Network partitioning methods, devices, equipment and storage media

Abstract

This application provides a network partitioning processing method, apparatus, device, and storage medium, relating to the field of communication technology. The method includes: determining a transition matrix of a target network based on traffic data of the target network; determining candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix; and determining a target logical subnet based on the modularity corresponding to each candidate logical subnet. The solution of this application improves the accuracy of network partitioning.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a network partitioning processing method, apparatus, device and storage medium. Background Technology

[0002] As research into complex networks deepens, it has been discovered that complex networks possess properties such as self-organization, self-similarity, attractors, small-world characteristics, and scale-free operation. With the continuous development of the internet and its applications, various network attack techniques are emerging, making network security issues increasingly serious. Currently, the main network security tools used include firewalls, antivirus software, and intrusion detection systems. However, firewalls, antivirus software, and intrusion detection systems are all boundary-oriented, requiring the division of the system into internal and external areas to isolate and protect the interior.

[0003] Therefore, how to more accurately implement network partitioning has become an urgent problem to be solved. Summary of the Invention

[0004] The purpose of this application is to provide a network partitioning processing method, apparatus, device, and storage medium to improve the accuracy of network partitioning.

[0005] To achieve the above objectives, embodiments of this application provide a network partitioning method, executed by a network device, comprising:

[0006] Based on the traffic data of the target network, determine the transition matrix of the target network;

[0007] Based on the eigenvalues ​​of the transition matrix, candidate logical subnets of the target network are determined;

[0008] The target logical subnet is determined based on the modularity of each candidate logical subnet.

[0009] Optionally, determining the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix includes:

[0010] Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined.

[0011] The candidate logical subnet is determined based on the sign of each element in the target feature vector;

[0012] The target eigenvalue is the eigenvalue of the entire transition matrix excluding the largest eigenvalue.

[0013] Optionally, each of the candidate logical subnets includes: a first candidate logical subnet and a second candidate logical subnet;

[0014] Determining the candidate logical subnet based on the sign of each element in the target feature vector includes:

[0015] If the sign of the current element is positive, the network node corresponding to the current element is assigned to the first candidate logical subnet;

[0016] If the sign of the current element is negative, the network node corresponding to the current element is assigned to the second candidate logical subnet.

[0017] Optionally, before determining the target logical subnet based on the modularity corresponding to each of the candidate logical subnets, the method further includes:

[0018] Based on the structure of the target network corresponding to the current candidate logical subnet, determine the first number of edges between nodes within different communities and the second number of edges between different communities;

[0019] Based on the first number of edges and the second number of edges, the modularity of the current candidate logical subnet is calculated.

[0020] Optionally, before determining the target logical subnet based on the modularity corresponding to each of the candidate logical subnets, the method further includes:

[0021] Based on the structure of the target network corresponding to the current candidate logical subnet, determine the weights of all edges;

[0022] The modularity of the current candidate logical subnet is calculated based on the weights.

[0023] Optionally, determining the target logical subnet based on the modularity corresponding to each of the candidate logical subnets includes:

[0024] The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

[0025] Optionally, after determining the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix, the method further includes:

[0026] If the modularity of each candidate logical subnet is less than or equal to the modularity obtained from the previous partitioning of the target network, the partitioning of the target network shall be stopped.

[0027] To achieve the above objectives, embodiments of this application provide a network partitioning processing apparatus, comprising:

[0028] The first processing module is used to determine the transition matrix of the target network based on the traffic data of the target network;

[0029] The second processing module is used to determine the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix.

[0030] The third processing module is used to determine the target logical subnet based on the module degree corresponding to each candidate logical subnet.

[0031] Optionally, the second processing module is further configured to:

[0032] Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined.

[0033] The candidate logical subnet is determined based on the sign of each element in the target feature vector;

[0034] The target eigenvalue is the eigenvalue of the entire transition matrix excluding the largest eigenvalue.

