Access point identification method and device, electronic equipment and storage medium

Abstract

The invention provides an access point identification method and device, electronic equipment and a storage medium, and relates to the technical field of internet surveying and mapping, and the method comprises the steps: carrying out the preliminary clustering of routing nodes based on geographic position information and autonomous system affiliation information, and obtaining at least one initial cluster; in each initial cluster, calculating a comprehensive score of each routing node based on the network topology out-degree and the average communication time delay, and performing secondary clustering on the routing nodes based on the comprehensive scores to obtain at least one sub-cluster; each sub-cluster is determined as an access point; and in each sub-cluster, selecting a routing node with the highest comprehensive score and the smallest network address value as a central identification node of the access point. According to the method and the device provided by the invention, the routing nodes are clustered for multiple times from coarse to fine, and the access points are accurately identified in a complex network structure; the method is beneficial for realizing the accurate attribution of the host nodes, and can more truly and accurately represent the physical and logic topological structures of the contemporary Internet.

Description

Technical Field

[0001] This application relates to the field of Internet mapping technology, and in particular to an access point identification method, device, electronic device, and storage medium. Background Technology

[0002] The physical foundation of the Internet consists of a vast number of Autonomous Systems (AS) and Points of Presence (POPs) distributed globally. An access point is a network service provider's network entry or aggregation node group deployed in a specific geographical location. It typically includes multiple interconnected routers, switches, and border devices used to carry regional traffic, provide network exit points, and facilitate interconnection. In Internet mapping, Content Delivery Network (CDN) site selection optimization, route stability analysis, and security protection systems, accurately identifying the scope, central nodes, and interconnections of access points is fundamental to understanding network structure and behavior patterns.

[0003] With the explosive growth of network scale, operators have deployed multiple access points within the same city or region to support different business traffic and service areas. A large autonomous system often contains dozens or even hundreds of physical access points, and their distribution no longer completely corresponds to geographical administrative regions or Autonomous System Number (ASN) boundaries. Therefore, network clustering methods in related technologies that rely on geographical location or ASN affiliation are no longer able to accurately represent the true topology of the contemporary Internet.

[0004] Therefore, how to accurately identify access points in complex network structures and precisely represent the true topology of the modern Internet has become a technical problem that the industry urgently needs to solve. Summary of the Invention

[0005] This application provides an access point identification method, apparatus, electronic device, and storage medium to solve the technical problem of how to accurately identify access points in complex network structures and precisely represent the real topology of the contemporary Internet.

[0006] This application provides an access point identification method, including: Obtain the geographical location information, autonomous system affiliation information, network topology out-degree and average communication latency of each routing node in the network; Based on the geographic location information and the autonomous system affiliation information, the routing nodes are initially clustered to obtain at least one initial cluster; In each initial cluster, a comprehensive score for each routing node is calculated based on the network topology out-degree and the average communication latency. The routing nodes are then further clustered based on the comprehensive score to obtain at least one sub-cluster. Each sub-cluster is identified as an access point. In each sub-cluster, the routing node with the highest overall score and the smallest network address value is selected as the central identifier node of the access point.

[0007] In some embodiments, the method further includes: Traverse the connection records centered on the host node to determine the connection relationship between the host node and at least one routing node in an associated access point; Based on the link quality parameters of each connection relationship, the connection weight of the host node to each associated access point is determined; the link quality parameters include the link communication latency of the connection relationship and the network topology in-degree of the connected routing node. Based on the connection weights of the host node to each associated access point, calculate the attribution confidence of the host node to each associated access point. The host node is assigned to the access point with the highest home confidence level that exceeds a preset threshold.

[0008] In some embodiments, the method further includes: If the attribution confidence is less than the preset threshold, obtain the two-hop neighbor node of the host node; The access point to which the host node belongs is determined based on the attribution confidence of the host node's two-hop neighbor nodes for each associated access point.

[0009] In some embodiments, the method further includes: If the access point to which the host node belongs cannot be determined based on the two-hop neighbor nodes of the host node, the host node shall be assigned to the access point with the highest confidence level among the candidate access points that have the same geographical location information and autonomous system affiliation information as the host node.

[0010] In some embodiments, the method further includes: The router-level connections in the network are mapped to access point-level connections to construct the access point-level network skeleton of the network; All router-level connections connecting the same two access points are merged, and the weight of the access point-level connections is calculated based on the number of merged connections and the average communication latency.

[0011] In some embodiments, the method further includes: In the access point-level network skeleton, if the autonomous system affiliation information of the two access points connected by any access point-level connection is different, then the access point-level connection is determined as a cross-domain boundary link.

