A method and system for evaluating the time synchronization accuracy of a network time server

Abstract

This invention provides a method and system for evaluating the timing accuracy of a network time server, relating to the field of computer network technology. The method includes: Step 1, real-time collection of operational data of the network time server in a dynamic network environment, including core indicator data and auxiliary indicator data, and normalization of the core indicator data to generate a standardized core indicator dataset; Step 2, dynamic integration of the standardized core indicator dataset based on the fluctuation characteristics and correlations of each core indicator to generate a comprehensive core indicator representation value; Step 3, dynamic compensation and correction of the comprehensive core indicator representation value in conjunction with auxiliary indicator data to generate a comprehensive timing accuracy evaluation value. This invention achieves comprehensive, accurate, and dynamic evaluation and anomaly diagnosis of the timing accuracy of a network time server, effectively improving the quality of network time services and operational efficiency.

Description

Technical Field

[0001] This invention relates to the field of computer network technology, and in particular to a method and system for evaluating the time synchronization accuracy of a network time server. Background Technology

[0002] The time synchronization accuracy of a Network Time Server (NTS) is a core element for ensuring the collaborative operation of enterprise distributed systems. Traditional evaluation methods rely on a single indicator (such as average time deviation) and calculate errors by statistically analyzing the round-trip delay of messages from time synchronization protocols (such as NTP and PTP). However, these methods do not incorporate the dynamic geometric characteristics of the network topology, leading to evaluation failures in scenarios with sudden traffic surges (such as DDoS attacks and large-scale data backups).

[0003] Specifically, in dynamic network environments, existing technologies cannot combine the physical location and topological connections of key nodes (core switches, border routers) to quantify the spatial propagation impact of network congestion on time synchronization accuracy.

[0004] For example, when a transmission path becomes congested due to an attack, traditional methods cannot characterize the congestion propagation effect through geometric features such as changes in distance between nodes and symmetric attenuation, nor can they locate abnormal core nodes, resulting in assessment results deviating from the actual scenario and delaying fault repair. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for evaluating the time synchronization accuracy of a network time server, which quantifies the impact of congestion propagation on time synchronization accuracy in real time.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0007] A first aspect includes a method for evaluating the time synchronization accuracy of a network time server, the method comprising:

[0008] Step 1: Collect real-time operational data of the network time server in a dynamic network environment, including core indicator data and auxiliary indicator data, and normalize the core indicator data to generate a standardized core indicator dataset.

[0009] Step 2: Based on the fluctuation characteristics and correlations of each core indicator, dynamically integrate the standardized core indicator dataset to generate a comprehensive core indicator representation value.

[0010] Step 3: Combine auxiliary indicator data to dynamically compensate and correct the comprehensive core indicator values ​​to generate a comprehensive time synchronization accuracy evaluation value.

[0011] Step 4: Based on the four target detection points in the auxiliary index data, obtain the vertex coordinates of the formed quadrilateral region, and calculate the side length ratio change rate, diagonal angle offset and symmetry attenuation coefficient of the quadrilateral according to the real-time changes of the quadrilateral vertex coordinates, and generate a dynamic geometric correction factor.

[0012] Step 5: The dynamic geometric correction factor is fused with the comprehensive time synchronization accuracy evaluation value. The confidence interval of the evaluation value is adjusted by the coverage shrinkage rate of the quadrilateral and the vertex offset direction to generate the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range.

[0013] Step 6: Based on the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range, combined with the preset congestion state classification threshold, output the network time server's time synchronization accuracy stability level, abnormal event warning information, and network topology abnormal area location report based on quadrilateral vertex offset direction identification in the current dynamic network environment.

[0014] Secondly, a network time server time synchronization accuracy evaluation system includes:

[0015] The data acquisition module is used to collect core and auxiliary indicator data from the network time server in a dynamic network environment in real time.

[0016] The data processing module is used to normalize the core indicator data to generate a standardized core indicator dataset, and dynamically integrate them according to the fluctuation characteristics and correlation of each core indicator to generate a comprehensive core indicator representation value.

[0017] The dynamic compensation and correction module is used to combine the network delay distribution characteristics and clock frequency offset in the auxiliary indicator data to dynamically compensate and correct the comprehensive core indicator characterization value, and generate a comprehensive time synchronization accuracy evaluation value.

[0018] The dynamic geometric correction module is used to calculate in real time the side length ratio change rate, diagonal angle offset and symmetry attenuation coefficient of the quadrilateral vertex coordinates based on the position information of the four target detection points in the auxiliary index data, and generate dynamic geometric correction factors.

[0019] The evaluation and optimization module is used to integrate the dynamic geometric correction factor with the comprehensive time synchronization accuracy evaluation value. It adjusts the confidence interval of the evaluation value by adjusting the coverage shrinkage rate of the quadrilateral and the vertex offset direction, and generates the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range.

[0020] The report generation module is used to output a report on the stability level of timing accuracy, early warning information of abnormal events, and the location of network topology abnormal areas based on quadrilateral vertex offset direction identification, based on the optimized comprehensive timing accuracy evaluation value and dynamic confidence range, combined with the preset congestion state classification threshold.

[0021] Thirdly, a computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0022] The above-described solution of the present invention has at least the following beneficial effects:

[0023] Real-time collection of core and auxiliary indicator data is performed, and the core indicator data is normalized to ensure comprehensive and standardized data acquisition. Core indicator data reflects key factors in time synchronization accuracy, while auxiliary indicator data provides information such as network environment. By analyzing the fluctuation characteristics and correlations of core indicators and dynamically integrating them, a comprehensive core indicator representation value is generated. This allows for the uncovering of hidden relationships between core indicators, reflecting the overall state of core elements related to time synchronization accuracy and avoiding the limitations of single indicator analysis. The comprehensive core indicator representation value is dynamically compensated and corrected by combining auxiliary indicator data, taking into account the impact of factors such as network environment on time synchronization accuracy, making the comprehensive time synchronization accuracy evaluation value closer to actual conditions.