[0035] Optionally, each of the candidate logical subnets includes: a first candidate logical subnet and a second candidate logical subnet;

[0036] The second processing module is also used for:

[0037] If the sign of the current element is positive, the network node corresponding to the current element is assigned to the first candidate logical subnet;

[0038] If the sign of the current element is negative, the network node corresponding to the current element is assigned to the second candidate logical subnet.

[0039] Optionally, the device further includes:

[0040] The fourth processing module is used to determine the first number of edges between nodes within different communities and the second number of edges between different communities based on the structure of the target network corresponding to the current candidate logical subnet.

[0041] The fifth processing module is used to calculate the modularity of the current candidate logical subnet based on the first number of edges and the second number of edges.

[0042] Optionally, the device further includes:

[0043] The sixth processing module is used to determine the weights of all edges based on the structure of the target network corresponding to the current candidate logical subnet;

[0044] The seventh processing module is used to calculate the modularity of the current candidate logical subnet based on the weights.

[0045] Optionally, the third processing module is further configured to:

[0046] The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

[0047] Optionally, the device further includes:

[0048] The eighth processing module is used to stop partitioning the target network when the modularity of each candidate logical subnet is less than or equal to the modularity obtained from the previous partitioning of the target network.

[0049] To achieve the above objectives, embodiments of this application provide a network device including a processor, wherein the processor is used for:

[0050] Based on the traffic data of the target network, determine the transition matrix of the target network;

[0051] Based on the eigenvalues ​​of the transition matrix, candidate logical subnets of the target network are determined;

[0052] The target logical subnet is determined based on the modularity of each candidate logical subnet.

[0053] To achieve the above objectives, embodiments of this application provide a network device, including a transceiver, a processor, a memory, and a program or instructions stored in the memory and executable on the processor; when the processor executes the program or instructions, it implements the network partitioning processing method described above.

[0054] To achieve the above objectives, embodiments of this application provide a readable storage medium having a program or instructions stored thereon, which, when executed by a processor, implement the steps in the network partitioning processing method described above.

[0055] To achieve the above objectives, embodiments of this application provide a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the network partitioning processing method described above.

[0056] The beneficial effects of the above technical solution in this application are as follows:

[0057] The method in this embodiment first determines the transition matrix of the target network based on the traffic data of the target network; then, it determines the candidate logical subnets of the target network based on the eigenvalues ​​of the determined transition matrix; finally, it determines the target logical subnet of the target network based on the modularity corresponding to each candidate logical subnet, thus completing the partitioning of the target network, improving the accuracy of network partitioning, and ensuring the efficiency of partitioning. Attached Figure Description

[0058] Figure 1 This is a flowchart of a network partitioning method according to an embodiment of this application;

[0059] Figure 2 This is a schematic diagram of the module structure of the network partitioning processing device according to an embodiment of this application;

[0060] Figure 3 This is a structural diagram of the network-side device according to an embodiment of this application;

[0061] Figure 4 This is a structural diagram of a network-side device according to another embodiment of this application. Detailed Implementation

[0062] To make the technical problems, technical solutions and advantages of this application clearer, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments.

[0063] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0064] In the various embodiments of this application, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0065] In addition, the terms "system" and "network" are often used interchangeably in this article.

[0066] In the embodiments provided in this application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0067] For ease of understanding, the following describes some aspects of the embodiments of this application:

[0068] Network security situational awareness has become a hot research topic in the field of network security due to its ability to proactively defend against attacks based on the overall security status of the network system. While network security situational awareness can accurately provide security strategies, the sheer size and complexity of networks lead to the state explosion problem in attack behavior prediction and analysis, preventing the provision of real-time security defense strategies. To address the state explosion problem in network security situational awareness, community detection is applied to the network to improve its real-time performance.

[0069] Community detection is an important research area in complex networks. It aims to discover community structures within complex networks using specific data. As a complex network, community detection algorithms can be used to analyze the network system's structure, thereby deepening our understanding of the network. Currently, there are several main types of community detection algorithms.