[0012] In some embodiments, the secondary clustering of the routing nodes based on the comprehensive score to obtain at least one sub-cluster includes: If the routing nodes within the initial cluster exhibit multiple density centers, the initial cluster is divided into multiple sub-clusters; each sub-cluster corresponds to a density center. If the routing nodes within the initial cluster are sparsely distributed, the initial cluster is determined to be a single sub-cluster.

[0013] This application provides an access point identification device, including: The acquisition module is used to acquire the geographical location information, autonomous system affiliation information, network topology out-degree and average communication latency of each routing node in the network; The preliminary clustering module is used to perform preliminary clustering of the routing nodes based on the geographical location information and the autonomous system affiliation information to obtain at least one initial cluster; The secondary clustering module is used to calculate the comprehensive score of each routing node in each initial cluster based on the network topology out-degree and the average communication latency, and to perform secondary clustering of the routing nodes based on the comprehensive score to obtain at least one sub-cluster; each sub-cluster is determined as an access point. The identification module is used to select the routing node with the highest comprehensive score and the smallest network address value in each sub-cluster as the central identification node of the access point.

[0014] This application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the access point identification method.

[0015] This application provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the access point identification method.

[0016] The access point identification method, apparatus, electronic device, and storage medium provided in this application combine multi-dimensional features such as geographic location information, autonomous system affiliation information, network topology out-degree, and average communication latency to perform multiple clusterings of routing nodes from coarse to fine, accurately identifying access points in complex network structures. Furthermore, by determining a central identification node with network representativeness and identification stability through comprehensive scoring and network address sequence number, it overcomes the shortcomings of coarse-grained and inaccurate identification methods in related technologies, which is conducive to the accurate attribution of host nodes and can more realistically and accurately represent the physical and logical topology of the contemporary Internet. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the access point identification method provided in this application.

[0020] Figure 2 This is a flowchart illustrating the access point identification and host attribution method provided in this application.

[0021] Figure 3 This is a schematic diagram of the access point identification device provided in this application.

[0022] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first," "second," etc., used in this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that comprises a series of steps, units, or modules is not necessarily limited to those explicitly listed, but may include other steps, units, or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0025] In the field of internet mapping, related technologies employ a geolocation-based approach, relying on geolocation to identify point-of-access (POP) locations. This primarily involves using public geographic databases to map Internet Protocol (IP) addresses to latitude and longitude coordinates, and then grouping nodes into the same cluster based on spatial distance thresholds. While this method is simple to implement and computationally efficient, it assumes geographical proximity (i.e., logical proximity) and ignores the hierarchical nature of network connections and the internal multi-access point layout within network operators.

[0026] Furthermore, related technologies also employ an Autonomous System Number (ASN)-based approach, relying on Border Gateway Protocol (BGP) routing information or WHOIS registration data to classify network nodes within the same ASN into the same category. While this method reflects the administrative domain of an ASN, its granularity is too coarse: a large operator's ASN often covers multiple provinces or even countries, failing to reflect differences in access points at the city or even data center level.

[0027] These methods all struggle to accurately identify access points and precisely represent the true topology of the modern Internet.

[0028] In order to address the shortcomings of related technologies, Figure 1 This is a flowchart illustrating the access point identification method provided in this application, as follows: Figure 1 As shown, the method includes steps 110, 120, 130 and 140.

[0029] Step 110: Obtain the geographical location information, autonomous system affiliation information, network topology out-degree and average communication latency of each routing node in the network.

[0030] Specifically, the access point identification method provided in this application is executed by an access point identification device or system. This device can be implemented in software, such as an access point identification program; or it can be a device that executes the access point identification method, such as a terminal, computer, or server.

[0031] In this embodiment of the application, a routing node can refer to any network device in the Internet that executes routing protocols and is responsible for forwarding data packets, such as a router or a Layer 3 switch. Each routing node is typically identified by one or more IP addresses.

[0032] Geographic location information refers to the physical geographic location corresponding to the IP address of a routing node. This information can include country, province (or state), city, and more precise latitude and longitude coordinates. This information can be obtained by querying public or commercial IP geographic location databases, such as MaxMind GeoLite2 and IP2Location. By mapping IP addresses to geographic coordinates, physical space attributes can be established for each routing node.

[0033] Autonomous System (AS) affiliation information refers to the AS number to which a routing node belongs. An AS is a collection of IP networks and routers managed by a single technical authority. This information is typically obtained by resolving Border Gateway Protocol (BGP) routing information or querying the WHOIS registry database, and it reflects the administrative domain affiliation of the routing node.

