A network management client topology management system
Patent Information
- Application Number
- CN202611152756.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-31
- Publication Date
- 2026-10-16
Smart Images
Figure CN122933434A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of network management technology, and in particular to a network management client topology management system. Background Technology
[0002] Existing network management client topology management systems are widely used in the monitoring and maintenance of data exchange networks. Their core task is to construct a visualized network topology by collecting neighbor identifiers and port statuses of network devices. In actual physical deployment environments, the signal quality of network links, especially long-distance transmission links, is extremely sensitive to changes in ambient temperature. When there are significant gradient changes in ambient temperature, the boundary state of the link's physical layer can easily become unstable, leading to temporary inconsistencies in the perception of adjacency relationships between devices at both ends of the link, resulting in half-adjacency or ghost link phenomena. These phenomena often occur sequentially along shared cabling paths such as shafts or cable trays in physical space, causing abnormal states in the network topology to move along physical paths over time, posing a technical challenge to the accurate attribution of network faults.
[0003] Existing technologies for topology management and fault identification primarily employ a discrete refresh mechanism based on real-time states. This means that the topology is updated whenever port connectivity or neighbor changes are detected, ignoring the crucial role of ambient temperature in driving link states. This approach severs the intrinsic connection between link logical states and physical environment changes, failing to identify non-persistent semi-adjacent states caused by temperature gradients. Consequently, a large number of unrelated link oscillation alarms are generated during environmental disturbances. Furthermore, there is a lack of exploration into the fault propagation patterns along shared physical paths, making it impossible to verify whether wavefront propagation behavior conforming to the relationship between physical distance and time lag exists among multiple link anomalies. This results in the network management system being unable to automatically converge the impact range when facing such physical layer interference, making it difficult to present the minimum impact area of the problem in an intuitive and convergent manner on the topology map. Summary of the Invention
[0004] The purpose of this invention is to address the problems in existing technologies, such as ignoring the impact of environmental temperature changes on link status, lacking verification of the propagation law of anomalies on shared physical paths, and being unable to automatically converge the minimum influence domain of topology anomalies. Therefore, this invention proposes a network management client topology management system.
[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution: A network management client topology management system, comprising: The alignment acquisition module is used to perform alignment sampling on the devices at both ends of the network link according to a preset sampling period, and to obtain the neighbor identifiers, port status and temperature at both ends of the link to form an observation sequence. The phase-locked gate module is used to determine the phase-locked time window that satisfies the temperature gradient phase-locked condition based on the rate of change and direction of change of temperature in the observation sequence. The event extraction module is used to identify the state where the neighbor identifiers at both ends of the network link are inconsistent within the phase-locked loop time window, and extract half-adjacent events that meet the preset duration to form a half-adjacent event set. The wavefront fitting module is used to obtain the set of links on the preset shared physical path and the discrete path distance corresponding to each link in the set. Based on the mapping relationship between the occurrence time of the semi-adjacency event of each link in the set and the discrete path distance, the wavefront propagation fitting verification is performed. The confidence rendering module is used to calculate the edge confidence of the link in the link set based on the semi-adjacent event set if the result of the wavefront propagation fitting verification meets the preset conditions, and to determine and render the minimum influence domain based on the edge confidence.
[0006] Preferably, the phase-locked time window that satisfies the temperature gradient phase-locked condition is determined based on the rate of change and direction of change of temperature in the observation sequence, including: Calculate the sum of the temperature changes at both ends of the link between the current sampling time and the previous sampling time, and divide the sum by the sampling period to obtain the temperature gradient; Determine whether the temperature gradient is greater than a preset gradient threshold; Determine whether the direction of temperature change at both ends of the link is consistent; If the temperature gradient is greater than the preset gradient threshold and the temperature changes at both ends of the link are in the same direction, then the current sampling time is determined to be within the phase-locked time window.
[0007] Preferably, a semi-adjacent event set is formed, including: An adjacency inconsistency indication is determined when the neighbor identifiers at both ends of the link are not equal; The duration during which the adjacency inconsistency indicator remains true is recorded. Events that occur within the phase-locked loop time window and whose duration is between the preset minimum and maximum half-adjacency durations are selected to form a half-adjacency event set.
