Distributed network latency measurement method and system

Abstract

The application discloses a distributed network delay measurement method and system, relates to the technical field of network delay measurement, and predicts and generates an abnormal report after a network abnormal event is assigned with a corresponding topic, generates an abnormal density of the abnormal event from network abnormal event data, selects key influencing factors from correlation data when the abnormal density exceeds expectation, and generates a corresponding response scheme by matching the target characteristics to the network abnormal event; after abnormal detection is performed on network quality, if the network quality is abnormal, a network quality analysis report is generated from visualized network quality data, a current network measurement scene is identified, the current network measurement scene meets preset rule conditions, a corresponding measurement scheme is output and optimized, and the measurement scheme is executed to complete the current delay measurement process. The measurement scheme and the actual scene are more adaptive, and interference of abnormal measurement environment and network quality on the network measurement process is effectively prevented.

Description

Technical Field

[0001] This invention relates to the field of network latency measurement technology, specifically to a distributed network latency measurement method and system. Background Technology

[0002] With the rapid development of information technology, computer networks have become the core infrastructure for modern enterprise and organizational operations, carrying massive amounts of data transmission, application services, and business processes. Especially driven by emerging technologies such as cloud computing, big data, and the Internet of Things, network architectures are becoming increasingly complex, their scale is constantly expanding, and the dynamism and diversity of network environments are significantly enhanced. This complex network environment places higher demands on the stability and reliability of network quality, making real-time monitoring and precise management of network performance crucial for ensuring efficient network operation. To comprehensively understand network status, defining different types of network measurement scenarios is particularly important. Various measurement scenarios, such as latency measurement, bandwidth measurement, packet loss rate measurement, and device performance measurement, provide multi-dimensional data collection and analysis methods for specific network anomalies, ensuring that the network management system can flexibly apply the most suitable measurement schemes for different operating environments and device states, achieving comprehensive monitoring and optimization of network performance.

[0003] Chinese invention patent application CN106961365A discloses a network latency measurement method based on the TCP protocol. This method includes: establishing a first optimization objective function for forward latency; establishing a second optimization objective function for return latency; weighting the first and second objectives to obtain a total optimization objective function; and using an ant colony algorithm to iteratively optimize and search the total objective function to solve for an estimated value of the network latency. This invention proposes a network latency measurement method based on the TCP protocol that eliminates the need to send a large number of probe packets or perform global clock calibration, providing a simple and efficient way to calculate network transmission latency.

[0004] Based on the above applications and existing technologies:

[0005] While existing technologies have made some progress in network anomaly detection and response, significant shortcomings remain in dynamically identifying and classifying network anomalies and defining corresponding network measurement scenarios. Specifically, traditional static rule-based response mechanisms rely on preset thresholds and conditions, lacking the ability to adapt to real-time data changes and struggling to dynamically adjust measurement scenarios to cope with complex and ever-changing network environments. This results in untimely or inaccurate responses from existing systems when faced with new or complex network anomalies, hindering the rapid location and resolution of network problems.

[0006] Furthermore, existing measurement schemes are mostly statically designed and lack optimization mechanisms, making it difficult to quickly generate optimal measurement schemes under multiple constraints, thus affecting the efficiency and reliability of network management. Therefore, there is an urgent need for an automated system that can dynamically identify and classify network anomalies and optimize corresponding network measurement scenarios and response measures based on real-time data and intelligent algorithms. This system would enhance the intelligence and adaptability of network management and meet the pressing needs of modern complex network environments for efficient and accurate management.

[0007] To this end, the present invention provides a distributed network latency measurement method and system. Summary of the Invention

[0008] (a) Technical problems to be solved

[0009] To address the shortcomings of existing technologies, this invention provides a distributed network latency measurement method and system. It predicts and generates anomaly reports by assigning corresponding topics to network anomaly events. The system generates anomaly density from network anomaly event data. When the anomaly density exceeds expectations, key influencing factors are selected from correlation data, and corresponding response schemes are generated by matching target features with network anomaly events. After detecting network quality anomalies, if anomalies are found, a network quality analysis report is generated from visualized network quality data. The current network measurement scenario is identified. If the current network measurement scenario meets preset rules, a corresponding measurement scheme is output and optimized, and the measurement scheme is executed to complete the current latency measurement process. This makes the measurement scheme more adaptable to the actual scenario and effectively prevents abnormal measurement environments and network quality from interfering with the network measurement process; thus solving the technical problems described in the background art.