[0034] Out-degree in network topology refers to the number of edges that directly connect a routing node to other nodes in a network topology graph. This metric is an important topological characteristic for measuring the breadth and influence of a node in the network. In a specific implementation, a routing-level topology graph can be constructed by actively probing or passively analyzing network traffic data, and the out-degree of each node can be counted from it.

[0035] Average communication latency refers to the average network communication delay from one or more fixed measurement points to a target routing node, usually measured by Round-Trip Time (RTT). This metric reflects the proximity or reachability of a node in the network; lower latency generally indicates a closer network distance. This latency can be obtained by performing multiple measurements on the target routing node using network measurement tools such as Ping and averaging the results to eliminate jitter and errors from single measurements.

[0036] By acquiring information from the above four dimensions, this embodiment of the application constructs a comprehensive feature vector for each routing node. This vector simultaneously includes the node's physical location, administrative affiliation, topological importance, and network proximity, laying the foundation for subsequent accurate identification.

[0037] Step 120: Based on geographic location information and autonomous system affiliation information, perform preliminary clustering of routing nodes to obtain at least one initial cluster.

[0038] Specifically, identical geographical location information indicates that routing nodes are located in the same city area. Same autonomous system affiliation information indicates that routing nodes belong to the same autonomous system's management domain.

[0039] Routing nodes with the same geographical location information and / or the same autonomous system affiliation information can be grouped into the same initial cluster, achieving preliminary clustering. Preliminary clustering ensures that the clustering results first satisfy the geographical nature of physical deployment and the consistency of network management.

[0040] Considering that a large operator (corresponding to an autonomous system) may deploy multiple physically independent access points within the same large city (corresponding to a geographical location), and the node distribution density varies greatly, relying solely on geographical location information and autonomous system affiliation information would result in an excessively large initial cluster, potentially containing multiple logical access points. Therefore, the initial clusters obtained through this step are still coarse-grained, potentially incorrectly grouping multiple logically independent access points into a single cluster. Thus, a further stage of refined clustering is required.

[0041] Step 130: In each initial cluster, calculate the comprehensive score of each routing node based on the network topology out-degree and average communication delay, and perform secondary clustering of the routing nodes based on the comprehensive score to obtain at least one sub-cluster; each sub-cluster is determined as an access point.

[0042] Specifically, the composite score is used to quantify the coreness or importance of each routing node within the initial cluster. This score combines the node's topological influence (network topology out-degree) and network centrality (average communication latency).

[0043] In one specific embodiment, the comprehensive score It can be calculated using the following formula: .

[0044] in, The out-degree of the network topology; This represents the average communication delay. and This is a weighting coefficient that can be preset based on experience or dynamically adjusted according to the network environment. It is used to balance the importance of two factors: topology out-degree and communication latency.

[0045] A higher out-degree of a routing node indicates more downstream nodes it connects to and a greater topological influence; a lower average communication latency indicates a more core network location and better reachability. Therefore, nodes with higher overall scores are more likely to be core devices in their local network.

[0046] Then, based on the calculated comprehensive score, the routing nodes within the initial cluster are further clustered. Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and Ordering Points To Identify the Clustering Structure (OPTICS) algorithms can be used here. These algorithms use the comprehensive score of a node as a key input dimension (or as a node weight), aggregating nodes with similar scores and close proximity in the network topology to form sub-clusters.

[0047] Each sub-cluster obtained through secondary clustering is identified as an access point in this embodiment. An access point physically corresponds to an access point or aggregation node group deployed by a network service provider in a specific geographical location, such as a collection of network devices within a data center. In this way, even within the same city and the same operator's initial cluster, if there are multiple device groups with different deployment densities and core nodes, the method provided in this embodiment can effectively distinguish them and identify them as multiple independent access points.

[0048] Step 140: In each sub-cluster, select the routing node with the highest comprehensive score and the smallest network address value as the central identifier node of the access point.

[0049] Specifically, after identifying each access point (i.e., sub-cluster), in order to facilitate subsequent management, reference and analysis, a unique and stable representative node, namely the central identifier node, can be assigned to each access point.

[0050] The selection of this node follows two principles: (1) The node with the highest comprehensive score is selected first within the sub-cluster. This ensures that the selected central identifier node is the most core and representative node in the network function of the access point.

[0051] (2) The node with the smallest network address value (IP int) is used as the deterministic arbitration standard when multiple nodes with the same highest comprehensive score exist within a subcluster. The node's IP address can be converted into an unsigned integer representing the network address sequence number, and then the node with the smallest integer value is selected as the central identifier node. This rule ensures that, regardless of when the method is executed, the exact same central identifier node will always be obtained for the same input, ensuring the stability and uniqueness of the identifier.