[0008] Preferably, wavefront propagation fitting verification is performed, including: Extract the first half-adjacency moment from the half-adjacency event set of each link in the link set; A linear regression model was established with discrete path distance as the independent variable and the first semi-adjacent time as the dependent variable. The wavefront propagation speed and wavefront initiation time were obtained by fitting the model. Calculate the mean square of the fitting residuals of the linear regression model.
[0009] Preferably, the conditions for determining that the wavefront propagation fitting verification is valid must be met simultaneously: the wavefront propagation speed is greater than zero, the number of links participating in the fitting is greater than the preset minimum number of effective edges, and the mean square of the fitting residual is less than the preset residual threshold.
[0010] Preferably, calculating the edge confidence of a link includes: Edge confidence is calculated based on the exponential decay function only when both physical ports at both ends of the link are in the enabled state. The number of events in the semi-adjacent event set is used as a variable for the decay term, and a preset decay factor is used as the coefficient of the decay term; the more events there are, the lower the edge confidence.
[0011] Preferably, determining and rendering the minimum influence domain includes: Links in the link set whose edge confidence is lower than a preset low confidence edge threshold are defined as the minimum influence domain; In the network management client topology diagram, only links belonging to the minimum influence domain are highlighted, and temperature gradient-related semi-adjacent wavefront anomalies are marked. In the display area associated with the minimum influence domain, the wavefront initiation time and wavefront propagation speed, obtained by wavefront propagation fitting verification, are output.
[0012] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention establishes an intrinsic correlation mechanism between the link logic state and physical environment temperature changes through the collaborative work of the alignment acquisition module and the phase-locked loop (PLL) gating module. The system accurately locks the PLL time window by utilizing the rate and direction of temperature change in the observation sequence, and then identifies and extracts semi-adjacent events caused by the environmental temperature gradient within this time window. This physical-driven event extraction method effectively overcomes the limitations of discrete refresh based solely on port on / off or neighbor changes while ignoring environmental factors. It can accurately identify non-persistent abnormal states during environmental disturbances and avoids a large number of unrelated link oscillation alarms caused by severing the physical environment connection.
[0013] 2. By establishing a linear regression model between the occurrence time of semi-adjacent events and the distance of discrete paths, this invention can verify whether the fault exhibits wavefront propagation behavior that conforms to the relationship between physical distance and time lag. Based on this, the edge confidence is calculated to automatically define the minimum influence domain. This not only solves the problem of difficulty in automatically converging the fault influence range under physical layer interference, but also highlights the core abnormal area in the topology map and outputs the wavefront start time and propagation speed, providing maintenance personnel with an intuitive and convergent view of fault root cause analysis. Attached Figure Description
[0014] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a functional module diagram of a network management client topology management system provided in an embodiment of the present invention. Detailed Implementation
[0015] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.
[0016] Example: This example provides a network management client topology management system. See [link to example]. Figure 1 Specifically, including: The alignment acquisition module is used to perform alignment sampling on the devices at both ends of the network link according to a preset sampling period, and to obtain the neighbor identifiers, port status and temperature at both ends of the link to form an observation sequence. In an embodiment of the present invention, according to a preset sampling period, aligned sampling is performed on the devices at both ends of the network link to obtain the neighbor identifiers, port statuses, and temperature values at both ends of the link, forming an observation sequence, including: According to the preset sampling period, the devices at both ends of the network link are aligned and sampled; Obtain the neighbor identifiers, port status, and temperature values at both ends of the link to form an observation sequence; Specifically, the neighbor identifier refers to the connection identifier between a device and other devices in the network, reflecting the physical or logical connection relationship between devices; the port status refers to the working status of a device port, usually enabled or disabled, indicating whether the port is active; the temperature value refers to the temperature value of the device port or the device itself, reflecting the temperature conditions of the device's working environment.
