A network visualization method based on hierarchical layout and clustering algorithm

Abstract

The application relates to a network visualization method based on hierarchical layout and a clustering algorithm, and belongs to the field of data visualization.The method comprises the following steps: calling an API interface to obtain circuit data and processing the circuit data, using a dynamic clustering algorithm to perform clustering processing on the data, layering the data processed by the precondition clustering algorithm, each cluster serving as a layer, determining the sequence of each layer according to the structure of the clustering result, mapping the clustering result to hierarchical levels to form node sets of different layers, optimizing the node sequence of each layer to reduce the crossing of edges, calculating the coordinates of each node according to the layering and crossing reduction result, performing drawing according to the calculated node coordinates to convert the processed data into a topology diagram, and acquiring alarm data in real time, performing correlation analysis on the alarm data, and identifying potential fault causes.The method can efficiently process various complex network data and realize intuitive display and dynamic analysis of network structures.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of data visualization and network management, and particularly relates to a network visualization method based on hierarchical layout and clustering algorithm. BACKGROUND

[0002] With the development of society, network devices and data volume present an explosive growth, and network topology structure becomes more and more complex. Traditional network visualization methods often have difficulty in processing large-scale data, and have problems such as poor layout effect, slow data loading, poor interactivity, etc., which makes it difficult for users to intuitively understand the overall structure and key connections of the network.

[0003] Traditional network topology display methods are mostly force-directed layout, which can intuitively show the relationship between nodes, but for networks with obvious hierarchical structure, force-directed layout may not effectively highlight the relationship between levels; a large number of nodes and connections make the topology graph complex, making it difficult to intuitively display and analyze. SUMMARY

[0004] In view of the above deficiencies of the prior art, the purpose of the present application is to provide a network visualization method based on hierarchical layout and clustering algorithm, which can efficiently process various complex network data and realize intuitive display and dynamic analysis of network structure.

[0005] The present application proposes a network visualization method based on hierarchical layout and clustering algorithm, comprising: S1, calling an API interface to obtain circuit data and processing the data to adapt to clustering algorithm processing, using a dynamic clustering algorithm to process the data for clustering, and generating a clustering result;

[0006] S2, layering according to the data processed by the precondition clustering algorithm, each cluster being a layer; determining the order of each level according to the structure of the clustering result, sorting the levels from low to high according to the bandwidth and delay of the feature vector of the clustering algorithm, mapping the clustering result to the level, forming a node set of different levels, and storing the sorted data for the next step of optimization;

[0007] S3, optimizing the node order of each level to reduce the crossing of edges;

[0008] S4, calculating the coordinates of each node according to the results of layering and reducing crossing, comprising: obtaining a two-dimensional array according to the results of layering and reducing crossing, and then determining the horizontal coordinates of each layer according to the two-dimensional array and the pre-set node spacing; then processing each layer, calculating the vertical coordinates of each device according to the array length and the middle index of each layer, for the next step of operation;

[0009] S5, drawing according to the calculated node coordinates to convert the processed data into a topology graph supporting user interactive viewing;

[0010] S6, after drawing, acquiring alarm data in real time, performing correlation analysis on the alarm data, identifying potential fault causes, and labeling abnormal devices.

[0011] Further, in S1, the called circuit data is cleaned, converted and aggregated to make the processed data suitable for clustering algorithm processing.

[0012] Further, in S1, the dynamic clustering algorithm is used to cluster the data, including:

[0013] The data is preprocessed, and the preprocessed data is preliminarily clustered using a K-means algorithm to obtain K initial clusters. In each initial cluster obtained by the K-means algorithm processing, a density clustering DBSCAN is used to further refine the clustering to obtain a density clustering result. According to the density clustering result, a density clustering structure that the K-means fails to accurately cluster is found. According to the changes in node and connection attributes, the clustering result is dynamically updated. Iterative calculation is performed until the clustering result is stable, and the final clustering result is output.