[0070] I. Graph Clustering Algorithm

[0071] The Karnofsky Performances (KL) algorithm is an optimization method based on graph partitioning and a typical example of iterative bisection. It introduces a gain function ΔM, which represents the difference between the number of edges within two communities and the number of edges between communities. This algorithm is designed for partitioning networks with two communities of known sizes. As a bisection algorithm, it requires knowledge of the number of communities to determine the termination condition. The KL algorithm is applicable to networks with various structures and offers high accuracy. However, it requires prior knowledge of the final number of communities and the number of nodes in each community. Furthermore, it needs to calculate the gain function after swapping all node pairs to determine the node pairs to be swapped, resulting in a high computational complexity of O(N^3).

[0072] Spectral partitioning is a method for community detection that utilizes the eigenvalues ​​and eigenvectors of the Laplacian matrix of a graph. It originates from solving the graph segmentation problem and employs a fast iterative approach to partition the network. Like the Karnofsky Performance Work (KL) algorithm, spectral partitioning requires prior knowledge of the number of communities and can only partition two communities at a time. This method is only effective for networks with a clear community structure; if the community structure is not obvious, it often fails to yield ideal partitioning results.

[0073] II. Hierarchical Clustering

[0074] Hierarchical clustering, belonging to the field of sociology, is an algorithm that analyzes the similarity between social networks or the strength of connections between edges. It is one of the classic community detection algorithms, utilizing the network's topology to build a hierarchical tree, thus constructing a hierarchical community structure. Depending on how edges are handled, hierarchical clustering is divided into partitioning and agglomerative methods. Partitioning hierarchical clustering follows a top-down approach. It first considers all nodes in the network as a single community, then continuously removes the "weakest" edges, dividing the network into two parts, repeating this process until the network is divided into multiple communities. In contrast, agglomerative hierarchical clustering follows a bottom-up approach. It first considers each node in the network as a separate community, calculates the similarity between node pairs, and merges the nodes with the highest similarity until all nodes in the network are assigned to the same community. However, the above algorithms are agglomerative algorithms based on a greedy approach, only accepting the better candidate solution in the current situation during the merging process.

[0075] Modularity-based optimization methods use modularity as the objective function and the modularity function as the metric to select the community structure with the highest modularity as the partitioning result. The Louvain algorithm is a representative heuristic modularity optimization method that can quickly detect communities in the network without requiring prior knowledge of the number of communities. However, the Louvain algorithm is also based on a greedy strategy, which can lead to local optima. Among modularity-based optimization methods, there is also a large and important class of algorithms based on intelligent optimization, including simulated annealing, evolutionary algorithms, and particle swarm optimization.

[0076] However, community detection algorithms have low accuracy. The traditional spectral bisection method mentioned above usually extracts the community structure of the network by continuously partitioning the network according to the distribution of elements in a specific feature vector. However, this method cannot guarantee that each partition is optimal. Other optimized community detection algorithms, while improving accuracy, neglect complexity and algorithm efficiency, and cannot balance accuracy and efficiency.

[0077] like Figure 1 As shown, a network partitioning method according to an embodiment of this application is executed by a network device and includes:

[0078] Step 11: Determine the transition matrix of the target network based on the traffic data of the target network;

[0079] Step 12: Determine the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix;

[0080] Step 13: Determine the target logical subnet based on the modularity corresponding to each candidate logical subnet.

[0081] Following steps 11-13 above, the network device first determines the transition matrix of the target network based on the traffic data of the target network; then, it determines the candidate logical subnets of the target network based on the eigenvalues ​​of the determined transition matrix; finally, it determines the target logical subnet of the target network based on the modularity corresponding to each candidate logical subnet, thus completing the partitioning of the target network, improving the accuracy of network partitioning, and ensuring the efficiency of partitioning.

[0082] It should be noted that in this embodiment, a candidate logical subnet and a target logical subnet can also be referred to as a community.

[0083] Optionally, in this embodiment, the traffic data of the target network is obtained through traffic acquisition technology. Traffic acquisition technology is the foundation for network node interaction data and provides a data source for traffic analysis. In this way, the tightness of the connections between nodes within the network can be understood from the traffic data, enabling more accurate partitioning.

