficient can be used to calculate the similarity to obtain the similarity matrix. The calculation formula for the elements of the similarity matrix is as follows:

[0136]

[0137] N(i) represents the set of neighbor node devices of node device i, |N(i)| represents the number of its neighbor node devices, N(j) represents the set of neighbor node devices of node device j, |N(j)| represents the number of its neighbors, |N(i) ∩ N(j)| is the number of the intersection of the neighbors of node device i and node device j, and |N(i) ∪ N(j)| is the number of the union of the neighbors of node device i and node device j. This metric expresses the node similarity by comparing the similarity of node neighbors, representing the similarity of each node device in the topological structure.

[0138] In this example, the hierarchical clustering method is used to generate the clustering results. Since the number of node devices is large, hierarchical labels can be assigned to 600 clustering result data, of which 200 are used as the training set and 200 are used as the test set. The support vector machine classification algorithm is selected for classifier training. The goal of the classification problem is to learn a linear classifier y = f(x) on the training set to minimize the error between the predicted label f(x) and the true label y. The following loss function is used:

[0139] l(y, f(x)) = max(0, 1 - y · f(x))

[0140] Among them, y is the true label and f(x) is the predicted label. If the prediction is correct, the error is 0; if the prediction is incorrect, the error is 1 - y·f(x). Here, y·f(x) represents the product between the true label and the predicted label.

[0141] The training results are as follows: the macro-precision (macro-P) reaches 0.8055, the macro-recall (macro-R) reaches 0.8111, and the macro F1-score (macro-F1) reaches 0.7919. According to the training results, it can be determined that the accuracy of the trained classifier basically meets the requirements.

[0142] Use the trained classifier to classify the clustering results to determine the hierarchical or regional distribution of the device groups, that is, the node clusters. The classification results of the port dataset are as follows:

[0143]

[0144]

[0145] Each example provided by this application performs clustering analysis through various types of similarity metrics, and then classifies according to the clustering analysis results to determine the hierarchical structure of each node device in the target communication network, making the network topology easy to understand, and better showing the potential relationships and rules in the ToB communication network topology, making network management and maintenance more efficient and convenient.

[0146] Figure 6 It is a schematic structural diagram of an electronic device provided by an embodiment of this application. As Figure 6 shown, the electronic device 2000 includes a memory 2100 and a processor 2200. The number of the memory 2100 and the processor 2200 can be one or more, Figure 6 taking one memory 2101 and one processor 2201 as an example; the memory 2101 and the processor 2201 in the network device can be connected through a bus or other means, Figure 6 taking the connection through a bus as an example.

[0147] The memory 2101, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the method provided by any embodiment of this application. The processor 2201 realizes the device management method provided by any of the above embodiments by running the software programs, instructions, and modules stored in the memory 2101.

[0148] The memory 2101 may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system and application programs required for at least one function. In addition, the memory 2101 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some instances, the memory 2101 further includes a memory remotely disposed relative to the processor 2201, and these remote memories may be connected to the device through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0149] An embodiment of the present application further provides a computer-readable storage medium storing computer-executable instructions for executing the device management method provided in any embodiment of the present application.

[0150] An embodiment of the present application further provides a computer program product including a computer program or computer instructions. The computer program or computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, so that the computer device executes the device management method provided in any embodiment of the present application.

[0151] The system architecture and application scenarios described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will know that with the evolution of the system architecture and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are equally applicable to similar technical problems.

[0152] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0153] In a hardware implementation, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a single physical component may have multiple functions, or a function or step may be executed by several physical components in cooperation. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVDs) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed by a computer. Additionally, as is well known to those of ordinary skill in the art, communication media typically embodies computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0154] The terms "component", "module", "system", etc. used in this specification are used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable, an execution thread, a program, or a computer. By way of illustration, both an application running on a computing device and the computing device can be components. One or more components can reside within a process or execution thread, and components can be located on one computer or distributed between two or more computers. Additionally, these components can execute from various computer-readable media having various data structures stored thereon. Components can communicate, for example, according to a signal having one or more data packets (e.g., data from two components interacting with each other from a local system, a distributed system, or a network, such as the Internet interacting with other systems via a signal) through local or remote processes.