[0010] (II) Technical Solution

[0011] To achieve the above objectives, the present invention provides the following technical solution: a distributed network latency measurement method and system. The distributed network latency measurement method includes: collecting operational scenario data of network devices and performing real-time detection of network anomalies; if a network anomaly meets preset rule conditions, generating a corresponding event and issuing an alarm command.

[0012] After assigning corresponding topics to network anomalies, predictions are made and anomaly reports are generated. The anomaly density D(t,x) of the anomalies is generated from the network anomaly event data. When the anomaly density D(t,x) exceeds the expectation, a key factor extraction instruction is sent to the outside.

[0013] After selecting key influencing factors from the correlation data, the target features are screened out using a sparse regression model and variance inflation factor. The target features are then used to match network anomaly events and generate corresponding response schemes.

[0014] After performing anomaly detection on network quality, if network quality is found to be abnormal, the changes in network data of each network device are visualized.

[0015] After generating a network quality analysis report from visualized network quality data, the current network measurement scenario is identified. If the current network measurement scenario meets preset rule conditions, a corresponding measurement plan is output and optimized, and the measurement plan is executed to complete the current latency measurement process. Specifically, the network quality of each node is mapped to a corresponding color intensity I(Q), and the contrast of the intermediate region is enhanced through a smooth transition.

[0016]

[0017] Where: I(Q) is the mapped color intensity, ranging from [0, I... max ], usually I max =255, Q is the network quality index, normalized to [0,1]; k is the slope parameter, which controls the steepness of the Sigmoid function; Q0 is the midpoint parameter, which determines the center position of the Sigmoid function.

[0018] Furthermore, sensor networks are deployed on network devices to collect environmental conditions and network device operating status data in real time, combined with the obtained network device operating scenario data;

[0019] The data from different sources are converted into different formats. After synchronizing the data from multiple sources by timestamp, a Bayesian network is applied to fuse the data from multiple sources.

[0020] Furthermore, network anomaly events are defined and detected in real time based on an event-driven architecture. Anomaly event types are predefined, and each type of anomaly event should have the following attributes: event ID, type, timestamp, source, details, and priority. These are then aggregated to generate a set of network anomaly events. A rule engine is integrated into the real-time processing framework to define detection rules for network anomaly events based on preset thresholds and conditions.

[0021] Furthermore, upon receiving an alarm command, the Apache Kafka topic mechanism is used to assign different types of network anomaly events to corresponding topics, and multiple event receivers are configured so that the event receivers focus on the corresponding event topics and are responsible for receiving and processing network anomaly events.

[0022] Predict network anomalies using the trained anomaly prediction model; periodically generate anomaly reports for network anomalies, including event type, frequency of occurrence, handling effect, and possible anomalies in the next stage, and send the anomaly reports to the event receiver.

[0023] Furthermore, network anomaly event data within the current stage is collected and recorded, including anomaly time nodes, anomaly location nodes, and corresponding anomaly severity. The anomaly density D(t,x) of the anomaly events is generated from the network anomaly event data, as follows:

[0024]

[0025] Where: D(t,x) represents the anomaly density at time t and location x, T is the time window length, Ω is the spatial survey area; s(τ,y) represents the anomaly severity at time τ and location y, K... t (t-τ) is the time kernel function; K x (xy) is the spatial kernel function; x and y represent vectors of spatial location, τ is the integral variable representing time, and y is the integral variable representing spatial location.

[0026] Furthermore, after receiving the key factor extraction instruction, the influencing factors that are highly correlated with the network state are initially screened out; principal component analysis is applied to reduce the dimensionality of the initially screened influencing factors to obtain the corresponding key influencing factors.