[0017] In detail, according to the preset sampling period, the system performs alignment sampling on the devices at both ends of the network link through an automated process; the preset value of the sampling period is 10 seconds, which is determined based on the performance of the devices and the dynamic change frequency of the network topology to ensure that the collected data has sufficient timeliness and accuracy; the system connects the devices at both ends of the link through the network management system (NMS) and obtains the neighbor identifier, port status and temperature of the devices through the network interface of the link.
[0018] In practice, the system first automatically obtains the neighbor identifiers of devices at both ends of the link using LLDP (Link Layer Discovery Protocol) or a similar protocol. These identifiers indicate the connection relationship between devices, including the identifiers of each device and other directly connected devices. The system then checks the port status at both ends of the link, indicating whether the port is enabled or disabled. Enabled indicates the port is active and can perform normal data communication, while disabled indicates the port is inactive and cannot exchange data. Real-time data from temperature sensors in the devices is obtained. The temperature reflects the ambient temperature of the device or port, typically expressed in degrees Celsius, with a range determined by the device's temperature measurement range, generally between -40°C and 85°C. All this data is stored in a preset format, such as a multidimensional vector format, and used to generate an observation sequence for subsequent data processing and topology analysis. This observation sequence provides detailed and real-time data support for network link status monitoring, fault diagnosis, and performance analysis.
[0019] The phase-locked gate module is used to determine the phase-locked time window that satisfies the temperature gradient phase-locked condition based on the rate of change and direction of change of temperature in the observation sequence. In an embodiment of the present invention, determining a phase-locked time window that satisfies the temperature gradient phase-locked condition based on the rate of change and direction of change of temperature in the observation sequence includes: Calculate the sum of the temperature changes at both ends of the link between the current sampling time and the previous sampling time, and divide the sum by the sampling period to obtain the temperature gradient; Determine whether the temperature gradient is greater than a preset gradient threshold; Determine whether the direction of temperature change at both ends of the link is consistent; If the temperature gradient is greater than the preset gradient threshold and the temperature changes at both ends of the link are in the same direction, then the current sampling time is determined to be within the phase-locked time window. Specifically, the phase-locked loop (PLL) time window refers to the period during which the temperature changes of the devices at both ends of the link are significant and synchronized. Within this time window, the temperature changes of the devices at both ends of the link follow the same direction, and the magnitude of the changes is greater than a preset temperature gradient threshold. The temperature gradient is obtained by calculating the ratio of the temperature change magnitude at both ends of the link to the sampling period, and is used to reflect the rate of temperature change of the devices. If the temperature changes at both ends of the link follow the same direction, it indicates that the devices have experienced similar thermal changes under the same environmental conditions. Determining the PLL time window helps to identify periods when the devices experience large temperature fluctuations that may affect the stability of the link.
[0020] Specifically, in this embodiment, the system calculates the sum of the temperature changes at both ends of the link between the current sampling time and the previous sampling time, and divides this sum by a preset sampling period to obtain the temperature gradient. The sampling period is 10 seconds, and this period length is set according to the device performance and data change rate to ensure the timeliness and stability of data acquisition. The temperature gradient reflects the rate of temperature change of the devices at both ends of the link per unit time and is used to assess the degree of temperature fluctuation of the devices. At this time, the sum of the temperature changes at both ends of the link refers to the sum of the absolute values of the temperature changes of the devices at both ends within the sampling period.
[0021] The calculated temperature gradient is compared with a preset temperature gradient threshold. This threshold is set based on the thermal stability and operating characteristics of the equipment, typically 1°C / second. When the temperature gradient exceeds this threshold, it indicates that the temperature change rate of the equipment at both ends of the link is relatively fast, which may affect the stability of the link. In this case, the system further determines whether the temperature change direction at both ends of the link is consistent, i.e., whether the temperatures at both ends rise or fall simultaneously. If the temperature change direction at both ends of the link is consistent, it indicates that the equipment is in a similar thermal environment and the temperature changes are synchronized.
[0022] When the temperature gradient exceeds a preset threshold and the temperature changes at both ends of the link are in the same direction, the current sampling time is considered to have entered the phase-locked loop (PLL) window. The PLL window is defined as a period of rapid and synchronous temperature changes in the equipment; during this period, temperature variations can significantly impact the performance and stability of the link. Through this process, the system can identify critical periods where link problems may exist and provide crucial information for subsequent monitoring and diagnosis.