[0014] Further, in S3, the barycenter method is used to reduce the intersection of edges, including:

[0015] (1) obtaining the sorted two-dimensional array layer data;

[0016] (2) scanning the layers from top to bottom

[0017] From the second layer to the last layer, the position of the first layer is first represented by an index;

[0018] For each node s of the i-th layer, where i = 2 to n, n is the total number of layers, the loop starts;

[0019] Calculate the barycenter value of node s = the average value of the positions of all adjacent nodes of node s in layer i-1;

[0020] Sort the nodes in layer i according to the barycenter value and update the positions;

[0021] (3) scanning the layers from bottom to top

[0022] For each node s of the i-th layer, where i = n-1 to 1;

[0023] Calculate the barycenter value of node s = the average value of the positions of all adjacent nodes of node s in layer i+1;

[0024] Sort the nodes in layer i according to the barycenter value and update the positions;

[0025] (4) repeat the above steps (2) and (3) until the number of intersections of edges no longer decreases.

[0026] Further, in S4, nodes of the same level are arranged horizontally, and nodes of different levels are arranged vertically.

[0027] Further, in S5, the page component of the topology graph includes: a networking component, an end-to-end component, a topology display component, a device attribute component, a link attribute component, a port traffic component, and an alarm data component.

[0028] Further, the networking component is used to provide data information query and perform networking display.

[0029] The end-to-end component is used to display the topology graph between two ends by selecting nodes or devices of the two ends respectively.

[0030] The topology display component is used to perform data processing and topology graph display.

[0031] The device attribute component is used to display related information of devices.

[0032] The link attribute component is used to display related attributes of links.

[0033] The port traffic component is used to display port traffic information of two ends of a link.

[0034] The alarm data component is used to display alarm-related information.

[0035] Further, in S5, the topology view supports a custom view mode and a multi-view synchronous display mode, wherein,

[0036] The custom view mode supports user free switching between global view and local view.

[0037] The multi-view synchronous display supports multi-view synchronous display function, supports users to simultaneously view multiple network topology views of different dimensions, and provides omnidirectional network monitoring.

[0038] Further, in S5, the topology graph supports user node dragging and placement, and the layout of the network topology graph is adjusted by dragging the nodes.

[0039] Further, in S6, when an abnormal situation of a device is detected, alarm information is collected and processed, an alarm correlation analysis model is established, correlation analysis is performed on the alarm information by combining machine learning algorithm and expert knowledge base, potential fault reasons are identified, and processing suggestions are provided to the user.

[0040] The present application has the following advantages:

[0041] 1. Efficient Data Preprocessing and Dynamic Clustering Algorithm Optimization: This invention proposes an efficient data preprocessing mechanism based on cutting-edge algorithms, capable of intelligently optimizing massive amounts of data through advanced task scheduling algorithms. A dynamic clustering algorithm combining K-means and DBSCAN is employed to accurately decompose and cluster complex network data in real time, achieving adaptive adjustment of network nodes and their connections. This mechanism can respond to real-time dynamic changes in network data and automatically optimize clustering results, thereby ensuring clear network structure hierarchy and efficient visualization.

[0042] 2. Network Visualization Computation Method Based on Hierarchical Layout: This invention utilizes advanced hierarchical layout algorithms to finely optimize networks with complex hierarchical structures, greatly enhancing the intuitiveness and interpretability of the network structure. Through multi-dimensional data layout optimization, this method makes the visualization of complex networks more concise and clear, allowing users to more easily understand the internal hierarchical relationships and overall architecture of the network, significantly improving the efficiency and accuracy of data analysis.

[0043] 3. Multi-view Dynamic View Management: This invention provides multi-view dynamic view management functionality, allowing users to freely switch between global and local views, and supporting the simultaneous display of multi-dimensional network topology diagrams. Through a real-time synchronization update mechanism, it enables comprehensive monitoring and management of the network. This technology not only enhances the system's interactivity and flexibility but also provides users with a multi-layered, multi-angle in-depth data analysis experience, significantly improving the convenience and practicality of network operations.

[0044] 4. Interface Rendering and Interaction Performance Optimization: In scenarios involving large-scale datasets, this invention employs a hierarchical folding computation technique, significantly reducing the amount of data to be rendered and thus substantially improving interface rendering efficiency. Through this optimization, users can still enjoy a smooth operating experience when processing massive amounts of data. The aesthetics of the interface design and the responsiveness of the interaction are fully guaranteed under high load conditions, further optimizing the overall user experience.

[0045] 5. Intelligent Alarm Correlation Analysis and Root Cause Diagnosis: This invention integrates intelligent alarm correlation analysis, which intelligently highlights alarm information during anomaly detection and provides precise root cause analysis and professional repair suggestions. Through automated alarm information correlation and analysis, users can quickly locate and resolve potential problems, significantly improving the efficiency of fault diagnosis and handling. This technology effectively reduces network maintenance costs and significantly improves system reliability and stability.