[0027] By using a sparse regression model with minimum absolute contraction and LASSO selection operator regression, key influencing factors are further screened to obtain the screened key influencing factors.

[0028] Furthermore, the variance inflation factor (VIF) of the key influencing factors after screening is calculated. Key influencing factors with VIF values ​​exceeding a preset threshold are eliminated or merged, and the screened key factors are used as target features.

[0029] After labeling network anomaly events with target features, automatic response rules are defined based on the event type and attributes of the network anomaly events. When a network anomaly event occurs, a response plan is automatically generated according to the event type and automatic response rules.

[0030] Furthermore, each network device in the network is treated as an independent node in the flowchart, and network quality data of each independent node is measured and collected; real-time network quality data is used as input, and the trained abnormal network detection model is used to detect anomalies and obtain anomaly detection data.

[0031] Each node is assigned network quality data, environmental condition data, and equipment operating status data corresponding to the stage; directed edges connect each node to represent the network quality change status; heat map elements are overlaid on each node based on the flowchart, and color gradients are used to represent the degree of network quality.

[0032] Furthermore, network quality data is bound by time series, automatically identifying and marking key time points and stages of network quality data growth or decline, and automatically generating network quality analysis reports based on changes in network quality.

[0033] The system identifies and classifies current network quality anomalies, defines different types of network measurement scenarios, and builds a rule-based automated response engine. Based on current device operating data and environmental conditions, it uses a genetic algorithm to optimize the currently matched measurement scheme and quickly generate the optimal measurement scheme for specific anomaly scenarios.

[0034] The distributed network latency measurement system includes an abnormal event detection unit, which collects network device operation scenario data and performs real-time detection of network abnormal events. If the network abnormal event meets the preset rule conditions, it generates the corresponding event and issues an alarm command.

[0035] The event topic allocation unit assigns corresponding topics to network anomaly events, makes predictions and generates anomaly reports. It generates anomaly density of anomaly events from network anomaly event data. When the anomaly density exceeds expectations, it sends a key factor extraction instruction to the outside.

[0036] The factor identification unit selects key influencing factors from the correlation data, uses a sparse regression model and variance inflation factor to screen out target features, and generates corresponding response schemes by matching the target features with network anomaly events.

[0037] The network quality detection unit performs anomaly detection on the network quality. If anomalies are found, it visualizes the changes in network data of each network device.

[0038] The measurement scheme generation unit generates a network quality analysis report from the visualized network quality data, identifies the current network measurement scenario, and outputs and optimizes the corresponding measurement scheme when the current network measurement scenario meets the preset rule conditions. The measurement scheme is then executed to complete the current latency measurement process.

[0039] (III) Beneficial Effects

[0040] This invention provides a distributed network latency measurement method and system, which has the following beneficial effects:

[0041] 1. Collect and detect network anomalies in the current stage, mark and classify various types of network anomalies, and send alarms to the outside when network anomalies occur, so as to deal with the existing network anomalies in a timely manner.

[0042] 2. When an abnormal event occurs, add a topic describing the corresponding abnormal event. This will help select and match the appropriate personnel to handle the network abnormal event, thereby improving the efficiency of abnormal event handling.

[0043] 3. After collecting the status data of network anomalies, construct the anomaly density D(t,x). Based on the anomaly density D(t,x), comprehensively evaluate the states that continuously generate anomalies in the current stage, determine the density and overall severity of network anomalies, and determine whether it is necessary to adjust, control or maintain each network device based on the comprehensive judgment. This will guide subsequent network optimization, control and testing.

[0044] 4. When network anomalies need to be handled, key influencing factors are extracted to formulate or match corresponding handling strategies, and targeted optimization or control is carried out; several external factors that may cause the current network status to be abnormal are identified, and the reliability of influencing factors can be guaranteed by extracting and screening key factors at multiple levels; on this basis, the handling effect can be improved when dealing with network devices and network anomalies.