[0023] The event extraction module is used to identify the state where the neighbor identifiers at both ends of the network link are inconsistent within the phase-locked loop time window, and extract half-adjacent events that meet the preset duration to form a half-adjacent event set. In an embodiment of the present invention, within the phase-locked loop (PLL) time window, the state where the neighbor identifiers at both ends of the network link are inconsistent is identified, and half-adjacency events that satisfy a preset duration are extracted to form a set of half-adjacency events, including: An adjacency inconsistency indication is determined when the neighbor identifiers at both ends of the link are not equal; The duration during which the adjacency inconsistency indicator remains true is recorded. Events that occur within the phase-locked loop time window and whose duration is between the preset minimum and maximum half-adjacency durations are selected to form a half-adjacency event set. It's important to note that a half-adjacency event set refers to a set of events in a network link where the neighbor identifiers of devices at both ends of the link are inconsistent, and this inconsistency lasts for a certain period. Specifically, adjacency inconsistency indicates that the neighbor identifiers reported by devices at both ends of the link via a neighbor discovery protocol (such as LLDP) are different, meaning that the devices at both ends of the link have failed to correctly identify each other, usually indicating a potential problem with the physical or logical connection of the link. A half-adjacency event refers to a situation where, at a specific moment, the neighbor identifiers of devices at both ends of the link change and remain inconsistent for a short period. This set is formed by collecting and statistically analyzing the moments of these half-adjacency events. Each element in the event set represents a link experiencing an inconsistent adjacency state at a specific moment, and the duration of this inconsistent state is between a preset minimum and maximum half-adjacency duration. By analyzing the half-adjacency event set, it is possible to identify link state anomalies caused by problems such as configuration errors, unstable connections, or physical layer failures at both ends of the link, providing a basis for subsequent fault diagnosis and performance analysis.
[0024] Specifically, the adjacency inconsistency indication is determined based on the neighbor identifiers obtained by the devices at both ends of the link through a neighbor discovery protocol (such as LLDP). The neighbor identifiers of the devices at both ends of the link are compared after exchanging data packets via the protocol. When the neighbor identifiers at both ends of the link are not equal, the system generates an adjacency inconsistency indication and sets it to true, indicating that the neighbor relationship between the devices at both ends of the link is inconsistent. Adjacency inconsistency usually indicates a problem with the physical or logical connection of the link, which may lead to data transmission interruption or performance degradation.
[0025] The system records the duration of the adjacency inconsistency indication remaining true for the specified period. This duration refers to the time interval from when the adjacency inconsistency indication becomes true until it returns to false. The system continuously monitors the adjacency relationship of devices at both ends of the link and updates the duration according to a preset sampling period. If the duration of the adjacency inconsistency exceeds the preset minimum half-adjacency duration (e.g., 30 seconds) but is less than the maximum half-adjacency duration (e.g., 5 minutes), the event is considered a valid half-adjacency event.
[0026] During the half-adjacency event identification process, the system further filters events whose occurrence time falls within the phase-locked loop (PLL) time window. The PLL time window is a specific time period set based on the temperature synchronicity and temperature gradient of the devices at both ends of the link. When the occurrence time of an inconsistency event meets the conditions within the PLL time window, and its duration meets the preset minimum and maximum half-adjacency duration requirements, the system adds the event to the half-adjacency event set. This set contains all valid half-adjacency events that meet the above conditions, for subsequent analysis and diagnosis. Through this process, the system can identify potential link faults or performance anomalies, providing support for subsequent topology optimization and fault location.