[0052] By applying the above two principles, the selected central identifier node achieves network representativeness and identifier stability. This central identifier node can be used as the unique identifier of its access point. This greatly simplifies the subsequent description and analysis of the entire network topology.

[0053] The access point identification method provided in this application combines multi-dimensional features such as geographic location information, autonomous system affiliation information, network topology out-degree, and average communication latency to perform multiple clusterings of routing nodes from coarse to fine, accurately identifying access points in complex network structures. Furthermore, it determines a central identifier node with network representativeness and identification stability by using comprehensive scoring and network address sequence number. This overcomes the shortcomings of coarse-grained and inaccurate identification methods in related technologies, which is conducive to the accurate attribution of host nodes and can more realistically and accurately represent the physical and logical topology of the contemporary Internet.

[0054] It should be noted that each implementation method of this application can be freely combined, rearranged, or executed individually, and does not need to rely on or depend on a fixed execution order.

[0055] In some embodiments, routing nodes are further clustered based on a comprehensive score to obtain at least one sub-cluster, including: If the distribution of routing nodes within the initial cluster exhibits multiple density centers, the initial cluster is divided into multiple sub-clusters; each sub-cluster corresponds to a density center. When the distribution of routing nodes within the initial cluster is sparse, the initial cluster is determined to be a single sub-cluster.

[0056] Specifically, the distribution of routing nodes within the initial cluster exhibits multiple density centers, corresponding to a scenario where a large operator deploys multiple physically independent, large data centers or tandem exchanges within a major city (i.e., the initial cluster). In the network topology, this manifests as routing nodes within the initial cluster forming multiple high-density regions, or density centers, in a feature space measured by a comprehensive score. Each density center typically surrounds one or more core routing nodes with extremely high comprehensive scores. In this case, secondary clustering can employ a density-based clustering algorithm to divide the initial cluster into multiple sub-clusters. These sub-clusters naturally correspond one-to-one with the density centers present in the data.

[0057] The initial cluster has sparsely distributed routing nodes, corresponding to scenarios where operators deploy on a small scale in a medium-sized city or region, or where the nodes are relatively evenly distributed without forming a clear multi-center structure. In this case, forcibly dividing the initial cluster into multiple sub-clusters would lead to over-segmentation, resulting in meaningless and overly scattered access points, which is also inconsistent with physical reality. In this situation, secondary clustering directly determines the entire initial cluster as a single sub-cluster. This is equivalent to confirming that under this combination of geographical location and operator, only one unified access point exists. This approach avoids erroneous segmentation of sparse areas, ensuring the rationality and robustness of the clustering results.

[0058] The access point identification method provided in this application can flexibly cope with various complex network deployment realities. Whether it is a multi-center layout in a large city or a single-center layout in a small or medium-sized area, it can obtain accurate and logical identification results.

[0059] In some embodiments, the method further includes: Traverse the connection records centered on the host node to determine the connection relationship between the host node and at least one routing node in an associated access point; Based on the link quality parameters of each connection relationship, the connection weight of the host node to each associated access point is determined; the link quality parameters include the link communication latency of the connection relationship and the network topology in-degree of the connected routing nodes. Based on the connection weights of the host node to each associated access point, calculate the attribution confidence of the host node to each associated access point. Assign the host node to the access point with the highest home confidence that exceeds the preset threshold.

[0060] Specifically, a host node refers to an endpoint device in a network whose main function is to act as a source or destination for communication, rather than to perform routing and forwarding.

[0061] Connection records are data describing the connection relationships between host nodes and routing nodes. Connection records can be obtained through network measurement data. The system iterates through all connection records related to the IP address of a host node to be assigned to. All routing nodes in the network have been assigned to their respective access points; therefore, these connection records can determine which routing nodes within which access points a host node is connected to. A host node may be connected to only one routing node within an access point, or it may be connected to multiple routing nodes within the same or different access points simultaneously.

[0062] After determining the connection relationship between the host node and the routing nodes within each candidate access point, the embodiments of this application aim to quantify the link quality parameters of each connection and calculate the voting weight based on the link quality parameters.

[0063] Link communication latency refers to the communication delay (e.g., RTT) between host nodes and routing nodes. The lower the latency, the closer the network distance and the higher the connection quality.

[0064] Network topology in-degree refers to the in-degree of the routing nodes to which a host node is connected. A higher in-degree routing node typically indicates that it aggregates more upstream links and is a more important aggregation point in the network. A host's affiliation to a core routing node with a high in-degree is more representative than its affiliation to an edge routing node with a low in-degree.