[0084] Furthermore, it should be understood that the transition matrix can be used to analyze network traffic data, detect network security threats, and provide network performance data in both Chinese and English. Therefore, in this embodiment, the transition matrix is ​​first determined for the target network based on its traffic data, and then the logical subnet is determined based on the eigenvalues ​​of the transition matrix.

[0085] Determining the transition matrix involves: first, identifying the physical nodes in the target network. Physical nodes are crucial elements of the target network and can be computers, routers, switches, etc., each with a unique IP address; then, combining the traffic data of the target network, organizing the physical nodes into a matrix. The transition matrix describes the traffic transfer patterns between nodes.

[0086] Optionally, determining the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix includes:

[0087] Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined.

[0088] The candidate logical subnet is determined based on the sign of each element in the target feature vector;

[0089] The target eigenvalue is the eigenvalue of the entire transition matrix excluding the largest eigenvalue.

[0090] Here, the characteristic equation is derived from the generalized characteristic equation As=λDs. In the generalized characteristic equation, A is the adjacency matrix of the network, D is the diagonal matrix of the network, λ is the eigenvalue, and s is the eigenvector corresponding to λ. When the network is connected, D is invertible, and the generalized characteristic equation is transformed into Ts=λs, where T is the transition matrix of the network, T=D′A.

[0091] Furthermore, regarding the modularity Q, the logical subnets of the network can be obtained using the traditional spectral bisection method, leading to the derivation that the magnitude of Q is proportional to λ (e.g., ...). Thus, Q reaches its maximum value when λ is the largest eigenvalue of the characteristic equation. However, λ cannot be the largest eigenvalue of the characteristic equation because s = (1,1,…,1) must be an eigenvector of the characteristic equation. According to the Perron-Frobenius theorem, the eigenvalue corresponding to this eigenvector must be the largest eigenvalue of the characteristic equation. However, this eigenvector cannot satisfy a constraint condition in the derivation process. Therefore, in this embodiment, the target eigenvalues ​​are the eigenvalues ​​of the transition matrix other than the largest eigenvalue. Based on Ts = λs, and each target eigenvalue of T (excluding the largest eigenvalue), the target eigenvector corresponding to each target eigenvalue is obtained by solving the characteristic equation.

[0092] For a target feature value, multiple target feature vectors are determined, such as s1, s2, ..., sk (k is an integer greater than or equal to 1). Then, for each target feature vector, a candidate logical subnet for that target feature vector is determined.

[0093] Optionally, in this embodiment, each candidate logical subnet includes: a first candidate logical subnet and a second candidate logical subnet. Thus, by partitioning the target network once, two target logical subnets are obtained, which are candidate logical subnets for the same target feature value.

[0094] Optionally, in this embodiment, determining the candidate logical subnet based on the sign of each element in the target feature vector includes:

[0095] If the sign of the current element is positive, the network node corresponding to the current element is assigned to the first candidate logical subnet;

[0096] If the sign of the current element is negative, the network node corresponding to the current element is assigned to the second candidate logical subnet.

[0097] That is, for one of the multiple target feature vectors, such as s1, the network is divided into two candidate logical subnets according to the signs of each element in s1.

[0098] Thus, in this embodiment, during a partitioning process, since the transition matrix has multiple target feature values, and each target feature value can determine two candidate logical subnets, which can be understood as a set of candidate logical subnets, the target logical subnet is finally determined based on the modularity among the resulting multiple sets of candidate logical subnets. The modularity Q is a standard for measuring the quality of network partitioning.

[0099] Optionally, in this embodiment, before determining the target logical subnet based on the modularity corresponding to each candidate logical subnet, the method further includes:

[0100] Based on the structure of the target network corresponding to the current candidate logical subnet, determine the first number of edges between nodes within different communities and the second number of edges between different communities;

[0101] Based on the first number of edges and the second number of edges, the modularity of the current candidate logical subnet is calculated.