[0046] 6. Optimized Nodes and Connections: This invention avoids displaying all nodes at once on the initial page entry. Instead, it first displays the first few layers of nodes and connections, then asynchronously requests subsequent nodes, quickly displaying them when needed by the user, greatly improving rendering efficiency. For data processing, a dynamic clustering algorithm is used to automatically classify the data.

[0047] 7. Layout and Visualization: The layout of this invention adopts a hierarchical layout algorithm. Since the nodes are selected and displayed in layers, the problem of large amount of computation that may lead to performance bottlenecks can be solved, and too many nodes will not be calculated at once.

[0048] 8. Interaction Design: Users of this invention can interact with the graph (such as clicking on nodes to view detailed information, dragging nodes to rearrange them, etc.), customize the view, select the information to view, or open multiple windows at once for information comparison. Users can choose different view modes such as global view, local view, etc. User operations in one view (such as dragging, zooming in, zooming out, etc.) will be automatically updated in other views.

[0049] 9. Real-time fault monitoring: When an abnormal situation is detected, the present invention highlights the alarm information, introduces an intelligent alarm correlation analysis function, and combines the database to perform correlation analysis on the alarm information to provide the root cause of the fault and handling suggestions. Attached Figure Description

[0050] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. It is obvious that the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings.

[0051] Figure 1 This is a flowchart of a network visualization method based on hierarchical layout and clustering algorithms according to an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of a page component according to an embodiment of the present invention;

[0053] Figure 3 This is a flowchart illustrating the hierarchical layout of an embodiment of the present invention;

[0054] Figure 4 This is a flowchart of the center-of-gravity method according to an embodiment of the present invention. Detailed Implementation

[0055] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0056] Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts disclosed in this invention.

[0057] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The terms "installed," "connected," and "linked" should be interpreted broadly; for example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0058] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of methods and systems consistent with some aspects of the invention as detailed in the appended claims.

[0059] This invention proposes a network visualization method based on hierarchical layout and clustering algorithms. This method involves displaying device topology diagrams based on circuit data and realizing large-scale network data visualization through hierarchical layout and dynamic clustering algorithms. It is suitable for the display and analysis of various complex network structures, such as IoT, 5G networks, and data center networks.

[0060] like Figure 1As shown, the network visualization method based on hierarchical layout and clustering algorithm provided by this invention includes the following steps:

[0061] S1 calls the API interface to obtain circuit data and processes the data to be suitable for clustering algorithm processing. The dynamic clustering algorithm is used to cluster the data and generate clustering results.

[0062] This step begins with data preprocessing: circuit data is obtained by calling API interfaces, and then cleaned, transformed, and aggregated to add anomaly detection and automatic repair functions. Next, distributed computing is executed: the backend data uses the MapReduce and Apache Flink frameworks for large-scale data processing, optimizing the computational task scheduling algorithm to improve processing efficiency. A dynamic clustering algorithm is also performed: combining K-means and DBSCAN algorithms, the cluster centers and clustering results are dynamically adjusted by continuously updating the attribute data of nodes and links.

[0063] S11, Data Preprocessing

[0064] By calling the API interface to obtain circuit data, the circuit data is cleaned, transformed, and aggregated to make the processed data suitable for clustering algorithm processing, and to add abnormal data detection and automatic repair functions.

[0065] For missing bandwidth and latency data, select records with a large number of missing values ​​to delete, and use statistical methods to identify and process outliers to ensure data quality.

[0066] New features are generated based on the node labels, bandwidth, and latency of the nodes and connections.

[0067] Standardize the selected features and combine them into a feature vector:

[0068] Feature vector = [label, bandwidth, latency].

[0069] S12, Initial Clustering (K-means)

[0070] The preprocessed data is initially clustered using the K-means algorithm to obtain K initial clusters. This step enables rapid grouping of nodes, helping to reduce the computational complexity of DBSCAN.

[0071] The computational complexity of DBSCAN is typically O(N). 2 Because for each data point, it is necessary to calculate the distance to all other points to determine whether the points in the neighborhood satisfy the density condition, the result is N. 2By using K-means preliminary clustering, data points are quickly divided into K clusters. This allows DBSCAN density clustering to be performed within each cluster, rather than across the entire dataset. In summary, K-means preliminary clustering can reduce the computational complexity of DBSCAN from O(N^2) to O(N^2). 2 ) reduced to O(N 2 / K).