[0024] A quadrilateral is constructed based on four target detection points from auxiliary indicator data. Relevant geometric parameters are calculated to generate a dynamic geometric correction factor, introducing a new dimension to time synchronization accuracy evaluation from a geometric perspective. This captures the impact of network topology changes on time synchronization accuracy, enhancing the comprehensiveness of the evaluation method. The dynamic geometric correction factor is integrated with the comprehensive time synchronization accuracy evaluation value, adjusting the confidence interval of the evaluation value to generate optimized evaluation values ​​and dynamic confidence ranges. This further optimizes the evaluation results, providing not only numerical results but also a confidence range, offering richer information for decision-making. Based on the optimized results, combined with preset thresholds, time synchronization accuracy stability levels, abnormal event warnings, and network topology anomaly area location reports are output. This transforms the evaluation results into specific and actionable information, helping operations and maintenance personnel quickly understand the operational status of the network time server, promptly detect and handle anomalies, ensure time synchronization accuracy, and improve network service quality. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating a method for evaluating the time synchronization accuracy of a network time server, as provided in an embodiment of the present invention.

[0026] Figure 2 This is a schematic diagram of a network time server time synchronization accuracy evaluation system provided by an embodiment of the present invention. Detailed Implementation

[0027] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0028] like Figure 1 As shown, an embodiment of the present invention proposes a method for evaluating the time synchronization accuracy of a network time server, the method comprising the following steps:

[0029] Step 1: Collect real-time operational data of the network time server in a dynamic network environment, including core indicator data and auxiliary indicator data, and normalize the core indicator data to generate a standardized core indicator dataset.

[0030] Step 2: Based on the fluctuation characteristics and correlations of each core indicator, dynamically integrate the standardized core indicator dataset to generate a comprehensive core indicator representation value.

[0031] Step 3: Combine auxiliary indicator data to dynamically compensate and correct the comprehensive core indicator values ​​to generate a comprehensive time synchronization accuracy evaluation value.

[0032] Step 4: Based on the four target detection points in the auxiliary index data, obtain the vertex coordinates of the formed quadrilateral region, and calculate the side length ratio change rate, diagonal angle offset and symmetry attenuation coefficient of the quadrilateral according to the real-time changes of the quadrilateral vertex coordinates, and generate a dynamic geometric correction factor.

[0033] Step 5: The dynamic geometric correction factor is fused with the comprehensive time synchronization accuracy evaluation value. The confidence interval of the evaluation value is adjusted by the coverage shrinkage rate of the quadrilateral and the vertex offset direction to generate the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range.

[0034] Step 6: Based on the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range, combined with the preset congestion state classification threshold, output the network time server's time synchronization accuracy stability level, abnormal event warning information, and network topology abnormal area location report based on quadrilateral vertex offset direction identification in the current dynamic network environment.

[0035] In this embodiment of the invention, by collecting core indicators (time deviation, jitter, synchronization cycle stability) and auxiliary indicators (network latency distribution, clock frequency offset) in real time, and combining them with a dynamic compensation and correction mechanism, the impact of short-term network latency fluctuations and long-term clock frequency drift on timing accuracy can be effectively distinguished.

[0036] A quadrilateral dynamic geometric model constructed based on four target detection points quantifies the spatial propagation effect of network congestion on the topological path by using the side length ratio change rate, diagonal angle offset, and symmetry attenuation coefficient. For the first time, it realizes real-time correlation analysis between time synchronization accuracy anomalies and network topological geometric features.

[0037] By integrating dynamic geometric correction factors with time synchronization accuracy evaluation values, and combining the coverage shrinkage rate of quadrilaterals with vertex offset directions, the confidence interval of evaluation values ​​is dynamically adjusted to solve the false alarm and false negative problems of traditional static threshold methods in sudden traffic scenarios, thus improving the sensitivity of anomaly detection. Based on the coupling relationship between the quadrilateral vertex offset direction and the network topology path, the core switch nodes or border router nodes that cause time synchronization accuracy degradation are located, and a visual report containing the anomaly propagation path is generated, improving fault repair efficiency by more than 50%. Through sliding window smoothing algorithms, dynamic baseline updates, and multi-timescale feature fusion technology, real-time responses to network traffic mutations (such as DDoS attacks and large-scale data backups) are achieved, ensuring that the time synchronization accuracy evaluation is dynamically corrected within 5 seconds, meeting the high real-time operation and maintenance requirements.

[0038] In a preferred embodiment of the present invention, step 1 above involves real-time acquisition of operational data of the network time server in a dynamic network environment, including core indicator data and auxiliary indicator data, and normalization of the core indicator data to generate a standardized core indicator dataset. The core indicator data includes time deviation sequences, time jitter parameters, synchronization period stability coefficients, and maximum delay error values. The auxiliary indicator data includes network delay distribution characteristics, clock frequency offset, and location information of four preset target detection points. The four target detection points are respectively located at the core switch node, border router node, critical service server node, and the node where the time server is located in the network topology, and may include:

[0039] In this embodiment of the invention, the analysis process of the core indicator data is as follows:

[0040] Time Deviation Sequence: In practice, high-precision time measurement equipment, such as a high-precision time synchronizer, is used to establish a reliable connection with a high-precision atomic clock. During the operation of the network time server, the time is obtained from the network time server at set precise time intervals (e.g., every millisecond) through a stable network communication protocol (e.g., NTP, PTP). Due to network transmission delays and equipment processing time, the round-trip time is measured multiple times each time the time is obtained, and the average value is calculated to estimate the transmission delay. This delay is subtracted when calculating the difference from the standard time. The obtained time differences are arranged in chronological order to form a time deviation sequence. During data recording, if abnormal data occurs, such as a sudden and large jump in the difference exceeding the normal fluctuation range, interpolation can be used to correct it by referring to adjacent data to ensure the accuracy of the sequence.

[0041] Time jitter parameter: Based on the obtained time deviation sequence, the sequence is first cleaned to remove invalid data points caused by network failures, equipment malfunctions, etc., such as multiple consecutive identical abnormal deviation values. To more accurately reflect the characteristics of time jitter, the data is reasonably segmented according to the fluctuation period and trend of the time deviation sequence. For example, if the sequence has obvious periodic fluctuations, it can be segmented by period. In each segment, the variance or standard deviation is calculated. When calculating the variance, the square of the difference between each data point and the average value of the segment is first calculated, and then the average of these squared values ​​is calculated; the standard deviation is the square root of the variance. By comparing the calculation results with different pre-set level thresholds, the severity of time jitter is classified into levels such as slight jitter, moderate jitter, and severe jitter, thereby assessing the instability of the time deviation. The smaller the value, the more stable the time.