[0027] The wavefront fitting module is used to obtain the set of links on the preset shared physical path and the discrete path distance corresponding to each link in the set. Based on the mapping relationship between the occurrence time of the semi-adjacency event of each link in the set and the discrete path distance, the wavefront propagation fitting verification is performed. In embodiments of the present invention, a preset set of links on a shared physical path and the discrete path distance corresponding to each link in the link set are obtained. Based on the mapping relationship between the occurrence time of the semi-adjacency event of each link in the link set and the discrete path distance, wavefront propagation fitting verification is performed, including: Obtain the set of links on the preset shared physical path and the discrete path distance corresponding to each link in the set; Specifically, the system obtains a set of links on a pre-defined shared physical path by accessing the physical asset information of network devices. A shared physical path refers to a network topology where several links share the same physical path or installation location, such as a single cable, the same shaft, or the same cabinet. The system identifies all links located on the same physical path by querying device asset management information and organizes these links into a link set. Each link in the set represents a physical link connecting two devices in the network, and their physical paths are similar or shared.
[0028] To accurately represent the relative position of each link on the shared physical path, the system assigns a discrete path distance to each link in the link set. The discrete path distance refers to the relative distance between the link and the start or end point of the physical path. This distance is typically based on the hierarchical location of the device or connection point, the number of hops between devices, or the cable length. The system determines the discrete path distance of each link by querying the physical location hierarchy of the device, such as the server room, floor, or rack, or based on the physical connection path information between devices. In some scenarios, the discrete path distance can also be calculated based on the actual cable length or specific measurements of the device's installation location. By acquiring and calculating the discrete path distance of each link, the system can provide fundamental data support for subsequent wavefront propagation analysis, ensuring that the positional relationships between links are accurately represented, and providing strong evidence for link fault analysis or optimization.
[0029] Extract the first half-adjacency moment from the half-adjacency event set of each link in the link set; A linear regression model was established with discrete path distance as the independent variable and the first semi-adjacent time as the dependent variable. The wavefront propagation speed and wavefront initiation time were obtained by fitting the model. Calculate the mean square of the fitting residuals of the linear regression model; It's important to note that the half-adjacency moment refers to the moment when the adjacency identifiers of the devices at both ends of a link first become inconsistent. Specifically, the devices at both ends of a link communicate via a neighbor discovery protocol (such as LLDP), and the adjacency identifier represents the connection status of the devices at both ends. When the adjacency identifiers of the devices at both ends of a link are unequal, it indicates an abnormal link state, usually due to a problem with the physical or logical connection of the link. The half-adjacency moment is the point in time when this inconsistency first occurs; this moment reflects the start time of the link problem and is usually used as an initial indicator of link failure or performance anomalies. The wavefront propagation speed refers to the speed at which the adjacency inconsistency propagates along the shared physical path from the occurrence of a half-adjacency event. The wavefront propagation speed is used to quantify the propagation speed of the adjacency inconsistency in the network and is usually related to the physical location of the link, transmission delay, and link topology. By analyzing the half-adjacency events of multiple links, the wavefront propagation speed can be calculated, thus helping to analyze the propagation pattern and scope of link failures. The wavefront initiation moment refers to the initial point in time when a half-adjacency event occurs on a link, that is, the moment when the adjacency identifiers of the first two ends of the link become inconsistent. The wavefront initiation time, along with the wavefront propagation speed, is used to describe the propagation characteristics of link failures or performance anomalies in the topology.
[0030] Specifically, the system extracts the first half-adjacency moment from the half-adjacency event set of each link in the link set. The half-adjacency event set contains all events where links exhibit adjacency inconsistencies at a specific time. Based on the occurrence time of the event, the system selects the earliest occurring half-adjacency event from the half-adjacency event set of each link, i.e., the first half-adjacency moment. This moment represents a significant adjacency inconsistency on the link at a specific point in time, usually related to link performance anomalies or failures.
[0031] A linear regression model was established for wavefront propagation analysis. In this model, discrete path distance was used as the independent variable, and the first half-adjacency moment was used as the dependent variable. The discrete path distance was calculated from the physical location of the devices or the distance of the connection path between the devices. The first half-adjacency moment represents the moment when the link first experiences a half-adjacency event, reflecting the time when the link state changes. By pairing discrete path distance and the first half-adjacency moment, a linear regression method was used to fit these data points, thereby calculating the wavefront propagation speed and the wavefront initiation moment. The wavefront propagation speed represents the speed at which an adjacency inconsistency event propagates between links on a shared physical path, and the wavefront initiation moment represents the starting point of propagation.