[0072] S13, Density Clustering (DBSCAN)

[0073] For each initial cluster obtained by the K-means algorithm, DBSCAN is used to further refine the clusters to obtain density clustering results. Based on the density clustering results, density clustering structures that K-means failed to accurately cluster can be identified.

[0074] S14, Dynamic Adjustment

[0075] The clustering results are dynamically updated based on changes in node and connection attributes. The distances between nodes and connection weights are recalculated, and the cluster centers and clustering results are adjusted accordingly.

[0076] S15, iterate the calculation until the clustering results are stable, and output the final clustering results.

[0077] In summary, this invention employs a dynamic clustering algorithm: first, K-means algorithm is used for preliminary clustering, and then DBSCAN is used to further refine the clustering. This process can automatically adjust the clustering results according to the real-time changes in network data, ensuring the hierarchical nature of the network structure and the display effect.

[0078] S2, firstly, node layering is performed: the data processed by the precondition clustering algorithm is layered, with each cluster serving as a layer. Then, hierarchical sorting is performed: based on the structure of the clustering results, the order of each layer is determined, and the layers are sorted from low to high according to the bandwidth and latency of the clustering algorithm's feature vectors. The clustering results are mapped to hierarchical levels, forming node sets at different levels. The sorted data is then stored for further optimization.

[0079] S3 optimizes the node order at each level to reduce edge intersections. Minimize edge intersections to improve aesthetics. Ideally, edge paths should be straight lines.

[0080] Specifically, such as Figure 4 As shown, the centroid method is used to reduce edge intersections, including:

[0081] (1) Obtain the sorted two-dimensional array layer data.

[0082] (2) Scanning layers from top to bottom

[0083] The loop starts from the second level and continues to the last level, with the position of the first level indicated by a subscript.

[0084] For each node s in the i-th layer, where i = 2 to n, and n is the total number of layers, the loop begins; for example, if node s = 1 and the connected nodes in the previous layer are 1, 3, and 5, then the calculated centroid value is 3. After all the centroid values ​​of the nodes in this layer have been calculated, they are reordered, and the positions are initialized using subscripts for easy calculation.

[0085] The centroid value of node s is calculated as the average of the positions of all its neighboring nodes in layer i-1.

[0086] Sort the nodes in layer i according to their centroid values.

[0087] (3) Scanning layers from bottom to top (using a similar calculation method as scanning from top to bottom)

[0088] For each node s in the i-th layer, where i = n-1 to 1;

[0089] The centroid value of node s is calculated as the average of the positions of all its neighboring nodes in layer i+1.

[0090] Sort the nodes in layer i according to their centroid values.

[0091] (4) Repeat steps (2) and (3) above until the number of edge intersections no longer decreases.

[0092] S4, based on the results of layering and reducing intersections, calculate the coordinates of each node.

[0093] Specifically, based on the results of layering and reducing cross-intersection, a two-dimensional array can be obtained. Then, based on the two-dimensional array and the pre-set node spacing, the x-coordinate of each layer can be determined. Then, for each layer, the array length and middle index are processed. If the length is even, then middle = length / 2 - 1 (rounded down and subtracted by 1, because indices usually start from 0); if the length is odd, then middle = (length - 1) / 2.

[0094] (1) If the array length is even:

[0095] If inds is less than or equal to middle, then the device is located in the upper half, and its vertical axis position (y) is calculated using the formula: y = 60 + (inds - middle) * 120. Here, 60 is assumed to be the y-coordinate reference for the first device in the upper half.

[0096] If inds is greater than middle, the device is located in the lower half, and its vertical position y is calculated as: y = -60 + (inds - middle) * 120. It's important to note that (inds - middle) is actually equal to (inds - middle - 1) because the counting starts from the first device in the lower half.

[0097] (2) If the array length is odd:

[0098] y is calculated directly using a symmetrical distribution method. The central device is located at y = 0 (or any other reference point). The formula for calculating the y value of other devices is: y = (inds - middle) * 120. (Note that the reference point should be adjusted if the center is not 0).

[0099] In this way, the present invention can calculate the ordinate of each device in order to proceed with the next step of the operation.

[0100] In this step, nodes at the same level are arranged horizontally, and nodes at different levels are arranged vertically.