[0045] 5. By visualizing network devices and corresponding network quality data, the current network status can be displayed in real time, which also facilitates source tracing when network anomalies occur, improving the targeted nature of the handling.

[0046] 6. When the current network quality is abnormal, after considering the environmental conditions of the current network equipment, the key factors causing the abnormality, and the current network measurement scenario, output and optimize in a targeted manner to obtain a corresponding measurement plan. When the network needs to be measured, the measurement plan should be more adapted to the actual scenario, effectively preventing the current abnormal measurement environment and network quality from interfering with the current network measurement process and affecting the reliability and effectiveness of network measurement. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the distributed network latency measurement method of the present invention;

[0048] Figure 2 This is a schematic diagram of the distributed network delay measurement system of the present invention. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] Please see Figure 1 This invention provides a distributed network latency measurement method, including,

[0051] Step 1: After collecting operational scenario data of network devices, perform real-time detection of network anomalies. If a network anomaly meets the preset rule conditions, generate the corresponding event and issue an alarm command.

[0052] Step one includes the following:

[0053] Step 101: Construct an electronic map covering the network area and mark each network device on the electronic map; deploy sensor networks on the network devices, such as environmental sensors and device status monitoring sensors, to collect environmental conditions and network device operating status data in real time, and combine them with the obtained network device operating scenario data.

[0054] The data from different sources are converted into different formats. After synchronizing the data from multiple sources by timestamp, a Bayesian network is applied to fuse the data from multiple sources.

[0055] When using it, when measuring a distributed network, the environmental conditions of the network devices are detected, and targeted measures are taken when abnormal environmental conditions occur.

[0056] Step 102: Define network anomaly events based on event-driven architecture and detect them in real time. Predefine the types of anomaly events, where each type of anomaly event should have the following attributes: event ID, type, timestamp, source, details, and priority; for example, network equipment failure events and power supply anomaly events, etc. After summarizing, generate a set of network anomaly events.

[0057] A rule engine is integrated into the real-time processing framework to define detection rules for network anomalies based on preset thresholds and conditions. When real-time data meets a certain rule condition, a corresponding event is generated and an alarm command is issued.

[0058] When using this method, refer to steps 101 and 102:

[0059] After determining the current environmental conditions, collect and detect existing network anomalies in the current stage, mark and classify various types of network anomalies, and send out alarms when network anomalies occur, so as to deal with existing network anomalies in a timely manner.

[0060] Step 2: After assigning corresponding topics to network anomaly events, make predictions and generate anomaly reports. Generate the anomaly density D(t,x) of the anomaly events from the network anomaly event data. When the anomaly density D(t,x) exceeds the expectation, send a key factor extraction instruction to the outside.

[0061] Step two includes the following:

[0062] Step 201: After receiving the alarm command, use the Apache Kafka topic mechanism to assign different types of network anomaly events to corresponding topics, configure multiple event receivers, and the event receivers pay attention to the corresponding event topics and are responsible for receiving and processing network anomaly events. For example, the event receiver for network device failure anomaly events is responsible for recording fault information, triggering maintenance processes, and adjusting production parameters to adjust the current network status.

[0063] When using this technology, considering that there are various types of network anomalies, a topic describing the anomaly is added when it occurs. This allows for the selection and matching of appropriate personnel to handle network anomalies, which can improve the efficiency of anomaly handling.

[0064] Step 202: After labeling the historical event data, train the neural network with the labeled data to obtain the trained anomaly prediction model; use the trained anomaly prediction model to predict possible future network anomalies and obtain prediction data.

[0065] Regularly generate anomaly reports for network anomalies, including event type, frequency of occurrence, handling effect, and possible anomalies in the next stage, and send the anomaly reports to the event receiver; by predicting anomalies, take proactive measures to handle potential network anomalies in advance.