[0042] Synchronization Period Stability Coefficient: Define the synchronization period of the network time server, which can be determined based on the synchronization protocol used (such as the default synchronization period of NTP or a custom period). Within each synchronization period, continuously monitor changes in time deviation. In addition to calculating the difference in time deviation between adjacent synchronization periods, use a sliding window approach to calculate indicators such as the rate of change and fluctuation range of time deviation within the window. Perform statistical analysis on these differences and indicators, calculating not only the mean and standard deviation, but also statistics such as the median and coefficient of variation. For example, the coefficient of variation reflects the relative dispersion of the data and helps describe the stability of time deviation within the synchronization period from different perspectives. Plot a line graph of the stability coefficient over time, and use trend analysis algorithms (such as moving average and exponential smoothing) to predict the stability of future synchronization periods, promptly identifying potential stability issues.

[0043] Maximum delay error value:

[0044] During time synchronization between the network time server and other nodes, high-precision timestamp recording devices (such as hardware-based high-precision timestamp chips) are used to accurately record time when sending time requests and receiving time responses. Due to various uncertainties in network transmission, such as network congestion and routing switching, each synchronization not only records the single latency but also relevant network status information, such as current network bandwidth utilization and packet loss rate. The measurement data is quality-assessed; when abnormally high latency values ​​occur, it is analyzed whether they are caused by transient network failures (such as brief link interruptions) or equipment performance bottlenecks. By comparing the latency distribution of multiple adjacent synchronization cycles, it is determined whether abnormally high values ​​are within the normal fluctuation range. The moving average of the maximum latency error value over a period of time is calculated; when this value exceeds a certain threshold, a network latency warning signal is issued so that timely measures can be taken to optimize network performance.

[0045] The process of analyzing auxiliary indicator data:

[0046] Network delay distribution characteristics: Select the appropriate probe packet type based on network type (e.g., Ethernet, wireless network) and measurement requirements (e.g., link delay measurement, end-to-end delay measurement). For example, use smaller ICMP probe packets when measuring link delay; use TCP or UDP probe packets similar to actual business data when performing end-to-end delay measurements. Appropriately set the sending frequency and quantity of probe packets, using stratified sampling to send probe packets at different time periods (e.g., peak and off-peak periods) and on different network paths. Record the round-trip time and the routing path information traversed by the probe packet each time it is sent. Perform statistical analysis on a large amount of delay data, calculating statistics such as mean, median, standard deviation, and percentiles, and plot the probability density function and cumulative distribution function of the delay data to visually display the distribution pattern of network delay. Classify the delay data according to different dimensions (e.g., source and destination node locations, network protocol type, etc.), calculate the delay statistics for each category, and identify patterns and differences in delay distribution. Compare with historical delay data to analyze the changing trends of network delay distribution characteristics and promptly detect abnormal changes in network performance.

[0047] Clock frequency offset: Using high-precision frequency measurement equipment, such as atomic frequency standards or highly stable crystal oscillators as reference sources, the clock frequency of the network time server is measured under stable environmental conditions (including constant temperature, humidity, and electromagnetic environment). Direct counting can be used, where a counter counts the periods of the clock signal and the frequency is calculated based on the counting results; alternatively, a phase comparison method can be used, comparing the phase of the network time server's clock signal with a standard clock signal and calculating the frequency offset based on the change in phase difference. When calculating the frequency offset, the accuracy and stability of the measuring equipment are considered, and error analysis and correction are performed. Measurement results are recorded at regular time intervals (e.g., once per minute), and a curve of frequency offset versus time is plotted. When the frequency offset exceeds a preset threshold, timely frequency calibration measures are taken, such as adjusting the clock's division factor through software or using hardware calibration equipment to ensure the accuracy of the network time server's clock frequency.

[0048] Location information of the four target detection points:

[0049] Obtain detailed network topology documentation, including network device models, connection relationships, IP addresses, and other information. For core switch nodes, border router nodes, and critical business server nodes, use network management software (such as SNMP management tools) to obtain the device's geographical location information (if the device is configured with such information). If the device lacks direct geographical location information, infer its approximate location through the logical structure of the network topology and network traffic analysis. For the node where the time server is located, obtain accurate coordinate information directly from the device's physical installation location. After obtaining the location coordinates, compare them with Geographic Information System (GIS) data, or use GPS positioning equipment to conduct on-site measurements of some nodes for coordinate calibration. As the network topology changes and devices are migrated, establish a dynamic management mechanism to update the location information of the four target detection points in a timely manner, ensuring that this information can be accurately used in network analysis and time synchronization accuracy assessment.

[0050] Normalizing core indicator data aims to transform data of different magnitudes and units to the same range. For each core indicator, the maximum and minimum values ​​in the indicator data are identified, and then each data point is mapped to a specified interval such as [0, 1] or [-1, 1]. For example, for an indicator value... Its normalized value It can be done through formula The calculation yielded, where and These are the minimum and maximum values ​​in the data for this indicator. In this way, core indicator data such as time deviation series, time jitter parameters, synchronization period stability coefficient, and maximum delay error value are all converted into unified dimensionless data, forming a standardized core indicator dataset.

[0051] In a preferred embodiment of the present invention, step 2 above, which involves dynamically integrating the standardized core indicator dataset based on the fluctuation characteristics and correlations of each core indicator to generate a comprehensive core indicator representation value, may include:

[0052] Step 200: Perform sliding window segmentation on the time deviation series, analyze the standard deviation of short-term fluctuation amplitude and the positive and negative slope distribution of trend direction within each window, and generate dynamic fluctuation characteristics of time deviation.

[0053] Step 201: Based on the dynamic fluctuation characteristics of time deviation, extract the peak distribution characteristics of time jitter parameters, filter jitter events through a preset threshold, count the frequency and intensity of jitter events, and generate a quantitative parameter set of time jitter abnormal events;

[0054] Step 202: Based on the time jitter anomaly event quantization parameter set, the synchronization period stability coefficient is converted from the time domain to the frequency domain, the main oscillation frequency and its harmonic component energy ratio are calculated, and the synchronization period stability spectrum quantization index is generated.

[0055] Step 203: Combining the synchronization period stability spectrum quantification index, based on the historical baseline data of the maximum delay error value, calculate the probability of triggering extreme events exceeding the historical threshold, and generate the maximum delay risk level parameter by associating it with the real-time network load status.

[0056] Step 204: Input the dynamic fluctuation characteristics of time deviation, the quantitative parameter set of time jitter abnormal events, the quantitative index of synchronization period stability spectrum and the maximum delay risk level parameter into the dynamic time window, calculate the time-varying correlation coefficient matrix between each parameter, and construct a multi-dimensional dynamic correlation model.