[0065] In a specific embodiment, a connection weight can be calculated for each connection between a host node and a routing node. Its calculation formula can be: .

[0066] in, For link communication delay; In-degree of network topology; and These are preset weighting coefficients used to adjust the relative importance of link communication latency and network topology in-degree in the final weight calculation. The formula reflects this: lower latency, higher weight; larger in-degree, higher weight.

[0067] A host node may be connected to multiple routing nodes, which may belong to different access points. This application embodiment uses a weighted voting mechanism to comprehensively evaluate the probability that the host node belongs to each candidate access point.

[0068] The system first aggregates the connection weights from the host node to the same access point. For each candidate access point, its total weighted votes can be calculated as follows: .

[0069] in, Access point The total weighted votes. The summation operation here is performed on the host node to the votes belonging to... All connections to all routing nodes of this access point.

[0070] After obtaining the total weighted votes for each candidate access point, this embodiment of the application then calculates the attribution confidence score to obtain a standardized and easily comparable metric. The attribution confidence score is a value between 0 and 1, representing the degree of confidence the system has inferred that the host node belongs to a specific access point. Its calculation formula can be: .

[0071] in, Indicates access point The confidence level of attribution, It is the sum of the total weighted votes for all candidate access points.

[0072] After calculating the affiliation confidence of the host node with respect to all candidate access points, this embodiment of the application performs the final affiliation determination. The determination rule includes two conditions: (1) The access point with the highest home confidence is selected as the preferred home target.

[0073] (2) Exceeding a preset threshold: The highest attribution confidence value must be greater than a preset threshold. This threshold is used to avoid making incorrect judgments when there is insufficient evidence. For example, the preset threshold can be set to 0.5. This means that if the attribution confidence is greater than 0.5, the current attribution judgment is reliable; if the attribution confidence is less than or equal to 0.5, the current attribution judgment is unreliable.

[0074] If an access point meets both of the above conditions, the system will officially assign the host node to that access point. If the highest confidence level of attribution fails to exceed a preset threshold (for example, the highest confidence level of attribution is only 0.4), it indicates that the current attribution determination has high ambiguity, and the system can choose not to determine the attribution of the host node temporarily, but wait for the next attribution determination.

[0075] The access point identification method provided in this application introduces a confidence threshold mechanism to perform weighted voting evaluation on the connection relationship between host nodes and various access points, and calculates the host affiliation confidence by combining parameters such as link quality and network topology in-degree, which significantly improves the reliability and robustness of affiliation judgment.

[0076] In some embodiments, the method further includes: If the attribution confidence is less than a preset threshold, obtain the host node's two-hop neighbor nodes; The access point to which the host node belongs is determined by the confidence level of the two-hop neighbor nodes of the host node for each associated access point.

[0077] Specifically, if the attribution confidence is greater than a preset threshold, the attribution result can be obtained directly. If the attribution confidence is less than the preset threshold, the system can enter a second round of inference, using the attribution information of the host node's two-hop neighbor nodes to assist in the judgment, thereby improving the accuracy and coverage of the attribution.

[0078] If, after calculation, the highest attribution confidence of a host node is still less than a preset threshold (e.g., 0.5), the system triggers a second round of inference.

[0079] The system needs to identify the target host node's two-hop neighbors. In network topology, a two-hop neighbor is a node connected to the target host node through an intermediate node. Methods for obtaining two-hop neighbors may include: analyzing network traffic data to find other hosts that have accessed the same intermediate router as the target host node; or querying network configuration information to determine other devices in the same subnet as the target host node.

[0080] After obtaining the set of two-hop neighbor nodes of the target host node, this embodiment analyzes the attribution of these two-hop neighbor nodes to infer the attribution of the target host node. The basic assumption is that the attribution of a node is likely to be consistent with the attribution trend of its neighbor nodes.

[0081] The specific determination process is as follows: (1) To query the neighbor affiliation, the system will traverse all found two-hop neighbor nodes. For each two-hop neighbor node, query its own access point affiliation.

[0082] (2) Summarize the opinions of neighbors. The system only focuses on those two-hop neighbor nodes that already have a clear ownership. The ownership results of these neighbor nodes are regarded as a vote on the ownership of the target host node.

[0083] (3) Determine the final affiliation. Based on the voting results of the neighbors, the system will assign the target host node to the access point that received the most votes.

[0084] In one specific embodiment, the voting described above can be weighted. The weights can be determined based on the confidence levels of the two-hop neighbors and their respective access points. For example, if one two-hop neighbor node belonging to access point A has a confidence level of 0.9, and another two-hop neighbor node belonging to access point A has a confidence level of 0.6, the former's voting weight can be set higher than the latter's. This makes the attribution determination more reliable.