[0066] Step 203: Collect and record network anomaly event data within the current stage, including anomaly time nodes, anomaly location nodes, and corresponding anomaly severity, etc. Generate the anomaly density D(t,x) of anomaly events from the network anomaly event data, as follows:

[0067]

[0068] Where: D(t,x) represents the anomaly density at time t and location x; T is the time window length, representing the past time range considered; Ω is the spatial investigation area, representing the geographical or network topology range of the anomaly event; s(τ,y) represents the anomaly severity at time τ and location y; K... t (t-τ) is the time kernel function, usually a Gaussian kernel function, used to smooth out anomalous events in the time dimension; K x (xy) is the spatial kernel function, usually a Gaussian kernel function, used to smooth out anomalous events in the spatial dimension; x and y represent vectors of spatial location, the specific dimensions of which are determined according to the network topology (such as two-dimensional or three-dimensional coordinates); τ is the integral variable representing time, and y is the integral variable representing spatial location;

[0069] To achieve smooth processing in both time and space dimensions, the Gaussian kernel function is often used, and its specific form is as follows:

[0070]

[0071] Where: σ t σ is the bandwidth parameter of the time kernel function, controlling the smoothness over time; x is the bandwidth parameter of the spatial kernel function, which controls the smoothness in the spatial dimension; n is the number of spatial dimensions (e.g., n = 2 in two-dimensional space); ∥xy∥ is the Euclidean distance between position vectors x and y;

[0072] Based on historical data and anticipated management of network anomalies, a density threshold is pre-set;

[0073] When the obtained anomaly density D(t,x) exceeds expectations, it indicates that the current network anomaly events are relatively frequent and the density is high. At this time, a key factor extraction instruction is sent to the outside.

[0074] When using this method, refer to steps 201 to 203:

[0075] Based on the occurrence of continuous network anomalies, an anomaly density D(t,x) is constructed after collecting the status data of the network anomalies. The anomaly density D(t,x) is used to comprehensively evaluate the state of continuous anomalies in the current stage, and to determine the density and overall severity of network anomalies. Based on the comprehensive judgment, it can be determined whether it is necessary to adjust, control or maintain various network devices, which can guide subsequent network optimization, control and testing.

[0076] Step 3: After selecting key influencing factors from the correlation data, use sparse regression model and variance inflation factor to screen out target features, and generate corresponding response schemes by matching network anomaly events with the target features.

[0077] Step three includes the following:

[0078] Step 301: After receiving the key factor extraction instruction, correlation analysis is used to initially screen out influencing factors that are highly correlated with the network status in the operating environment and equipment operating status; principal component analysis is applied to reduce the dimensionality of the initially screened influencing factors, and multidimensional features are converted into a few principal components by constructing a principal component matrix to retain the main variation information of the data and obtain the corresponding key influencing factors.

[0079] When in use, key factors are extracted to identify the key factors affecting the current network anomaly from among many factors. When it is necessary to deal with the network anomaly, the extracted key influencing factors can be used to formulate or match corresponding processing strategies for targeted optimization or control.

[0080] Step 302: Using a sparse regression model with minimum absolute contraction and LASSO selection operator regression, further screen the key influencing factors to obtain the screened key influencing factors, as follows:

[0081] By introducing an L1 regularization term and penalized regression coefficients, insignificant feature coefficients are automatically compressed to zero, achieving feature selection and model simplification.

[0082]

[0083] Among them, y i x is the correlation coefficient. ij For the j-th feature, β j The regression coefficients are λ, and the regularization parameter is λ. The optimal value is determined through cross-validation.

[0084] Considering that after initially identifying key influencing factors, further screening of these factors can be conducted, it can help improve maintenance efficiency and reduce the corresponding workload;

[0085] Step 303: Calculate the variance inflation factor (VIF) of the key influencing factors after screening. For key influencing factors after screening with VIF values ​​exceeding a preset threshold (e.g., 10), remove or merge them to reduce the linear correlation between features. Use the key factors after screening as target features. in, The regression determination coefficient of feature j with all other features;

[0086] After labeling network anomalies with target features, automatic response rules are defined based on the event type and attributes of the network anomalies. When a network anomaly occurs, a response plan is automatically generated according to the event type and automatic response rules; for example, automatic control of the corresponding equipment or control of the current environmental conditions.