[0102] In other words, among multiple candidate logical subnets, one candidate logical subnet is selected as the current candidate logical subnet. Based on the structure of the target network formed by these two candidate logical subnets, the first number of edges between nodes within different communities and the second number of edges between different communities are determined. Then, based on the first number of edges and the second number of edges, the modularity of these two candidate logical subnets is calculated.

[0103] The first edge count is for a single community. For community i, the first edge count is the number of edges between nodes within community i. If there are p communities, then p first edge counts are determined. The second edge count is for two communities, representing the number of edges between nodes in communities i and j.

[0104] Specifically, the formula for calculating the modularity Q is: Among them, a i a represents the proportion of edges connected to nodes within community i out of the total number of edges. i =∑ j e ij e ij e represents the proportion of edges connecting nodes in community i to nodes in community j out of the total number of edges. ii This represents the proportion of edges within community i to the total number of edges. The modularity value represents the sum of the differences between the proportion of edges within communities to the total number of edges in the network and the proportion of the degree of nodes within communities to the total number of edges. The total number of edges, often simply called the total edge count, is the number of edges between all nodes, regardless of community.

[0105] Of course, the formula for calculating Q above can also be transformed into formula... m is the total number of edges; n is the number of nodes in the target network; A ij This indicates whether there is an edge between node i and node j; 0 indicates an edge exists, and 1 indicates no edge exists. i The degree of a node, δ(c) i ,c j The modulus indicates whether node i and node j belong to the same community; a value of 1 indicates they belong to the same community, and a value of 0 indicates they belong to the same community. Modularity can be understood as the proportion of edges falling within the same community minus the expected value of randomly assigning these edges to the same community. Its range is between -0.5 and 1. The closer the modularity is to 1, the tighter the connections between nodes within the community, the more obvious the community structure, and the better the community partitioning effect. Therefore, maximizing the modularity can optimize the network community partitioning result.

[0106] Modularity is defined for undirected, unweighted networks, while network traffic data needs to consider the frequency of traffic interactions. The network traffic data matrix represents a directed, weighted graph. Therefore, optionally, in this embodiment, before determining the target logical subnet based on the modularity corresponding to each candidate logical subnet, the following steps are also included:

[0107] Based on the structure of the target network corresponding to the current candidate logical subnet, determine the weights of all edges;

[0108] The modularity of the current candidate logical subnet is calculated based on the weights.

[0109] In other words, among multiple candidate logical subnets, one candidate logical subnet is selected as the current candidate logical subnet. Based on the structure of the target network formed by these two candidate logical subnets, the weights of all edges are determined, and then the modularity of these two candidate logical subnets is calculated based on the weights.

[0110] Specifically, Where W is the sum of the weights of all edges, w ij v represents the weight of the edge connecting node i and node j. i The strength of node i is represented by the sum of the weights of all edges connected to node i.

[0111] Optionally, in this embodiment, determining the target logical subnet based on the modularity corresponding to each candidate logical subnet includes:

[0112] The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

[0113] That is, after calculating the modularity of multiple candidate logic subnets, the candidate logic subnet corresponding to the maximum modularity can be used as the target logic subnet.

[0114] It should also be noted that in this embodiment, the target network can be the entire network or a logical subnet of the entire network after one or more partitions.

[0115] Optionally, in this embodiment, after determining the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix, the method further includes:

[0116] If the modularity of each candidate logical subnet is less than or equal to the modularity obtained from the previous partitioning of the target network, the partitioning of the target network shall be stopped.

[0117] Thus, the entire network is partitioned, and the entire network is used as the target network for the first partition, resulting in two logical subnets (target logical subnets), each corresponding to a community. Subsequently, the logical subnets are used as target networks for further partitioning, and this process is repeated until the partitioning reaches the cutoff condition, which is that the modularity of all candidate logical subnets of the target network is less than or equal to the modularity obtained from the previous partitioning of the target network.

[0118] Since the goal of this embodiment is to maximize modularity, if the partitioning of a certain logical subnet cannot achieve a greater modularity, then the partitioning of that subnet is stopped.