[0057] Step 205: Based on the multi-dimensional dynamic association model, perform feature space projection on the standardized core indicator dataset to extract dynamic feature vectors at different time granularities;

[0058] Step 206: Perform multi-time granularity superposition and normalization on the dynamic feature vector to generate a comprehensive core index characterization value that reflects the coupling effect of the dynamic network environment.

[0059] In this embodiment of the invention, firstly, the size and step size of the sliding window are determined. The window size is like the dimensions of a "small window" for observing time deviations, for example, set to 100 consecutive time points of data; the step size determines the distance the window slides each time, assuming a step size of 10 time points of data. Starting from the beginning of the time deviation sequence, the window is overlaid on the sequence, and the 100 time deviation data points within the window are processed. The standard deviation of these 100 data points is calculated. The standard deviation is calculated by first finding the mean of the data set, then calculating the square of the difference between each data point and the mean, then averaging these squared values, and finally taking the square root of this mean. The standard deviation measures the dispersion of the time deviation data within the window, that is, the magnitude of the fluctuation; the larger the standard deviation, the more drastic the fluctuation.

[0060] Linear fitting is performed on the 100 time deviation data points within the window, essentially finding a straight line that best represents the trend of these data changes. If the slope of this line is positive, it indicates that the time deviation is increasing within this window period; a negative slope indicates that the time deviation is decreasing; and a slope close to 0 means that the time deviation is relatively stable. After completing the analysis of one window, the window is slid to the next position according to the set step size (10 time points), and the process of calculating the standard deviation and analyzing the slope is repeated until the window has slid over the entire time deviation sequence. The distribution of the standard deviation and the positive and negative slope of the trend direction calculated for each window is recorded. This information constitutes the dynamic fluctuation characteristics of the time deviation.

[0061] Step 201: Based on the dynamic fluctuation characteristics of the time deviation obtained in Step 200, perform point-by-point analysis of the time jitter parameters. In this process, an observation window is set, the size of which depends on the stability of the network environment and the required accuracy of the analysis. For example, in a stable enterprise network, a 5-10 minute observation window is selected; while in a wide area network environment with greater fluctuations, it may need to be narrowed down to 1-2 minutes.

[0062] Within each observation window, a sliding comparison method is used to determine peak values. Specifically, the jitter parameter value of the current point is compared with a certain number of its immediate and next-nearest points (e.g., 5 points before and after). If the current point's value is greater than the values ​​of all its neighbors and exceeds a certain proportion (e.g., 1.5 times) of the average value within the window, it is identified as a local peak. The specific timestamps of these peaks are recorded, accurate to the millisecond, along with the corresponding peak sizes, retaining sufficient decimal places (e.g., 6 decimal places) to ensure precision. After analyzing all observation windows, the obtained peak values ​​are arranged in chronological order to form a peak sequence.

[0063] Specifically, determining the preset threshold involves considering multiple factors, including network type, business requirements, and historical data performance. A statistical data method is employed: analyzing time jitter parameter data over a past period (e.g., a week or a month) and calculating its average and standard deviation. The threshold can be set as the average plus three times the standard deviation.

[0064] After determining the preset threshold and peak distribution characteristics, the process of filtering jitter events begins. For each peak point, its value is compared with the preset threshold; if it exceeds the threshold, it is marked as a potential jitter event. To avoid false positives, these potential jitter events need secondary confirmation. Specifically, the jitter parameters are examined over a period of time (e.g., 30 seconds) before and after the peak point. If the jitter parameters show a clear upward and downward trend during this period, and the duration of exceeding the threshold reaches a certain length (e.g., 5 seconds), then it is confirmed as a valid jitter event. For multiple consecutive peak points, if they all exceed the threshold and the time interval is short (e.g., less than 10 seconds), they are merged into a single jitter event to avoid double-counting the same fluctuation process.

[0065] Jitter event frequency calculation: After confirming all jitter events, count the total number of jitter events that occurred within a specific time period (e.g., 1 hour, 1 day). Divide this total number by the length of the time period to obtain the jitter event frequency, which can be in units of times / hour or times / day.

[0066] Jitter event intensity cumulative value calculation: For each confirmed jitter event, its intensity is calculated. The intensity is calculated by integrating the portion of the event where the peak value exceeds a threshold, which is the difference between the peak value and the threshold, and multiplying this difference by the duration of the difference. The intensity values ​​of all jitter events are summed up to obtain the intensity cumulative value.

[0067] The obtained information, including the frequency of jitter events, cumulative intensity, and peak distribution characteristics, is organized to form a complete set of quantification parameters for time jitter anomalies. This parameter set can be organized in tabular form, and its specific contents include:

[0068] Basic statistical information includes total observation time, total number of jitter events, and average frequency.

[0069] Peak distribution characteristics: List the timestamps, amplitude values, and statistical characteristics of all peaks (such as maximum, minimum, average, median, etc.).

[0070] Jitter Event List: Records the start time, end time, duration, peak value, magnitude exceeding the threshold, and calculated intensity value for each jitter event.

[0071] Summary indicators: These include comprehensive indicators such as cumulative intensity values ​​and frequency change trends.

[0072] Step 202: After obtaining the quantization parameter set of time jitter anomalies, the synchronization period stability coefficient data is transformed from the time domain to the frequency domain. The time domain describes the change of the synchronization period stability coefficient over time, while the frequency domain transformation allows for analysis of the data from a frequency perspective, i.e., decomposing a sound into notes of different frequencies, and decomposing the synchronization period stability coefficient into components of different frequencies. In the transformed frequency domain data, the frequency with the most concentrated energy is found; this frequency is the dominant oscillation frequency, representing the most significant frequency component in the change of the synchronization period stability coefficient. Then, the energy proportion of each of the other harmonic components (i.e., frequency components other than the dominant oscillation frequency) is calculated. Recording the dominant oscillation frequency and the energy proportion of each harmonic component generates the spectral quantization index of synchronization period stability.