[0087] When using this method, refer to steps 301 to 303:

[0088] When network anomalies occur frequently in the current phase, several external factors that may cause the current network status to be abnormal can be identified, such as environmental conditions. By extracting and screening key factors at multiple levels, the factors that have the greatest impact on network anomalies can be identified, which can ensure the reliability of the influencing factors. Based on this, the processing effect of network equipment and network anomalies can be improved.

[0089] Step 4: After performing anomaly detection on network quality, if network quality is found to be abnormal, visualize the changes in network data of each network device.

[0090] Step four includes the following:

[0091] Step 401: Train the Isolation Forest algorithm using the labeled sample data to obtain the trained anomaly detection model;

[0092] Treat each network device in the network as an independent node in the flowchart, and measure and collect network quality data for each independent node, including network performance indicators, connection quality indicators, and service quality indicators.

[0093] Using real-time network quality data as input, the trained anomaly detection model is used to detect anomalies, verify whether there are anomalies in the current network quality data, and obtain anomaly detection data; thus, the existing network anomaly events are verified based on the actual anomaly detection data.

[0094] It can also verify the anomaly detection data and obtain the corresponding detection feedback data, such as reliability and accuracy; it introduces an adaptive threshold setting method based on dynamic statistical analysis, and adjusts the anomaly detection threshold in real time by combining the changing trend of historical data and detection feedback data to reduce false alarms and false negatives.

[0095] Step 402: Assign network quality data, environmental condition data, and equipment operating status data for each node corresponding to the stage; connect each node with directed edges to represent the network quality change status; overlay heat map elements on each node based on the flowchart, and use color gradients to represent the degree of network quality.

[0096] When using this method, refer to steps 401 and 402:

[0097] When in use, based on the current specific network quality data, the network devices and corresponding network quality data are visualized to display the current network status in real time. This also facilitates tracing the source of network anomalies and further improves the targeting of the processing.

[0098] Step 5: After generating a network quality analysis report from the visualized network quality data, identify the current network measurement scenario. If the current network measurement scenario meets the preset rule conditions, output the corresponding measurement plan and optimize it. Execute the measurement plan to complete the current delay measurement process.

[0099] Step five includes the following:

[0100] Step 501: Map the network quality of each node to the corresponding color intensity I(Q). Enhance color contrast using linear or nonlinear mapping functions to improve recognizability. The Sigmoid function enhances contrast in intermediate regions through smooth transitions and is suitable for scenarios requiring balanced contrast across the entire range.

[0101]

[0102] Where: I(Q) is the mapped color intensity, ranging from [0, I... max ], usually I max =255, Q is the network quality index, normalized to [0,1]; k is the slope parameter, which controls the steepness of the Sigmoid function; Q0 is the midpoint parameter, which determines the center position of the Sigmoid function.

[0103] By binding network quality data to a time series, the system automatically identifies and marks key time points and stages when network quality data increases or decreases, and automatically generates network quality analysis reports based on changes in network quality.

[0104] When used, it improves the readability and understanding of data through systematic data integration, provides powerful visualization tools, supports efficient handling of network anomalies, and ensures the consistency and efficiency of network anomaly handling.

[0105] Step 502: Identify and classify the current network quality anomalies, and define different types of network measurement scenarios, including abnormal network equipment operation, abnormal environmental conditions, and abnormal network security.

[0106] A rule-based automated response engine is built. When the current network measurement scenario meets the preset rule conditions, the corresponding measurement plan is re-output. Based on the current device operation data and environmental condition data as feedback, a genetic algorithm is used to optimize the currently matched measurement plan. For specific abnormal situations, the optimal measurement plan is quickly generated and the measurement plan is executed to complete the current delay measurement process.

[0107] When using this method, refer to steps 501 and 502:

[0108] When network quality is abnormal, the system takes into account the environmental conditions of the network devices, the key factors causing the abnormality, and the current network measurement scenario. It then outputs and optimizes the solution to obtain a corresponding measurement plan. When the network needs to be measured, the optimized measurement plan is executed to make the measurement plan more suitable for the actual scenario. This also effectively prevents the abnormal measurement environment and network quality from interfering with the current network measurement process and affecting the reliability and effectiveness of the network measurement.