[0101] S5: Draws a graph based on the calculated node coordinates, ensuring edges are as straight and short as possible, while maintaining a balanced layout. The processed data is then converted into a topology graph, which supports interactive viewing by the user.

[0102] Through a series of complex calculations, the processed data is transformed into an intuitive and easy-to-understand topology diagram. This calculation process not only requires high accuracy and real-time performance but also careful consideration of the topology layout, the connections between nodes, and the ease of user interaction. The hierarchical interface display of the topology component is another major highlight. Through a carefully designed user interface and interaction mechanism, users can intuitively view the topology diagram and understand the network structure and status. Simultaneously, the topology component supports simultaneous display of multiple views, such as intelligent alarm handling: when an anomaly is detected, alarm information is highlighted, and root cause analysis and remediation suggestions are provided.

[0103] like Figure 2 As shown, the page components of the topology diagram include: networking components, end-to-end components, topology display components, device attribute components, link attribute components, port traffic components, and alarm data components.

[0104] Page Interaction Process: This invention aims to clarify the logical difference between network topology and end-to-end topology. Network topology, from a global perspective, displays the layout and connections of the entire network through a hierarchical topology diagram. End-to-end topology, on the other hand, focuses on the connections between specific nodes or devices, revealing specific details of the network through a detailed network link diagram.

[0105] During the interaction, users select the province using the interface, and the system immediately retrieves the corresponding province's network data from the interface layer, presenting it visually as a topology map in the display component. This process requires not only efficient data retrieval but also accurate and real-time rendering of the topology map. The devices and links within the display component are not merely static graphical elements but also carriers of interactive functions. These elements are interconnected; users can trigger detailed information display for devices or links by clicking. Furthermore, the connections between devices (i.e., links) support a collapsible function, allowing users to adjust and optimize the topology map as needed for a clearer view and a more convenient analysis experience. Alarm information is also directly displayed on the interface; once an anomaly is detected, an alarm notification will be immediately triggered on the interface. Alarm information is presented to the user in an intuitive way.

[0106] The components are described below.

[0107] The networking component is used to provide data information queries and display network topologies. Specifically, the networking component provides data information queries and displays network topologies for certain provinces.

[0108] Compared to networking components, end-to-end components add the option to select nodes or devices at both ends. By selecting nodes or devices at each end, the topology between these two ends can be displayed.

[0109] The topology display component is used to perform data processing and display topology diagrams.

[0110] Specifically, the topology component, as an indispensable core component of the entire system, bears the key tasks of data processing and topology graph display. Its primary responsibility is to perform fine-grained processing of the massive amounts of data returned by the interface layer. This process involves multiple stages such as data parsing, cleaning, transformation, and aggregation, ensuring the accuracy and integrity of the data and laying a solid foundation for subsequent calculations and display. After data processing, the topology component first uses WebGL technology for graphics rendering, leveraging the parallel computing capabilities of the GPU to achieve efficient graphics rendering, optimize rendering algorithms, and reduce the amount of graphics computation. For loading, it uses chunked loading and progressive rendering: gradually loading and rendering large-scale data, introducing preloading technology to ensure the smoothness of the user interaction process. For layout, it uses a hierarchical layout, such as... Figure 3 As shown.

[0111] The device properties component is used to display relevant information about the device after interaction, making it easy for users to view on the interface.

[0112] The Link Attributes component is used to display the relevant attributes of a link.

[0113] The port traffic component is used to display port traffic information at both ends of the link.

[0114] The alarm data component is used to display alarm-related information.

[0115] In an embodiment of the present invention, the topology view supports a custom view mode and a multi-view synchronous display mode.

[0116] Custom view mode: It supports users to freely switch between the global view and the local view.

[0117] Multi-view synchronous display: It supports the multi-view synchronous display function, allows users to view multiple network topology views of different dimensions simultaneously, updates in real-time synchronization, and provides all-round network monitoring. The multi-views are synchronized in real-time. Operations of users in one view (such as dragging, zooming in, zooming out, etc.) will be automatically updated in other views, providing a multi-view linkage analysis function. Users can compare and analyze between different views. It also supports viewing various information.

[0118] In the present invention, the topology diagram supports users to drag and place nodes. By dragging nodes, the layout of the network topology diagram can be adjusted, enhancing the interaction function.

[0119] Regarding node dragging and placement:

[0120] (1) All nodes on the interface support users to customize dragging and placement to meet the viewing habits and needs of different users.