[0032] To evaluate the accuracy of the regression model, the system calculates the mean square of the residuals of the linear regression model. The mean square of the residuals is the average of the sum of squares of the differences between the predicted values and the actual observed values of the fitted model, and is used to measure the model's fit. A smaller mean square of residuals indicates a better model fit, accurately reflecting the characteristics of wavefront propagation; a larger mean square of residuals indicates a poor model fit, which may be due to data noise or unreasonable model specification.
[0033] The confidence rendering module is used to calculate the edge confidence of the link in the link set based on the semi-adjacent event set if the result of the wavefront propagation fitting verification meets the preset conditions, and to determine and render the minimum influence domain based on the edge confidence. In an embodiment of the present invention, the edge confidence of a link is calculated based on a set of semi-adjacent events in the link set, and the minimum influence domain is determined and rendered based on the edge confidence, including: When the wavefront propagation speed is greater than zero, the number of links involved in the fitting is greater than the preset minimum number of effective edges, and the mean square of the fitting residuals is less than the preset residual threshold, the edge confidence of the links is calculated in the link set based on the set of semi-adjacent events. It should be noted that edge confidence is a quantity used to characterize the degree of credibility of a link edge in the network management client topology graph as a real and stable connection under the current observation conditions. Its formation is based on the adjacency consistency stability reflected by the port states at both ends of the link and the set of half-adjacency events. The edge confidence decreases numerically as the number of events in the set of half-adjacency events increases and is suppressed to unusable when the port state is not enabled. This reflects whether the link edge is in a short-term adjacency asynchronous boundary state triggered by the phase-locked loop time window, thus providing a quantifiable basis for subsequently determining the minimum influence domain. This allows the topology to distinguish between link edges that have long-term stable connectivity and those that are prone to ghost connections under physical disturbances.
[0034] Specifically, after completing the wavefront propagation fitting, the wavefront propagation velocity, the number of links involved in the fitting, and the mean square of the fitting residuals are read from the fitting output. Based on these, a condition determination is performed. The wavefront propagation velocity must be greater than zero to exclude non-propagation cases where time and discrete path distance are inversely related. The number of links involved in the fitting must be greater than a preset minimum effective number of edges to reduce the risk of misfitting due to random noise. The mean square of the fitting residuals must be less than a preset residual threshold to ensure the propagation model's explanatory power for the first half-adjacency time. The minimum effective number of edges is set to 5, a value based on the premise that a shared physical path typically covers at least several adjacent link segments. The system can exhibit a continuous propagation relationship, and when the number of samples is less than 5, the linear fitting shows a significantly increased sensitivity to outliers. The residual threshold is set to 30 seconds, which matches the 10-second sampling period and allows time jitter of no more than three sampling periods to still be considered as the same wavefront propagation event. A strict inequality is used to ensure that the propagation direction is consistent with the definition of discrete path distance when the wavefront propagation velocity is greater than zero. When the above three conditions are met simultaneously, the system calculates the edge confidence score for each link in the link set. The edge confidence score is used to quantify the credibility of the link as a topological edge and emphasizes that the accumulation of semi-adjacency within the phase-locked time window will reduce the credibility. The formula for the edge confidence score is: In the formula, The edge confidence of the link. This indicates the number of events in the set of half-adjacent events formed within the phase-locked loop time window of this link. This represents the preset attenuation factor. Indicates an indicator function, and These represent the port states at both ends of the link at the current rendering time. The value indicates the enabled state. An exponential decay form is used because an increase in the number of half-adjacent events typically indicates that the link is in a more frequent, short-term adjacency asynchronous state. The confidence level should decrease rapidly in a non-linear manner to highlight its impact on topology reliability. Simultaneously, the port status serves as a hard constraint to ensure that the link should not be assigned a high confidence level when any port is disabled. The decay factor is preferably set to 0.7. This value ensures that when the number of half-adjacent events reaches 3, the edge confidence decays to approximately 0.12. This effectively distinguishes recurring half-adjacent links from occasional inconsistent links within a phase-locked loop time window, thus providing a stable quantitative input for subsequently determining the minimum influence domain in the link set based on the low-confidence edge threshold.