[0119] It should also be noted that the method in this embodiment is based on the partitioning process of a community detection algorithm. This algorithm can divide the network into several communities based on different characteristics. Thus, according to different user needs and the different service functions of nodes, the network system can be segmented, revealing structural features in functional modules or connection relationships. Network system community detection mainly observes the tightness of connections between nodes within the network system. Based on the cognitive physics characteristics of nodes, and through large-scale network topology simulation, it can be found that viruses can spread rapidly from critical nodes to the entire network, much faster than propagation originating from edge nodes. By using networked data mining methods to uncover community networks and backbone networks, protection targets can be determined at the appropriate granularity based on network size, the nature of security threats, and activity status. This makes protection strategies adaptive and targeted, reducing protection costs and improving protection efficiency.

[0120] Compared with traditional partitioning, the method of this embodiment not only has a high-quality community structure, but also has high efficiency. It can effectively detect communities in the network and improve the accuracy of partitioning structure while ensuring efficiency.

[0121] like Figure 2 As shown in the figure, a network partitioning processing apparatus according to an embodiment of this application includes:

[0122] The first processing module 210 is used to determine the transition matrix of the target network based on the traffic data of the target network;

[0123] The second processing module 220 is used to determine the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix.

[0124] The third processing module 230 is used to determine the target logical subnet based on the module degree corresponding to each candidate logical subnet.

[0125] Optionally, the second processing module is further configured to:

[0126] Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined.

[0127] The candidate logical subnet is determined based on the sign of each element in the target feature vector;

[0128] The target eigenvalue is the eigenvalue of the entire transition matrix excluding the largest eigenvalue.

[0129] Optionally, each of the candidate logical subnets includes: a first candidate logical subnet and a second candidate logical subnet;

[0130] The second processing module is also used for:

[0131] If the sign of the current element is positive, the network node corresponding to the current element is assigned to the first candidate logical subnet;

[0132] If the sign of the current element is negative, the network node corresponding to the current element is assigned to the second candidate logical subnet.

[0133] Optionally, the device further includes:

[0134] The fourth processing module is used to determine the first number of edges between nodes within different communities and the second number of edges between different communities based on the structure of the target network corresponding to the current candidate logical subnet.

[0135] The fifth processing module is used to calculate the modularity of the current candidate logical subnet based on the first number of edges and the second number of edges.

[0136] Optionally, the device further includes:

[0137] The sixth processing module is used to determine the weights of all edges based on the structure of the target network corresponding to the current candidate logical subnet;

[0138] The seventh processing module is used to calculate the modularity of the current candidate logical subnet based on the weights.

[0139] Optionally, the third processing module is further configured to:

[0140] The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

[0141] Optionally, the device further includes:

[0142] The eighth processing module is used to stop partitioning the target network when the modularity of each candidate logical subnet is less than or equal to the modularity obtained from the previous partitioning of the target network.

[0143] It should be noted that this device is an apparatus that applies the above-described network partitioning processing method. The implementation of the above-described method embodiments is applicable to this device and can achieve the same technical effect.

[0144] like Figure 3 As shown, a network device 300 according to an embodiment of this application includes a processor 310 and a transceiver 320, wherein the processor is used for:

[0145] Based on the traffic data of the target network, determine the transition matrix of the target network;

[0146] Based on the eigenvalues ​​of the transition matrix, candidate logical subnets of the target network are determined;

[0147] The target logical subnet is determined based on the modularity of each candidate logical subnet.

[0148] Optionally, the processor is further configured to:

[0149] Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined.

[0150] The candidate logical subnet is determined based on the sign of each element in the target feature vector;

[0151] The target eigenvalue is the eigenvalue of the entire transition matrix excluding the largest eigenvalue.

[0152] Optionally, each of the candidate logical subnets includes: a first candidate logical subnet and a second candidate logical subnet;

[0153] The processor is also used for:

[0154] If the sign of the current element is positive, the network node corresponding to the current element is assigned to the first candidate logical subnet;

[0155] If the sign of the current element is negative, the network node corresponding to the current element is assigned to the second candidate logical subnet.