[0085] The access point identification method provided in this application further enhances the stability of the network attribution by introducing the voting influence of two-hop neighbor nodes, making the final network topology profile more complete and accurate.

[0086] In some embodiments, the method further includes: If the two-hop neighbor nodes of the host node cannot determine the access point to which the host node belongs, the host node will be assigned to the access point with the highest confidence level among the candidate access points that have the same geographical location information and autonomous system affiliation information as the host node.

[0087] Specifically, if the access point to which a host node belongs cannot be determined based on its two-hop neighbor nodes, this application provides a binding mandatory attribution strategy to ensure that every host node in the network can obtain an attribution, thereby guaranteeing the integrity of the final network topology.

[0088] The system first determines a range of candidate access points. This range is defined by two key attributes: having the same geographical location information and autonomous system affiliation information as the target host node. This range is based on common sense: a user's network access is typically provided by a local operator in their geographic area. Therefore, forcibly assigning them to an access point outside this range is illogical.

[0089] Within the aforementioned candidate range, the system will re-examine the target host node's confidence level in its affiliation with these candidate access points. Although these confidence levels may not have exceeded the preset threshold at that time, their relative magnitudes still reflect the relative strength of the connection.

[0090] The system directly selects the access point with the highest attribution confidence within the candidate range and forces the target host node to belong to it.

[0091] Based on the above embodiments, this application provides a multi-level, progressive attribution determination method: When the attribution confidence is higher than the preset threshold, a direct judgment is made based on the attribution confidence of the direct connection between the host node and the routing node; when the attribution confidence is low but there is a certain bias, a second round of judgment is made based on the attribution confidence of each associated access point by the two-hop neighbor node; when the host is still in an ambiguous state, forced attribution is made based on geographical location information and autonomous system attribution information.

[0092] The access point identification method provided in this application performs forced attribution based on geographical location information and autonomous system attribution information, ensuring the integrity and geographical consistency of the attribution results.

[0093] In some embodiments, the method further includes: Map router-level connections in the network to access point-level connections to build the network's access point-level network skeleton; All router-level connections connecting the same two access points are merged, and the weight of the access point-level connections is calculated based on the number of merged connections and the average communication latency.

[0094] Specifically, a list or set of edges at the router level is a dataset that describes the direct connections between routers in a network.

[0095] The system iterates through each connection in the router-level connection list. For each connection, the system queries the access point identifiers (APIs) of both ends, distinguishing between input APIs (Access Point Identifiers). ) and output access point identifier ( Through the query process, each micro-level connection between routers is mapped to a macro-level connection between access points. After this mapping process, the original complex network graph composed of thousands of router nodes and connections is transformed into a new graph with access points as nodes and connections occurring between access points. This achieves the mapping of router-level connections to access point-level connections, thus obtaining the access point-level network skeleton.

[0096] Connecting the same two access points ( and In other words, among all router-level connections (that is, connections between two access points), there may be parallel and redundant connections. These router-level connections are merged, and an index that can comprehensively reflect the connection strength between the two access points is calculated, namely the access point-level connection weight.

[0097] The specific merging and calculation process is as follows: (1) Merging and Statistics: The system groups all router-level connections that connect the same two access points into one group. Then, it counts the total number of connections in this group (i.e., the number of connections) and calculates the average communication latency of these connections. The number of connections reflects the redundancy and bandwidth capacity potential of the link between the two access points; a higher number usually indicates a more important connection. The average communication latency reflects the network distance and connection quality between them.

[0098] (2) Calculate the connection weight. Based on the above statistical results, calculate a comprehensive weight value. In a specific embodiment, the weight of the access point-level connection is... It can be calculated using the following formula: .

[0099] in, Indicates the number of connections. Indicates average communication latency. and It is a preset weighting coefficient used to balance the importance of the number of connections and the average communication latency.

[0100] The formula shows that the more physical links connecting two access points, or the lower the average communication latency between them, the higher the weight of the connection at the access point level, indicating that the connection between the two access points is closer and more important.

[0101] In a point-cluster-level network skeleton, nodes are access points, edges are connections at the access point level, and the weight of an edge is the weight of the connection at the access point level. The higher the weight, the more important the link between the two access points connected by that edge is in terms of traffic capacity and interconnection quality.

[0102] The access point identification method provided in this application maps router-level connections in the network to access point-level connections, constructing an access point-level network skeleton. This provides a solid data foundation and intuitive analysis tools for subsequent network performance analysis, fault diagnosis, capacity planning, and security policy formulation, greatly enhancing the ability to understand and manage large-scale complex networks.