[0155] Some embodiments of the present application have been illustrated above with reference to the accompanying drawings, which do not limit the scope of rights of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present application shall fall within the scope of rights of the present application.

Claims

1. A device management method, characterized in that, The method includes: extracting multiple node devices from a target communication network; extracting the node attributes of each of the node devices and determining the similarity between each of the node devices; clustering each of the node devices according to the similarity to obtain multiple node clusters; classifying the multiple node clusters according to a preset classification model to determine the preset levels corresponding to each of the node devices.

2. The device management method according to claim 1, wherein The extracting the node attributes of each of the node devices and determining the similarity between each of the node devices includes: extracting the node attributes of each of the node devices and establishing an attribute matrix; determining the similarity between each of the node devices according to the attribute matrix.

3. The device management method according to claim 2, wherein Before establishing the attribute matrix, the method further includes: obtaining the preset level; determining the node attribute types according to the preset level; The establishing the attribute matrix includes: selecting the node attributes corresponding to each of the node devices and the node attribute types from the extracted node attributes and constructing the attribute matrix.

4. The device management method according to claim 2, characterized in that Before extracting the node attributes of each of the node devices, the method further includes: obtaining the preset level; determining the node attribute types according to the preset level; The extracting the node attributes of each of the node devices and establishing an attribute matrix includes: extracting the node attributes corresponding to each of the node devices and the node attribute types according to the node attribute types and constructing the attribute matrix.

5. The device management method according to claim 2, wherein The determining the similarity between each of the node devices according to the attribute matrix includes: converting the node attributes in the attribute matrix into feature vectors through a pre-trained word vector model to obtain a node vector matrix; determining the similarity between each of the node devices according to the node vector matrix.

6. The device management method according to claim 1, wherein, The extracting the node attributes of each of the node devices and determining the similarity between each of the node devices includes: extracting the node attributes of each of the node devices and the connection relationships between each of the node devices; constructing an edge set according to the node attributes and the connection relationships; determining the similarity between each of the node devices according to the edge set through the Jaccard coefficient.

7. The device management method according to any one of claims 1 to 6, characterized in that The clustering each of the node devices according to the similarity to obtain multiple node clusters includes: constructing a similarity matrix according to the similarity, where each row of the similarity matrix represents the similarity between one node device and each of the other node devices; determining a degree matrix according to the similarity matrix; performing clustering through an unsupervised learning model based on a spectral clustering algorithm according to the similarity matrix and the degree matrix to obtain multiple node clusters.

8. The device management method according to claim 1, characterized in that, The classifying the multiple node clusters according to a preset classification model to determine the preset levels corresponding to each of the node devices includes: determining class labels according to the preset level; assigning one of the class labels to each node cluster to generate node cluster samples; training a preset classifier according to the node cluster samples to obtain a trained classifier; re-classifying the multiple node clusters through the trained classifier to determine the preset levels corresponding to each of the node devices.

9. The device management method according to claim 1, characterized in that When the number of the node clusters is less than a preset value, classifying the multiple node clusters according to a preset classification model to determine the corresponding preset levels of the respective node devices includes: Determining a node attribute type according to the preset level; Extracting the corresponding node attributes in each of the node clusters according to the node attribute type; Matching the corresponding preset levels of the respective node clusters based on the corresponding node attributes in each of the node clusters; Determining the preset level corresponding to the node cluster to which each node device belongs as the preset level corresponding to each node device.

10. An electronic device, characterized in that, Including: At least one processor; At least one memory for storing at least one program; When at least one of the programs is executed by at least one of the processors, implementing the device management method according to any one of claims 1 to 9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for executing the device management method according to any one of claims 1 to 9.

12. A computer program product, comprising a computer program or computer instructions, characterized in that, The computer program or the computer instructions are stored in a computer-readable storage medium, and a processor of a computer device reads the computer program or the computer instructions from the computer-readable storage medium, and the processor executes the computer program or the computer instructions, so that the computer device executes the device management method according to any one of claims 1 to 9.

Images