[0156] Optionally, the processor is further configured to:

[0157] Based on the structure of the target network corresponding to the current candidate logical subnet, determine the first number of edges between nodes within different communities and the second number of edges between different communities;

[0158] Based on the first number of edges and the second number of edges, the modularity of the current candidate logical subnet is calculated.

[0159] Optionally, the processor is further configured to:

[0160] Based on the structure of the target network corresponding to the current candidate logical subnet, determine the weights of all edges;

[0161] The modularity of the current candidate logical subnet is calculated based on the weights.

[0162] Optionally, the processor is further configured to:

[0163] The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

[0164] Optionally, the processor is further configured to:

[0165] If the modularity of each candidate logical subnet is less than or equal to the modularity obtained from the previous partitioning of the target network, the partitioning of the target network shall be stopped.

[0166] The network device in this embodiment first determines the transition matrix of the target network based on the traffic data of the target network; then, it determines the candidate logical subnets of the target network based on the eigenvalues ​​of the determined transition matrix; finally, it determines the target logical subnet of the target network based on the modularity corresponding to each candidate logical subnet, thus completing the partitioning of the target network, improving the accuracy of network partitioning, and ensuring the efficiency of partitioning.

[0167] Another embodiment of this application provides a network device, such as... Figure 4 As shown, it includes a transceiver 410, a processor 400, a memory 420, and a program or instructions stored in the memory 420 and executable on the processor 400; when the processor 400 executes the program or instructions, it implements the network partitioning processing method applied to the network device described above.

[0168] The transceiver 410 is used to receive and send data under the control of the processor 400.

[0169] Among them, Figure 4In this context, the bus architecture may include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 400) and memory (memory 420). The bus architecture may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 410 may be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium. The processor 400 is responsible for managing the bus architecture and general processing, and the memory 420 may store data used by the processor 400 during operation.

[0170] An embodiment of this application provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the steps in the network partitioning process described above and achieve the same technical effect. To avoid repetition, further details are omitted here.

[0171] The processor mentioned above is the processor in the network device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0172] This application also provides a computer program product, including computer instructions, which, when executed by a processor, implement the above-described... Figure 1 The various processes of the method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0173] It should be further noted that many of the functional components described in this specification are referred to as modules in order to emphasize the independence of their implementation.

[0174] In this embodiment, the module can be implemented in software so that it can be executed by various types of processors. For example, an identified executable code module may include one or more physical or logical blocks of computer instructions, which may be constructed as objects, procedures, or functions. Nevertheless, the executable code of the identified module does not need to be physically located together, but may include different instructions stored in different bits, which, when logically combined, constitute the module and achieve the module's intended purpose.

[0175] In practice, an executable code module can be a single instruction or many instructions, and can even be distributed across multiple different code segments, different programs, and across multiple memory devices. Similarly, operational data can be identified within the module and can be implemented in any suitable form and organized within any suitable type of data structure. This operational data can be collected as a single dataset or distributed across different locations (including different storage devices), and can exist, at least in part, solely as electronic signals within the system or network.

[0176] When a module can be implemented using software, considering the current level of hardware technology, modules that can be implemented in software can be implemented using hardware circuits by those skilled in the art to achieve the corresponding functions, without considering cost. These hardware circuits include conventional very-large-scale integrated circuits (VLSI) or gate arrays, as well as existing semiconductors such as logic chips and transistors, or other discrete components. Modules can also be implemented using programmable hardware devices, such as field-programmable gate arrays, programmable array logic, and programmable logic devices.

[0177] The exemplary embodiments described above are with reference to the accompanying drawings. Many different forms and embodiments are feasible without departing from the spirit and teachings of this application. Therefore, this application should not be construed as limiting the exemplary embodiments set forth herein. Rather, these exemplary embodiments are provided to make this application complete and convey the scope of this application to those skilled in the art. In these drawings, component dimensions and relative dimensions may be exaggerated for clarity. The terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. As used herein, unless clearly indicated otherwise, the singular forms “a,” “an,” and “the” are intended to include all such forms. It will be further understood that the terms “comprising” and / or “including”, when used in this specification, indicate the presence of the stated features, integers, steps, operations, components, and / or elements, but do not exclude the presence or addition of one or more other features, integers, steps, operations, components, and / or groups thereof. Unless otherwise indicated, when stated, a range of values ​​includes the upper and lower limits of the range and any subranges in between.