[0073] Step 203: Referring to the synchronization period stability spectrum quantization index obtained in Step 202, retrieve the historical baseline data of the maximum latency error value. The historical baseline data is the range and statistical information of the maximum latency error value collected and recorded over a long period under normal and stable network operation. Compare the current maximum latency error value data with the threshold in the historical baseline data to count the number of times the maximum latency error value exceeds the historical threshold within a certain time period. Divide the number of times it exceeds the threshold by the total number of counts to obtain the probability of an extreme event triggering the historical threshold. Simultaneously, obtain real-time network load status information, such as network bandwidth usage and the amount of data traffic in the network. Based on the extreme event trigger probability and real-time network load status, and according to pre-set risk level classification rules, divide the maximum latency risk into different levels, such as low risk, medium risk, and high risk, thereby generating a maximum latency risk level parameter.

[0074] Step 204: Set a dynamic time window. The size of this window can be adjusted according to actual needs; it acts like an "observation box" that moves over time. Place the dynamic fluctuation characteristics of time deviation, the quantitative parameter set of time jitter anomalies, the quantitative index of synchronization cycle stability spectrum, and the maximum delay risk level parameter obtained in steps 200-203 into this dynamic time window. Within the window, calculate the time-varying correlation coefficient between any two parameters. The correlation coefficient measures the degree of association between two parameters, ranging from -1 to 1. If the correlation coefficient is close to 1, it indicates that the changing trends of the two parameters are highly consistent; close to -1, it indicates that their changing trends are opposite; close to 0, it means that there is almost no association between the two parameters. Calculate the correlation coefficient between each pair of parameters at different time points, and arrange these correlation coefficients into a matrix. This matrix shows the mutual correlation of each parameter at different time points. Based on this time-varying correlation coefficient matrix, a multi-dimensional dynamic correlation model is constructed, which presents the complex dynamic relationships between the core indicators.

[0075] Step 205: Using the multi-dimensional dynamic association model constructed in Step 204, the standardized core indicator dataset is treated as a whole. Projecting this dataset into its feature space using the model is like "mapping" the dataset from one space to another. In the new space, the data features will be presented in different forms. Based on different time granularity requirements, such as short time (e.g., minutes), medium time (e.g., hours), and long time (e.g., days), corresponding dynamic features are extracted from the projected results. These dynamic features are key information about the data at different time scales. Organizing the dynamic features extracted at each time granularity into vector form yields the dynamic feature vectors for different time granularities.

[0076] Step 206 involves superimposing the dynamic feature vectors obtained in step 205 at different time granularities. This is like layering information from different time periods together, allowing them to merge and form a comprehensive vector containing information from multiple time granularities. Since the numerical values ​​and ranges of the dynamic feature vectors at different time granularities may differ, this comprehensive vector needs to be normalized. Normalization involves adjusting the values ​​in the vector to a fixed range, such as between 0 and 1, according to certain rules. Through normalization, the dimensional differences between feature vectors at different time granularities are eliminated, making them comparable. After multi-time granularity superposition and normalization, the final result is a comprehensive core indicator characterization value reflecting the coupling effect of the dynamic network environment. This value integrates the dynamic changes of each core indicator at different time granularities and the interaction relationships between them.

[0077] Meticulous monitoring of time deviation fluctuations and trends over short periods provides a clearer understanding of its dynamic characteristics, facilitating the timely detection of abnormal changes. Transforming abstract time jitter into concrete, quantifiable parameters allows for precise location of jitter anomalies, facilitating analysis of its impact on timing accuracy and providing robust data support for evaluating and optimizing timing precision. Analyzing synchronization period stability from a frequency domain perspective reveals potential periodic interference factors, enabling a more comprehensive and in-depth assessment of the synchronization period's influence on timing accuracy and helping to identify key frequency components affecting synchronization period stability.

[0078] By combining historical data and real-time network status, the system proactively predicts the risk of maximum latency errors, enabling operations and maintenance personnel to take timely measures to address potential severe latency issues and ensure the stability of time synchronization accuracy. The constructed multi-dimensional dynamic correlation model demonstrates the interrelationships and dynamic changes among various core indicators, breaking the limitations of isolated indicator analysis and helping to understand the comprehensive impact mechanism of each indicator on time synchronization accuracy, providing a theoretical basis for accurate evaluation. Extracting dynamic feature vectors at different time granularities meets the time scale requirements of various analysis scenarios, allowing for the acquisition of relevant key information whether focusing on short-term changes or long-term trends, thus improving the flexibility of evaluation.

[0079] In a preferred embodiment of the present invention, step 3 above, which combines auxiliary indicator data to dynamically compensate and correct the comprehensive core indicator representation value to generate a comprehensive time synchronization accuracy evaluation value, may include:

[0080] Step 300: Based on the intensity and duration of sudden fluctuations in the network delay distribution characteristics, set a dynamic abnormal jump threshold and identify abnormal jump intervals in the time deviation sequence that are ≥ the dynamic abnormal jump threshold.

[0081] Step 301: The sliding window smoothing algorithm is used to segment and compensate the abnormal jump interval, and the compensated time deviation correction sequence is calculated.

[0082] Step 302: Based on the compensated time deviation correction sequence, the cumulative drift error of the synchronization period stability coefficient is reverse-calibrated to generate the drift error compensation coefficient.

[0083] Step 303: The time deviation correction sequence and drift error compensation coefficient are superimposed on the comprehensive core index characterization value to generate the time synchronization accuracy evaluation value after preliminary compensation, and the real-time change rate of network delay distribution characteristics is calculated based on the real-time fluctuation amplitude of the time synchronization accuracy evaluation value.

[0084] Step 304: Dynamically adjust the window length and compensation intensity of the sliding window smoothing algorithm according to the real-time change rate of the network delay distribution characteristics, and generate a comprehensive timing accuracy evaluation value that includes dynamic network delay correction and clock frequency offset suppression.

[0085] In this embodiment of the invention, when analyzing network latency distribution characteristics, the first step is to determine the time range for data statistics; the choice of this range is crucial. If the time range is too short, it may be impossible to capture the complete network fluctuation pattern; if the time range is too long, it may contain too much irrelevant interference information. Taking the past 10 minutes as an example, within these 10 minutes, the network monitoring device continuously records each instance of network latency data. This data is like a string of beads, each bead representing the network latency value at a specific point in time. When calculating the mean and variance of the network latency data during this period, the mean is like the average weight of these beads, reflecting the general level of network latency. The variance reflects the degree of difference in the weight of these beads; the larger the variance, the more drastic the fluctuation in network latency. Calculating the mean and variance is like weighing these beads on a scale and then analyzing their weight distribution.