[0103] In some embodiments, the method further includes: In the access point-level network skeleton, if the autonomous system affiliation information of the two access points connected by any access point-level connection is different, then any access point-level connection is identified as a cross-domain boundary link.

[0104] Specifically, for any access point-level connection in the access point-level network skeleton, if the autonomous system affiliation information of the two connected access points is different, it indicates that the access point-level connection is a cross-domain boundary link, and the access point in the connection is a potential boundary router.

[0105] Meanwhile, by analyzing the changing trends of the weights (edge ​​weights) of access point-level connections over time, the system has the ability to detect cross-domain boundary enhancements or bottlenecks in advance, which is beneficial for operation and maintenance deployment, capacity planning, and security detection strategy formulation. Ultimately, the system can abstract the access point-level network skeleton at multiple levels, allowing the network layout to be aggregated and displayed at the geographical region or national level to adapt to the network insight needs of different application scenarios.

[0106] The access point identification method provided in this application embodiment, based on the construction of an access point-level network skeleton, further provides the ability to identify cross-domain boundary links, which is beneficial for identifying key interconnection links and potential bottleneck areas between operators.

[0107] Figure 2 This is a flowchart illustrating the access point identification and host attribution method provided in this application, as follows: Figure 2 As shown, the method includes: Step 210: Access point identification.

[0108] This step is the first stage of the overall process. Its main goal is to perform intelligent clustering on the underlying routing nodes to obtain the basic framework of the network topology, which combines physical and logical elements.

[0109] This step includes dividing host nodes and routing nodes, obtaining geographical location information and autonomous system affiliation information, initial clustering, secondary clustering within the initial cluster, and determining the central identifier node.

[0110] The output of the first phase forms an access point topology covering the entire network, providing a reliable framework for subsequent host attribution inference.

[0111] Step 220: Host node ownership.

[0112] This step is the second stage of the overall process. The system will expand its focus to host nodes on the terminal side or server side and infer the most suitable access point affiliation based on their connection relationship with the routing node.

[0113] This step includes calculating connection weights, calculating home confidence, and determining host affiliation.

[0114] The second phase not only improved the accuracy of host attribution, but also enhanced system reliability and the ability to avoid misclassification through a confidence mechanism.

[0115] Step 230: Construction of access point-level network skeleton and identification of cross-domain boundary links.

[0116] This step is the third stage of the overall process, which abstracts router-level links to access point-level networks to form a network backbone topology, used to identify key interconnection areas and cross-domain boundaries.

[0117] This step includes building an access point-level network skeleton and identifying cross-domain boundary links.

[0118] The third stage can identify cross-domain boundaries or bottlenecks in advance, enabling network layouts to be aggregated and displayed at the geographical region or national level, in order to meet the needs of different application scenarios for in-depth network insights.

[0119] The apparatus provided in the embodiments of this application is described below. The apparatus described below can be referred to in correspondence with the method described above.

[0120] Figure 3 This is a schematic diagram of the access point identification device provided in this application, as shown below. Figure 3 As shown, the device includes: The acquisition module 310 is used to acquire the geographical location information, autonomous system affiliation information, network topology out-degree and average communication latency of each routing node in the network; The preliminary clustering module 320 is used to perform preliminary clustering of routing nodes based on geographic location information and autonomous system affiliation information to obtain at least one initial cluster. The secondary clustering module 330 is used to calculate the comprehensive score of each routing node in each initial cluster based on the network topology out-degree and average communication latency, and to perform secondary clustering of the routing nodes based on the comprehensive score to obtain at least one sub-cluster; each sub-cluster is determined as an access point. The identification module 340 is used to select the routing node with the highest comprehensive score and the smallest network address value in each sub-cluster as the central identification node of the access point.

[0121] The access point identification device provided in this application combines multi-dimensional features such as geographic location information, autonomous system affiliation information, network topology out-degree, and average communication latency to perform multiple clusterings of routing nodes from coarse to fine, accurately identifying access points in complex network structures. Furthermore, it determines a central identification node with network representativeness and identification stability by using comprehensive scoring and network address sequence number. This overcomes the shortcomings of coarse-grained and inaccurate identification methods in related technologies, which is beneficial for achieving accurate attribution of host nodes and can more realistically and accurately represent the physical and logical topology of the contemporary Internet.