[0178] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A network partitioning processing method, characterized in that, Performed by network devices, including: Based on the traffic data of the target network, determine the transition matrix of the target network; Based on the eigenvalues ​​of the transition matrix, candidate logical subnets of the target network are determined; The target logical subnet is determined based on the modularity of each candidate logical subnet. The step of determining the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix includes: Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined. The candidate logical subnet is determined based on the sign of each element in the target feature vector; Wherein, the target feature value is the feature value other than the largest feature value among all the feature values ​​of the transition matrix; The step of determining the target logical subnet based on the modularity corresponding to each candidate logical subnet includes: The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

2. The method according to claim 1, characterized in that, Each of the candidate logical subnets includes: a first candidate logical subnet and a second candidate logical subnet; Determining the candidate logical subnet based on the sign of each element in the target feature vector includes: If the sign of the current element is positive, the network node corresponding to the current element is assigned to the first candidate logical subnet; If the sign of the current element is negative, the network node corresponding to the current element is assigned to the second candidate logical subnet.

3. The method according to claim 1, characterized in that, Before determining the target logical subnet based on the modularity corresponding to each candidate logical subnet, the process further includes: Based on the structure of the target network corresponding to the current candidate logical subnet, determine the first number of edges between nodes within different communities and the second number of edges between different communities; Based on the first number of edges and the second number of edges, the modularity of the current candidate logical subnet is calculated.

4. The method according to claim 1, characterized in that, Before determining the target logical subnet based on the modularity corresponding to each candidate logical subnet, the process further includes: Based on the structure of the target network corresponding to the current candidate logical subnet, determine the weights of all edges; The modularity of the current candidate logical subnet is calculated based on the weights.

5. The method according to claim 1, characterized in that, After determining the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix, the method further includes: If the modularity of each candidate logical subnet is less than or equal to the modularity obtained from the previous partitioning of the target network, the partitioning of the target network shall be stopped.

6. A network partitioning processing device, characterized in that, include: The first processing module is used to determine the transition matrix of the target network based on the traffic data of the target network; The second processing module is used to determine the candidate logical subnets of the target network based on the eigenvalues ​​of the transition matrix. The third processing module is used to determine the target logical subnet based on the modularity of each candidate logical subnet. The second processing module is further configured to: Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined. The candidate logical subnet is determined based on the sign of each element in the target feature vector; Wherein, the target feature value is the feature value other than the largest feature value among all the feature values ​​of the transition matrix; The third processing module is also used for: The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

7. A network device, characterized in that, Includes a processor, the processor being used for: Based on the traffic data of the target network, determine the transition matrix of the target network; Based on the eigenvalues ​​of the transition matrix, candidate logical subnets of the target network are determined; The target logical subnet is determined based on the modularity of each candidate logical subnet. The processor is further configured to: Based on the characteristic equation and each target eigenvalue of the transition matrix, a target feature vector corresponding to each target eigenvalue is determined. The candidate logical subnet is determined based on the sign of each element in the target feature vector; Wherein, the target feature value is the feature value other than the largest feature value among all the feature values ​​of the transition matrix; The processor is also used for: The candidate logical subnet corresponding to the maximum modularity is determined as the target logical subnet.

8. A network device, comprising: A transceiver, a processor, a memory, and a program or instructions stored in the memory and executable on the processor; characterized in that, when the processor executes the program or instructions, it implements the network partitioning processing method as described in any one of claims 1-5.

9. A readable storage medium having a program or instructions stored thereon, characterized in that, When the program or instructions are executed by the processor, they implement the steps in the network partitioning processing method as described in any one of claims 1-5.

10. A computer program product, characterized in that, It includes computer instructions that, when executed by a processor, implement the steps of the network partitioning processing method as described in any one of claims 1 to 5.

Images