[0086] Observing fluctuations in network latency data to determine if there are sudden spikes requires the same level of sensitivity as observing weather changes. For example, in an otherwise calm network environment, if at a certain moment the network latency suddenly rises from an average of 50ms to 200ms, and this high latency persists for 10 seconds, it's like a clear sky suddenly becoming overcast and experiencing heavy rain—this obvious change constitutes a sudden spike.

[0087] When setting dynamic anomaly thresholds based on the intensity and duration of sudden fluctuations, historical data should be referenced. For example, long-term observation might show that under normal circumstances, network latency increases generally do not exceed 30ms and last no more than 2 seconds. If a sudden fluctuation occurs with an intensity exceeding 50ms and a duration exceeding 3 seconds, to accurately identify the anomaly while avoiding oversensitivity, the threshold would be set to a value slightly lower than the intensity of the sudden fluctuation, such as 45ms. This threshold acts like a warning line; when data changes in the time deviation sequence cross this line, an anomaly is considered likely.

[0088] Identifying anomalous transition intervals in a time deviation series requires examining each data point sequentially, much like searching for special symbols in a code. When a data point's value is found to have a difference (i.e., change) greater than or equal to a set dynamic anomalous transition threshold, it is marked as such, equivalent to finding the first special symbol in the code. This process continues until a point with a change less than the threshold is found; the interval between these two points is the anomalous transition interval. If there are multiple such intervals in the time deviation series, they must be identified one by one, like a detective searching for clues.

[0089] Step 301: After identifying the anomalous jump interval, selecting an appropriately sized sliding window is crucial. The choice of window size should consider both the length of the anomalous jump interval and the data fluctuations. For example, selecting a window size that includes data from 10 time points acts like a movable magnifying glass, allowing observation of data details within the anomalous jump interval. The sliding window is placed at the beginning of the anomalous jump interval, and the data within the window is processed. The average value of the data within the window is calculated; this process is like mixing the 10 data points in a large bowl to obtain a value representing the average level of these 10 data points. This average value is then used to replace the original value of each data point within the window. This is equivalent to replacing the irregularly shaped beads with beads of the same size, achieving data smoothing. The window is slid forward in steps (e.g., moving one time point at a time), repeating the above operations of calculating the average and replacing data with each slide. This process is like slowly moving a magnifying glass across the anomalous jump interval, smoothing the data little by little. Once the sliding window covers the entire range of abnormal jumps, the previously volatile data becomes relatively smooth.

[0090] For longer anomalous jump intervals, they can be divided into several smaller segments for more precise processing. For example, if an anomalous jump interval contains data from 50 time points, we can imagine it as a long rope and divide it into 5 equal segments, each with 10 time points. Then, we can perform a sliding window smoothing operation on each segment in turn, just like combing the long rope segment by segment, and finally obtain the compensated time deviation correction sequence.

[0091] Step 302: When analyzing the synchronization cycle stability coefficient based on the compensated time deviation correction sequence, we need to treat each synchronization cycle as an independent micro-world. Within each micro-world, we observe the changes in the time deviation correction sequence and calculate the cumulative change in time deviation within each synchronization cycle. This is analogous to recording the total change in the weight of the beads in each micro-world. We compare these cumulative changes with the synchronization cycle stability coefficient to determine if the coefficient has drifted. If the synchronization cycle stability coefficient continuously increases or decreases over multiple cycles, inconsistent with the trend of the time deviation correction sequence, it indicates that it has drifted, like a train deviating from its track. When calculating the cumulative drift error, we need to find the sum of the differences between the actual and theoretical values ​​of the synchronization cycle stability coefficient. The theoretical value here is a reasonable value inferred from the time deviation correction sequence, much like the normal weight inferred from the weight change pattern of the beads. By calculating the sum of the differences between the actual and theoretical values, we can determine how much the synchronization cycle stability coefficient has deviated.

[0092] Then, based on the magnitude and direction of the cumulative drift error, a drift error compensation coefficient is generated. If the cumulative drift error is positive, it indicates that the synchronization period stability coefficient is too large, just like a train running too fast and deviating from its track. In this case, the drift error compensation coefficient is a negative value, used to reduce the synchronization period stability coefficient and bring it back to the normal track; the opposite is also true.

[0093] Step 303 involves superimposing the compensated time deviation correction sequence and drift error compensation coefficient into the comprehensive core indicator representation value. This is a data integration process. For the time deviation-related parts of the comprehensive core indicator representation value, the time deviation correction sequence is used for replacement, much like replacing old parts with new ones. For the parts related to the synchronization cycle stability coefficient, adjustments are made based on the drift error compensation coefficient, much like adjusting machine parameters, thus generating a preliminary compensated time synchronization accuracy evaluation value. When calculating the fluctuation range of the preliminary compensated time synchronization accuracy evaluation value over a short period (e.g., the past minute), this can be achieved by calculating the difference between the maximum and minimum values. This is analogous to measuring the undulation of ocean waves; the difference between the maximum and minimum values ​​is like the height difference between the highest and lowest points of the waves, while the variance more comprehensively reflects the intensity of the wave fluctuations. Based on this fluctuation range, combined with historical and current data on network delay distribution characteristics, the real-time rate of change of the network delay distribution characteristics is calculated. If the fluctuation range of the evaluation value increases and the network latency data also shows that the latency value is increasing, like the waves becoming more and more turbulent, then the real-time change rate is positive, indicating that the network latency distribution characteristics are deteriorating; conversely, if the fluctuation range of the evaluation value decreases and the real-time change rate is negative, then the network latency distribution characteristics are improving, like the waves gradually calming down.

[0094] Step 304: Based on the real-time rate of change of the calculated network latency distribution characteristics, dynamically adjust the window length and compensation intensity of the sliding window smoothing algorithm. If the real-time rate of change is large, it indicates that the network latency fluctuates violently, like waves in a storm. In this case, increase the length of the sliding window (e.g., from 10 time points to 20 time points) to better smooth the data, like using a larger net to catch the waves; at the same time, increase the compensation intensity, that is, replace the original data with the average value more significantly, to quickly eliminate the impact of abnormal fluctuations and allow the data to return to calm as soon as possible.

[0095] Conversely, if the real-time rate of change is small, it indicates that the network latency is relatively stable, like a calm lake. In this case, reducing the length of the sliding window (e.g., from 10 time points to 5 time points) increases the sensitivity to data changes, like using a finer net to catch small fish; reducing the compensation intensity avoids over-smoothing and losing the true information about data changes, preventing the filtering out of small fish as well.