[0121] (2) By dragging nodes, users can more intuitively adjust the layout of the network topology diagram, improving the efficiency of viewing and analysis.

[0122] In terms of interface display and style design, the present invention has refined the positions of device nodes and circuits. The logic and style details of the interface display are designed to ensure that users can intuitively and clearly understand the network topology structure.

[0123] (1) Device node display logic

[0124] Considering that the data volume of the subsequent layers of device nodes is huge, by default, these layers will be in a collapsed state to avoid overcrowding of the interface.

[0125] When the user clicks on the corresponding collapsible node, its subsequent device layers will be expanded and displayed. This interaction method helps users to more clearly focus on the current part of the network of concern. <​​​Each layer of device nodes will be distinguished by different icons to ensure users can quickly identify the node's level and type. Devices within the same server room will be grouped together to clearly indicate their spatial relationships. The icon design will be simple and clear, while maintaining visual consistency and aesthetics.

[0128] S6: After drawing the graph, acquire alarm data in real time, perform correlation analysis on the alarm data, identify potential causes of failure, and mark abnormal devices.

[0129] Specifically, when an abnormal situation is detected in the device, alarm information is collected and processed, and the alarm information is highlighted. An intelligent alarm correlation analysis function is introduced, establishing an alarm correlation analysis model. Combining machine learning algorithms and an expert knowledge base, the alarm information is correlated to identify potential causes of failure and provide users with detailed handling suggestions to help them quickly resolve problems.

[0130] In summary, through refined interface display logic and style design, this invention provides users with a clear, intuitive, and highly interactive network topology view. This will greatly improve user efficiency and experience when managing and analyzing complex network structures.

[0131] This invention proposes a network visualization method based on hierarchical layout and clustering algorithms to realize a circuit topology visualization scheme, employing the following technical solutions:

[0132] 1. The layered display design allows users to clearly see the relationship between different levels, and the layout uses a hierarchical layout algorithm.

[0133] 2. Introduce a custom view mode, allowing users to select different windows to view different information.

[0134] 3. Alarm traffic monitoring and analysis: The interface displays advanced device and link information in real time, and provides root cause analysis and remediation suggestions based on alarm data. Furthermore, the interface offers various convenient features such as dynamic zoom and interactive information prompts, search and filtering functions, and responsive design to further enhance the user experience and aesthetic appeal.

[0135] The method of this invention introduces hierarchical layout algorithms and clustering algorithms into the field of network visualization to achieve more efficient and aesthetically pleasing network topology display. It is suitable for networks with a clear hierarchical structure; the clustering algorithm, by grouping nodes and simplifying the data structure, makes the network graph display more hierarchical and logical.

[0136] This invention employs dynamic clustering analysis in data processing to group nodes of the same type together. For the interface layout, a hierarchical layout algorithm is used to better adapt the interface to various scenarios. In terms of interaction, it provides customizable view modes, allowing users to choose how to view the topology diagram. It also offers multi-view synchronous display, where user actions in one view (such as dragging, zooming, etc.) automatically update in other views, and multi-view linkage analysis, allowing users to compare and analyze different views. Users can also view port traffic and other related information. Real-time alarm monitoring provides intelligent alarm correlation analysis, offering users root cause analysis and troubleshooting suggestions.

[0137] This invention not only efficiently processes various complex network data, enabling intuitive visualization and dynamic analysis of network structures, but also adapts to multiple network types, including IoT, 5G, and data center networks, by combining a clear hierarchical layout with dynamic optimization of clustering algorithms. While ensuring the display of the overall network structure, it highlights key nodes and their hierarchical relationships, thereby improving user analysis efficiency and operational experience.

[0138] The network visualization method based on hierarchical layout and clustering algorithm provided by this invention can be applied to the following scenarios: development of telecommunications-related services, relationship and pattern analysis of complex network structures, and network topology monitoring and analysis.

[0139] (1) Building a topology map from a large amount of data: By adopting dynamic clustering and hierarchical layout, the device type can be quickly analyzed and then automatically laid out, which greatly improves the efficiency compared to traditional mapping.

[0140] (2) Multi-view analysis and comparison: Users can open different views on the page to compare and view the data, which also meets the needs of different users who want to view different information. Users can choose the observation window at their own discretion.

[0141] (3) Real-time monitoring: For faulty devices or links, the interface will quickly highlight them in red. Users can click on the alarm to obtain an analysis of the root cause of the fault and handling suggestions.