[0122] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this application, such as... Figure 4 As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430, and a communications bus 440, wherein the processor, communications interface, and memory communicate with each other via the communications bus. The processor can invoke logical commands stored in the memory to execute the methods described in the above embodiments, for example: Obtain the geographical location information, autonomous system affiliation information, network topology out-degree, and average communication latency of each routing node in the network; based on the geographical location information and autonomous system affiliation information, perform preliminary clustering of the routing nodes to obtain at least one initial cluster; within each initial cluster, calculate the comprehensive score of each routing node based on the network topology out-degree and average communication latency, and perform secondary clustering of the routing nodes based on the comprehensive score to obtain at least one sub-cluster; each sub-cluster is determined as an access point; within each sub-cluster, select the routing node with the highest comprehensive score and the smallest network address value as the central identifier node of the access point.

[0123] Furthermore, the logical commands in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several commands to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0124] The processor in the electronic device provided in this application embodiment can call logical instructions in the memory to implement the above method. Its specific implementation method is the same as the aforementioned method implementation method and can achieve the same beneficial effect, which will not be repeated here.

[0125] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the methods provided in the above embodiments.

[0126] The specific implementation method is the same as the aforementioned method implementation method and can achieve the same beneficial effects, so it will not be repeated here.

[0127] This application provides a computer program product, including a computer program that, when executed by a processor, implements the method described above.

[0128] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0129] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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. Such 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 this application.

Claims

1. An access point identification method, characterized in that, include: Obtain the geographical location information, autonomous system affiliation information, network topology out-degree and average communication latency of each routing node in the network; Based on the geographic location information and the autonomous system affiliation information, the routing nodes are initially clustered to obtain at least one initial cluster; In each initial cluster, a comprehensive score for each routing node is calculated based on the network topology out-degree and the average communication latency. The routing nodes are then further clustered based on the comprehensive score to obtain at least one sub-cluster. Each sub-cluster is identified as an access point. In each sub-cluster, the routing node with the highest overall score and the smallest network address value is selected as the central identifier node of the access point.

2. The access point identification method according to claim 1, characterized in that, The method further includes: Traverse the connection records centered on the host node to determine the connection relationship between the host node and at least one routing node in an associated access point; Based on the link quality parameters of each connection relationship, the connection weight of the host node to each associated access point is determined; the link quality parameters include the link communication latency of the connection relationship and the network topology in-degree of the connected routing node. Based on the connection weights of the host node to each associated access point, calculate the attribution confidence of the host node to each associated access point. The host node is assigned to the access point with the highest home confidence level that exceeds a preset threshold.

3. The access point identification method according to claim 2, characterized in that, The method further includes: If the attribution confidence is less than the preset threshold, obtain the two-hop neighbor node of the host node; The access point to which the host node belongs is determined based on the attribution confidence of the host node's two-hop neighbor nodes for each associated access point.

4. The access point identification method according to claim 3, characterized in that, The method further includes: If the access point to which the host node belongs cannot be determined based on the two-hop neighbor nodes of the host node, the host node shall be assigned to the access point with the highest confidence level among the candidate access points that have the same geographical location information and autonomous system affiliation information as the host node.

5. The access point identification method according to claim 1, characterized in that, The method further includes: The router-level connections in the network are mapped to access point-level connections to construct the access point-level network skeleton of the network; All router-level connections connecting the same two access points are merged, and the weight of the access point-level connections is calculated based on the number of merged connections and the average communication latency.

6. The access point identification method according to claim 5, characterized in that, The method further includes: In the access point-level network skeleton, if the autonomous system affiliation information of the two access points connected by any access point-level connection is different, then the access point-level connection is determined as a cross-domain boundary link.

7. The access point identification method according to claim 1, characterized in that, The secondary clustering of the routing nodes based on the comprehensive score to obtain at least one sub-cluster includes: If the routing nodes within the initial cluster exhibit multiple density centers, the initial cluster is divided into multiple sub-clusters; each sub-cluster corresponds to a density center. If the routing nodes within the initial cluster are sparsely distributed, the initial cluster is determined to be a single sub-cluster.

8. An access point identification device, characterized in that, include: The acquisition module is used to acquire the geographical location information, autonomous system affiliation information, network topology out-degree and average communication latency of each routing node in the network; The preliminary clustering module is used to perform preliminary clustering of the routing nodes based on the geographical location information and the autonomous system affiliation information to obtain at least one initial cluster; The secondary clustering module is used to calculate the comprehensive score of each routing node in each initial cluster based on the network topology out-degree and the average communication latency, and to perform secondary clustering of the routing nodes based on the comprehensive score to obtain at least one sub-cluster; each sub-cluster is determined as an access point. The identification module is used to select the routing node with the highest comprehensive score and the smallest network address value in each sub-cluster as the central identification node of the access point.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the access point identification method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the access point identification method according to any one of claims 1 to 7.

Images