[0096] After dynamically adjusting the sliding window smoothing algorithm, the time deviation correction sequence is processed again, combined with the analysis of clock frequency offset. The clock frequency offset is like the rotational speed deviation of a small gear; based on its magnitude, the time synchronization accuracy evaluation value is fine-tuned accordingly to compensate for the error caused by the clock frequency offset. Finally, a comprehensive time synchronization accuracy evaluation value is generated, incorporating dynamic network latency correction and clock frequency offset suppression, resulting in a more accurate reflection of the network time server's time synchronization accuracy.

[0097] By setting a dynamic abnormal jump threshold based on network latency distribution characteristics, the system can flexibly adapt to fluctuations under different network environments, accurately identify abnormal jump intervals in the time deviation sequence, provide precise targets for subsequent corrections, avoid misjudging normal fluctuations as abnormalities, and improve the accuracy of the assessment. A sliding window smoothing algorithm is used to compensate for abnormal jump intervals in segments, effectively eliminating the impact of abnormal jumps on the time deviation sequence, making the time deviation data smoother and more stable, reducing data noise interference, and laying the foundation for more accurate subsequent assessments of time synchronization accuracy. Reverse calibration of the cumulative drift error of the synchronization period stability coefficient can correct deviations in the synchronization period stability coefficient caused by various factors, ensuring that it truly reflects the synchronization period stability of the network time server, and improving the reliability of this indicator in time synchronization accuracy assessment. The corrected time deviation sequence and compensation coefficient are superimposed on the comprehensive core indicator representation value to generate a preliminary compensated time synchronization accuracy evaluation value, and the real-time change rate of the network latency distribution characteristics is calculated. This achieves preliminary correction of time synchronization accuracy and real-time perception of network latency changes, providing a basis for further optimization and assessment. The sliding window smoothing algorithm is dynamically adjusted based on the real-time rate of change of network delay distribution characteristics, enabling the algorithm to adapt to changes in the network environment. It can effectively correct time deviations under different network conditions, while suppressing the influence of clock frequency offset. Ultimately, it generates a more accurate and reliable comprehensive time synchronization accuracy evaluation value, improving the adaptability and accuracy of the time synchronization accuracy evaluation method in dynamic network environments.

[0098] In a preferred embodiment of the present invention, step 4 above, which involves obtaining the vertex coordinates of the formed quadrilateral region based on the four target detection points in the auxiliary index data, and calculating the side length ratio change rate, diagonal angle offset, and symmetry attenuation coefficient of the quadrilateral based on the real-time changes in the vertex coordinates, and generating a dynamic geometric correction factor, may include:

[0099] Step 400: Initialize the historical baseline data as the average coordinates of the quadrilateral vertices under no-congestion network conditions, and dynamically update the historical baseline data according to the current network traffic status to generate adaptive baseline reference values.

[0100] Step 401: Calculate the difference rate between the real-time side length and the adaptive baseline reference value to obtain the side length ratio change rate;

[0101] Step 402: Based on the side length ratio change rate and combined with the packet loss rate surge event of the network topology transmission path, calculate the baseline deviation of the diagonal angle, and dynamically amplify or suppress the baseline deviation according to the real-time delay fluctuation amplitude to generate the diagonal angle offset.

[0102] Step 403: Analyze the decay law of quadrilateral vertex symmetry with network congestion state, and generate symmetry decay coefficient based on the degree of deviation of vertex coordinates from the ideal symmetry position;

[0103] Step 404: Define the ideal symmetrical position as the geometric center symmetrical coordinates of the quadrilateral vertices, and combine the diagonal angle offset to calculate the Euclidean distance between the coordinates of each vertex and the ideal symmetrical position in real time, and generate the symmetry deviation.

[0104] Step 405: Based on historical data of network congestion events, establish a mapping relationship between symmetry deviation and congestion intensity, and generate a symmetry attenuation coefficient;

[0105] Step 406: Normalize the side length ratio change rate, diagonal angle offset and symmetry attenuation coefficient, and dynamically adjust the correction ratio of each parameter based on the real-time transmission priority of the network topology path to obtain the corrected parameters.

[0106] Step 407: The corrected parameters are fused to generate a dynamic geometric correction factor that reflects the propagation path of network topology anomalies.

[0107] In this embodiment of the invention, during the initial stage of network deployment or during periods when the network is stable and congestion-free (such as late nighttime hours during off-peak hours), coordinate data of four target detection points (core switch node, border router node, critical business server node, and the node where the time server is located) are continuously collected. After multiple collections, the average coordinates of each node on the horizontal and vertical axes are calculated to obtain the average coordinates of the four vertices of the quadrilateral, which is used as historical baseline data.

[0108] As the network operates, network traffic status is monitored in real time. When network traffic changes, such as during peak business hours when traffic increases significantly, or when a sudden surge in traffic occurs (e.g., a large-scale data download task starts), the historical baseline data is updated based on the current traffic situation. A weighted average approach can be used, assigning higher weights to recent coordinate data so that the updated baseline data can more quickly reflect the actual network status and generate an adaptive baseline reference value. For example, recent data might have a weight of 0.7, and historical data a weight of 0.3. The current coordinate data and historical baseline data are then weighted and averaged according to this weight to obtain a new baseline reference value.

[0109] Step 401: Based on the real-time coordinates of the four target detection points, calculate the real-time lengths of the four sides of the quadrilateral using the distance formula between two points. For example, for one side of the quadrilateral, the coordinates of its two endpoints are (…). , )and( , If the real-time length of a side is calculated as the square root of the sum of the squares of the differences between the x and y coordinates of the two points, then the length of that side is equal to the sum of the squares of the squares of the differences between the x and y coordinates of the two points. After calculating the real-time lengths of the four sides, the real-time length of each side is compared with the length of the corresponding side in the adaptive baseline reference. When calculating the difference rate, the difference between the real-time length and the baseline length is divided by the baseline length to obtain the percentage change rate of the side length for each side. For example, if the real-time length of a certain side is... The baseline length is Its side length ratio changes at a rate of Through such calculations, we can clearly understand the changes in each edge relative to the baseline state and determine the changes in the network topology along the edge length dimension.

[0110] Step 402: Given the real-time coordinates of the four vertices of the quadrilateral, for example, vertex A ( , ), B ( , ), C( ′, ′), D( ′, For the diagonal AC connecting vertices A and C, calculate its slope using the slope formula. Similarly, calculate the slope of the diagonal BD connecting vertices B and D. Then, the angle between the two diagonals is calculated using the relationship between slope and angle (such as the tangent function). The real-time angles obtained in this way accurately reflect the angular relationship between the diagonals of the quadrilateral.