[0109] Please see Figure 2 This invention provides a distributed network latency measurement system, comprising:

[0110] The abnormal event detection unit collects network device operation scenario data and performs real-time detection of network abnormal events. If the network abnormal event meets the preset rule conditions, it generates the corresponding event and issues an alarm command.

[0111] The event topic allocation unit assigns corresponding topics to network anomaly events, makes predictions and generates anomaly reports. It generates anomaly density of anomaly events from network anomaly event data. When the anomaly density exceeds expectations, it sends a key factor extraction instruction to the outside.

[0112] The factor identification unit selects key influencing factors from the correlation data, uses a sparse regression model and variance inflation factor to screen out target features, and generates corresponding response schemes by matching the target features with network anomaly events.

[0113] The network quality detection unit performs anomaly detection on the network quality. If anomalies are found, it visualizes the changes in network data of each network device.

[0114] The measurement scheme generation unit generates a network quality analysis report from the visualized network quality data, identifies the current network measurement scenario, and when the current network measurement scenario meets the preset rule conditions, it outputs and optimizes the corresponding measurement scheme, and executes the measurement scheme to complete the current delay measurement process.

[0115] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0116] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0117] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0118] 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0119] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A distributed network delay measurement method, characterized in that: include, After collecting operational scenario data of network devices, network anomaly events are detected in real time. If a network anomaly event meets the preset rule conditions, a corresponding event is generated and an alarm command is issued. After assigning corresponding topics to network anomaly events, predictions are made and anomaly reports are generated. The anomaly density of anomaly events is generated from the network anomaly event data. In abnormal density When the results exceed expectations, issue instructions to extract key factors to external parties. Key influencing factors are selected from correlation data, and target features are screened using sparse regression models and variance inflation factors. The target features are then used to match network anomalies and generate corresponding response schemes. Perform anomaly detection on network quality; if anomalies are found, visualize the changes in network data of each network device. Network quality analysis reports are generated from visualized network quality data to identify current network measurement scenarios, including abnormal network equipment operation, abnormal environmental conditions, and abnormal network security. When the current network measurement scenario meets the preset rule conditions, the corresponding measurement plan is output and optimized, and the measurement plan is executed to complete the current latency measurement process; in this process, the network quality of each node is mapped to the corresponding color intensity. The contrast in the middle area is enhanced through a smooth transition, where... ； in: The mapped color intensity, ranging from [0, ... ], , This is a network quality metric, normalized to [0, 1]. The slope parameter controls the steepness of the sigmoid function. The midpoint parameter determines the center position of the Sigmoid function.

2. The distributed network latency measurement method according to claim 1, characterized in that: Deploy sensor networks on network devices to collect environmental conditions and network device operating status data in real time, and combine the acquired network device operating scenario data; convert the operating scenario data from different sources into different formats, synchronize the time of the multi-source data through timestamp synchronization, and then apply a Bayesian network to fuse the multi-source data.

3. The distributed network latency measurement method according to claim 2, characterized in that: Network anomaly events are defined and detected in real time based on an event-driven architecture. Anomaly event types are predefined, and each type of anomaly event should have the following attributes: event ID, type, timestamp, source, details, and priority. After being aggregated, a set of network anomaly events is generated. A rule engine is integrated into the real-time processing framework to define detection rules for network anomaly events based on preset thresholds and conditions.

4. The distributed network latency measurement method according to claim 3, characterized in that: Upon receiving an alarm command, the Apache Kafka topic mechanism is used to assign different types of network anomaly events to corresponding topics. Multiple event receivers are configured so that the event receivers focus on the corresponding event topics and are responsible for receiving and processing network anomaly events. Predict network anomalies using a trained anomaly prediction model; Regularly generate anomaly reports for network anomalies, including event type, frequency of occurrence, and handling effect, as well as possible anomalies in the next stage, and send the anomaly reports to the event receiver.