[0035] Links in the link set whose edge confidence is lower than a preset low confidence edge threshold are defined as the minimum influence domain; In the network management client topology diagram, only links belonging to the minimum influence domain are highlighted, and temperature gradient-related semi-adjacent wavefront anomalies are marked. In the display area associated with the minimum influence domain, the wavefront initiation time and wavefront propagation speed obtained by wavefront propagation fitting verification are output; It should be noted that the minimum impact domain refers to the smallest subset of links selected by edge confidence judgment within the set of links on the shared physical path. This subset consists of links with edge confidence scores below a preset low-confidence edge threshold. It is used to characterize the connection range that contributes most unstablely to the reliability of topology connectivity and has the most concentrated impact on services under the current observation conditions. The decrease in edge confidence comes from the accumulation of events in the set of semi-adjacent events within the phase-locked time window and the degradation of connection consistency reflected by port state constraints. Therefore, the minimum impact domain can delineate the abnormal areas that need to be focused on with the minimum number of link boundaries and avoid misinterpreting abnormal areas. Irrelevant links are included in the rendering scope; temperature gradient-related semi-adjacent wavefront anomalies refer to the short-term inconsistency of neighbor identifiers at both ends of a link within the temperature gradient phase-locked time window, forming a set of semi-adjacent events. Moreover, these semi-adjacent events exhibit a temporal propagation relationship related to discrete path distance among multiple links on a shared physical path, and this anomaly is verified by wavefront propagation fitting. This anomaly reflects that links are more prone to asynchronous adjacency recognition and time-delayed diffusion along the physical path during periods of rapid temperature change in the same direction. As a result, on the topology map, it manifests as a semi-adjacent alarm feature where local link groups appear in a wavefront pattern in time.
[0036] Specifically, after calculating the edge confidence of each link within the link set, links with edge confidence below a preset low-confidence edge threshold are defined as the minimum influence domain. The low-confidence edge threshold is set to 0.30, a value determined based on the interpretability of exponentially decaying edge confidence. This ensures that, with all ports enabled, when the number of events in the half-adjacent event set within the same phase-locked time window reaches two or more, the edge confidence usually falls below this threshold. This distinguishes recurring half-adjacent links from occasional inconsistent adjacency links and prevents the minimum influence domain from overexpansion. The generation of the minimum influence domain is limited by the link set. The system traverses each link in the link set and reads its edge confidence. Links with edge confidence less than 0.30 are added to the minimum influence domain. Simultaneously, a data structure consistent with the topology rendering engine is constructed in memory to record the device identifiers at both ends of the link, the port identifiers, and the corresponding edge confidence values, facilitating subsequent rendering updates only on this subset.
[0037] The rendering phase of the network management client's topology map adopts a layered refresh mechanism. The basic topology layer maintains the original layout and the drawing state of non-minimum influence domain links unchanged. The alarm overlay layer only applies a highlight style to the links in the minimum influence domain and displays fixed text labels next to the links indicating the semi-adjacent wavefront anomalies related to the temperature gradient. This ensures that maintenance personnel can intuitively see the minimum set of abnormal links within the same topology view without being interfered with by unrelated links. The display area associated with the minimum influence domain is implemented using a details panel or information overlay linked to the topology map. The system writes the wavefront start time and wavefront propagation speed output from the wavefront propagation fitting verification as structured fields into this display area. The wavefront start time is used to indicate when the anomaly on the shared physical path begins to appear and corresponds to the timestamp of the phase-locked time window. The wavefront propagation speed is used to indicate how fast the anomaly propagates along the discrete path distance direction and is used to explain the time lag relationship of the first semi-adjacent time of different links within the minimum influence domain. The display area is updated synchronously when the user selects any minimum influence domain link or when the system automatically pops up an alarm card, thus achieving a consistent presentation of minimum influence domain highlighting and wavefront parameter output.