[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A network visualization method based on hierarchical layout and clustering algorithms, characterized in that, include: S1, call the API interface to obtain circuit data and process the data to be suitable for clustering algorithm processing, use dynamic clustering algorithm to cluster the data and generate clustering results; S2, the data processed by the precondition clustering algorithm is layered, with each cluster as a layer; the order of each layer is determined according to the structure of the clustering result; the layers are sorted from low to high according to the bandwidth and latency of the feature vector of the clustering algorithm; the clustering result is mapped to the layer level to form a set of nodes at different layers; the sorted data is then stored for further optimization. S3 optimizes the node order at each level to reduce edge intersections; S4. Based on the results of layering and reducing cross intersections, calculate the coordinates of each node, including: obtaining a two-dimensional array based on the results of layering and reducing cross intersections, and then determining the horizontal coordinate of each layer based on the two-dimensional array and the pre-set node spacing; then processing each layer and calculating the vertical coordinate of each device based on the array length and intermediate index of each layer for the next step of operation. S5. Draw a graph based on the calculated node coordinates to convert the processed data into a topology graph, which supports interactive viewing by the user. S6: After drawing the graph, acquire alarm data in real time, perform correlation analysis on the alarm data, identify potential causes of failure, and mark abnormal devices.

2. The network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S1, the called circuit data is cleaned, transformed, and aggregated to make the processed data suitable for clustering algorithm processing.

3. The network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S1, the process of clustering data using a dynamic clustering algorithm includes: The data is preprocessed, and the K-means algorithm is used to perform initial clustering on the preprocessed data to obtain K initial clusters. In each initial cluster obtained by the K-means algorithm, density clustering (DBSCAN) is used to further refine the clusters to obtain density clustering results. Based on the density clustering results, density clustering structures that K-means failed to accurately cluster are identified. The clustering results are dynamically updated according to changes in node and connection attributes. Iterative calculations are performed until the clustering results stabilize, and the final clustering results are output.

4. The network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S3, the centroid method is used to reduce edge intersections, including: (1) Obtain the sorted two-dimensional array layer data; (2) Scanning layers from top to bottom The loop starts from the second level and continues to the last level, with the position of the first level indicated by a subscript. For each node s in the i-th layer, where i = 2 to n and n is the total number of layers, start the loop; The centroid value of node s is calculated as the average of the positions of all its neighboring nodes in layer i-1. Sort the nodes in layer i according to their centroid values; (3) Scanning layers from bottom to top For each node s in the i-th layer, where i = n-1 to 1; The centroid value of node s is calculated as the average of the positions of all its neighboring nodes in layer i+1. Sort the nodes in layer i according to their centroid values ​​and then... (4) Repeat steps (2) and (3) above until the number of edge intersections no longer decreases.

5. The network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S4, nodes at the same level are arranged horizontally, and nodes at different levels are arranged vertically.

6. The network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S5, the page components of the topology diagram include: networking components, end-to-end components, topology display components, device attribute components, link attribute components, port traffic components, and alarm data components.

7. A network visualization method based on hierarchical layout and clustering algorithm according to claim 6, characterized in that, The networking component is used to provide data information query and to display the networking configuration; The end-to-end component is used to display the topology between the two ends by selecting nodes or devices at each end respectively; The topology display component is used to perform data processing and display topology graphs. The device attribute component is used to display relevant information about the device; The link attribute component is used to display the relevant attributes of the link; The port traffic component is used to display port traffic information at both ends of the link; The alarm data component is used to display alarm-related information.

8. The network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S5, the topology map supports custom view modes and multi-view synchronous display modes, wherein, Custom view mode: allows users to freely switch between global and local views; Multi-view synchronous display: Supports multi-view synchronous display, allowing users to view multiple network topology views from different dimensions simultaneously, with real-time synchronous updates, providing comprehensive network monitoring.

9. A network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S5, the topology graph allows users to drag and place nodes, adjusting the layout of the network topology graph by dragging and dropping nodes.

10. A network visualization method based on hierarchical layout and clustering algorithm according to claim 1, characterized in that, In S6, when an abnormal situation is detected in the device, alarm information is collected and processed, an alarm correlation analysis model is established, and machine learning algorithms and expert knowledge bases are combined to perform correlation analysis on the alarm information, identify potential causes of failure, and provide users with handling suggestions.

Images