5. The distributed network latency measurement method according to claim 4, characterized in that: Collect and record network anomaly event data within the current phase, including anomaly time nodes, anomaly location nodes, and corresponding anomaly severity. Generate anomaly density from the network anomaly event data. The method is as follows: ； in: In time and location The abnormal density at the location The time window length, For spatial research area; In time and location The degree of abnormality at the site, For time kernel function, For spatial kernel functions; It is a spatial position vector. For the integral variable representing time, Let be the integral variable representing spatial location.

6. The distributed network latency measurement method according to claim 5, characterized in that: After receiving the key factor extraction instruction, the influencing factors that are highly correlated with the network status are initially screened out; principal component analysis is applied to reduce the dimensionality of the initially screened influencing factors to obtain the corresponding key influencing factors. The sparse regression model is used to screen key influencing factors by employing minimum absolute contraction and LASSO selection operator regression, and the screened key influencing factors are obtained.

7. The distributed network latency measurement method according to claim 6, characterized in that: Calculate the variance inflation factor (VIF) of the key influencing factors after screening. Key influencing factors with VIF values ​​exceeding the preset threshold are eliminated or merged. The key factors after screening are used as target features. After labeling network anomaly events with target features, automatic response rules are defined based on the event type and attributes of the network anomaly events. When a network anomaly event occurs, a response plan is automatically generated according to the event type and automatic response rules.

8. The distributed network delay measurement method according to claim 7, characterized in that: Each network device in the network is treated as an independent node in the flowchart. Network quality data of each independent node is measured and collected. Real-time network quality data is used as input, and an anomaly detection model is used to detect anomalies and obtain anomaly detection data. Each node is assigned network quality data, environmental condition data, and equipment operating status data corresponding to the stage. Directed edges connect each node to represent the network quality change status. Heatmap elements are overlaid on each node based on the flowchart, and color gradients are used to represent the degree of network quality.

9. The distributed network latency measurement method according to claim 8, characterized in that: By binding network quality data to a time series, the system automatically identifies and marks key time points and stages of network quality data growth or decline, and automatically generates network quality analysis reports based on changes in network quality. The system identifies and classifies current network quality anomalies, defines different types of network measurement scenarios, and builds a rule-based automated response engine. Using current device operating data and environmental condition data as feedback, it uses a genetic algorithm to optimize the currently matched measurement scheme and quickly generate the optimal measurement scheme for specific anomaly scenarios.

10. A distributed network delay measurement system, characterized in that: include, The abnormal event detection unit collects network device operation scenario data and performs real-time detection of network abnormal events. If the network abnormal event meets the preset rule conditions, it generates the corresponding event and issues an alarm command. The event topic allocation unit assigns corresponding topics to network anomaly events, makes predictions and generates anomaly reports. It generates anomaly density of anomaly events from network anomaly event data. When the anomaly density exceeds expectations, it sends a key factor extraction instruction to the outside. The factor identification unit selects key influencing factors from the correlation data, uses a sparse regression model and variance inflation factor to screen out target features, and generates corresponding response schemes by matching the target features with network anomaly events. The network quality detection unit performs anomaly detection on the network quality. If anomalies are found, it visualizes the changes in network data of each network device. The measurement scheme generation unit generates a network quality analysis report from visualized network quality data, and then identifies the current network measurement scenario, including abnormal network equipment operation, abnormal environmental conditions, and abnormal network security. When the current network measurement scenario meets the preset rule conditions, the corresponding measurement plan is output and optimized, and the measurement plan is executed to complete the current latency measurement process; In this process, the network quality of each node is mapped to the corresponding color intensity. The contrast in the middle area is enhanced through a smooth transition, where... ； in: The mapped color intensity, ranging from [0, ... ], , This is a network quality metric, normalized to [0, 1]. The slope parameter controls the steepness of the sigmoid function. The midpoint parameter determines the center position of the Sigmoid function.

Images