[0038] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A network management client topology management system, characterized in that, include: The alignment acquisition module is used to perform alignment sampling on the devices at both ends of the network link according to a preset sampling period, and to obtain the neighbor identifiers, port status and temperature at both ends of the link to form an observation sequence. The phase-locked gate module is used to determine the phase-locked time window that satisfies the temperature gradient phase-locked condition based on the rate of change and direction of change of temperature in the observation sequence. The event extraction module is used to identify the state where the neighbor identifiers at both ends of the network link are inconsistent within the phase-locked loop time window, and extract half-adjacent events that meet the preset duration to form a half-adjacent event set. The wavefront fitting module is used to obtain the set of links on the preset shared physical path and the discrete path distance corresponding to each link in the set. Based on the mapping relationship between the occurrence time of the semi-adjacency event of each link in the set and the discrete path distance, the wavefront propagation fitting verification is performed. The confidence rendering module is used to calculate the edge confidence of the link in the link set based on the semi-adjacent event set if the result of the wavefront propagation fitting verification meets the preset conditions, and to determine and render the minimum influence domain based on the edge confidence.
2. The network management client topology management system according to claim 1, characterized in that, The observation sequence consists of: neighbor identifiers at both ends of the link, port status at both ends of the link, and temperature values at both ends of the link.
3. The network management client topology management system according to claim 1, characterized in that, Based on the rate of change and direction of change of temperature in the observation sequence, determine the phase-locked time window that satisfies the temperature gradient phase-locking condition, including: Calculate the sum of the temperature changes at both ends of the link between the current sampling time and the previous sampling time, and divide the sum by the sampling period to obtain the temperature gradient; Determine whether the temperature gradient is greater than a preset gradient threshold; Determine whether the direction of temperature change at both ends of the link is consistent; If the temperature gradient is greater than the preset gradient threshold and the temperature changes at both ends of the link are in the same direction, then the current sampling time is determined to be within the phase-locked time window.
4. The network management client topology management system according to claim 1, characterized in that, Form a semi-adjacent event set, including: An adjacency inconsistency indication is determined when the neighbor identifiers at both ends of the link are not equal; The duration during which the adjacency inconsistency indicator remains true is recorded. Events that occur within the phase-locked loop time window and whose duration is between the preset minimum and maximum half-adjacency durations are selected to form a half-adjacency event set.
5. A network management client topology management system according to claim 1, characterized in that, Perform wavefront propagation fitting validation, including: Extract the first half-adjacency moment from the half-adjacency event set of each link in the link set; A linear regression model was established with discrete path distance as the independent variable and the first semi-adjacent time as the dependent variable. The wavefront propagation speed and wavefront initiation time were obtained by fitting the model. Calculate the mean square of the fitting residuals of the linear regression model.
6. A network management client topology management system according to claim 5, characterized in that, The conditions for validating wavefront propagation fitting are to be met simultaneously: the wavefront propagation velocity is greater than zero, the number of links involved in the fitting is greater than the preset minimum number of effective edges, and the mean square of the fitting residuals is less than the preset residual threshold.
7. A network management client topology management system according to claim 6, characterized in that, Calculate the edge confidence of the link, including: Edge confidence is calculated based on the exponential decay function only when both physical ports at both ends of the link are in the enabled state. The number of events in the semi-adjacent event set is used as a variable for the decay term, and a preset decay factor is used as the coefficient of the decay term; the more events there are, the lower the edge confidence.
8. A network management client topology management system according to claim 7, characterized in that, Determine and render the minimum impact area, including: Links in the link set whose edge confidence is lower than a preset low confidence edge threshold are defined as the minimum influence domain; In the network management client topology diagram, only links belonging to the minimum influence domain are highlighted, and temperature gradient-related semi-adjacent wavefront anomalies are marked. In the display area associated with the minimum influence domain, the wavefront initiation time and wavefront propagation speed, obtained by wavefront propagation fitting verification, are output.
9. The method according to claim 1, characterized in that, The set of links on the shared physical path is bound based on the physical asset information of the network devices; the distance of the discrete path is determined based on the physical location level of the device or the cable length.