[0111] The calculated real-time angle Angle with the diagonal of the adaptive baseline reference value Compare them. Calculate the difference between them, that is, the baseline deviation of the included angle. =∣ - | This deviation value measures the degree of change in the angle between the diagonals of the current quadrilateral relative to the baseline state.

[0112] Real-time monitoring of packet loss rate along the network topology transmission path is crucial. A sudden increase in packet loss rate indicates a network transmission anomaly. This suggests increased instability in the network, making changes in the diagonal angle more likely due to network problems. Further analysis of real-time latency fluctuations reveals that large fluctuations indicate more network instability. To more accurately reflect the impact of network topology changes on the quadrilateral's shape angle, the baseline deviation is amplified, for example, by multiplying the deviation value by a coefficient greater than 1. Conversely, small latency fluctuations indicate relative network stability, leading to the suppression of the baseline deviation, for example, by multiplying the deviation value by a coefficient less than 1. This dynamic adjustment of the baseline deviation ultimately generates a diagonal angle offset, which more accurately reflects the impact of network topology changes on the quadrilateral's shape angle.

[0113] Step 403 involves observing and analyzing numerous network congestion events to record the changes in the symmetry of the quadrilateral vertices under different congestion levels. For example, when network congestion is low, the quadrilateral's shape is relatively regular, and its vertex symmetry is good; as congestion increases, the quadrilateral's shape may become distorted, and its vertex symmetry gradually deteriorates. Through repeated observations and summaries, the decay law of quadrilateral vertex symmetry with network congestion state is obtained.

[0114] Calculate the degree of deviation of the vertex coordinates from the ideal symmetrical position:

[0115] Based on the real-time coordinates of the four target detection points, the ideal symmetrical position is determined as the geometric center symmetry coordinates of the quadrilateral's vertices. The geometric center coordinates of the quadrilateral's four vertices are calculated by taking the average of the four vertex x-coordinates as the center x-coordinate. =4 + + ′+ The average value of the ordinates is used as the central ordinate. =4 + + ′+ For each vertex coordinate, calculate its Euclidean distance to the coordinates symmetric to the geometric center. Let vertex A( , For example, Euclidean distance = + By calculating the Euclidean distance of each vertex, the degree of deviation of each vertex from its ideal symmetrical position can be obtained.

[0116] Based on the calculated deviation and the summarized attenuation rules, a preliminary symmetry attenuation coefficient is generated. The greater the deviation, the worse the symmetry of the quadrilateral, and the larger the corresponding symmetry attenuation coefficient. For example, a relationship between the deviation and the symmetry attenuation coefficient can be defined. If the deviation exceeds a certain threshold, the symmetry attenuation coefficient is increased according to certain rules to reflect the structural changes in network topology caused by congestion.

[0117] Step 404 reiterates that the ideal symmetrical position is the geometric center symmetry coordinates of the quadrilateral's vertices. By calculating the geometric center coordinates of the four vertices, a reference point is obtained, representing the center position of the quadrilateral in its ideal symmetrical state. For each vertex coordinate, its Euclidean distance to the geometric center symmetry coordinates is accurately calculated. This Euclidean distance is obtained by taking the square root of the sum of the squares of the differences between the x and y coordinates. For example, for vertex B... , Euclidean distance = + After calculating the Euclidean distances to all vertices, these distances reflect the deviation of each vertex from its ideal symmetrical position.

[0118] The calculated Euclidean distance is comprehensively considered in conjunction with the diagonal angle offset. This is because the diagonal angle offset also reflects changes in the quadrilateral's shape and affects symmetry assessment. For example, when the diagonal angle offset is large, even if the Euclidean distance of a particular vertex is relatively small, the change in the diagonal angle will still affect the symmetry deviation of that vertex. The Euclidean distance after considering the diagonal angle offset is used as the symmetry deviation of that vertex. The symmetry deviations of multiple vertices collectively reflect the overall symmetry deviation of the quadrilateral. This comprehensive consideration allows for a more complete assessment of the quadrilateral's symmetry.

[0119] Step 405: Collect historical data on network congestion events, including the symmetry deviation of the quadrilateral under different congestion intensities. During actual network operation, record the congestion intensity (e.g., measured by network traffic volume, packet loss rate, etc.) at the time of each congestion event, along with the coordinates of each vertex of the quadrilateral, and then calculate the symmetry deviation. Analyze the collected data and establish a mapping relationship between the symmetry deviation and congestion intensity using statistical methods (e.g., creating a table of correspondences).

[0120] For example, collect a large amount of data related to network congestion events, covering network congestion conditions in different time periods and under different network environments. Classify the intensity of network congestion, for example, into levels such as mild, moderate, and severe congestion.

[0121] When analyzing the relationship between symmetry deviation and congestion intensity, a large amount of quadrilateral vertex coordinate data was processed for each congestion intensity level. Taking the mild congestion level as an example, the symmetry deviation of numerous quadrilaterals was statistically analyzed at this level, recording the degree of deviation of each quadrilateral vertex from its ideal symmetrical position. For example, some quadrilateral vertices had smaller Euclidean distances from their ideal symmetrical positions, while others had larger ones. By processing this data, the range of symmetry deviation under the mild congestion level was obtained, such as the proportion of quadrilaterals whose Euclidean distance falls within a certain interval.

[0122] For the moderate congestion level, the vertex coordinates of a large number of quadrilaterals are also analyzed. By calculating the Euclidean distance between each quadrilateral vertex and its ideal symmetrical position, the distribution of symmetry deviation under this congestion level can be obtained. For example, it was found that in the case of moderate congestion, the symmetry deviation of most quadrilaterals is concentrated within a specific numerical range.

[0123] Under the severe congestion level, a large amount of quadrilateral vertex coordinate data was analyzed in detail. By calculating the Euclidean distance between the vertex and the ideal symmetrical position, and combining factors such as the diagonal angle offset, the range of symmetry deviation under this congestion level was obtained.

[0124] By analyzing a large amount of quadrilateral vertex coordinate data under different congestion levels, the correspondence between each congestion level and the symmetry deviation range can be determined. For example, in mild congestion, the symmetry deviation is small; in moderate congestion, the symmetry deviation is moderate; and in severe congestion, the symmetry deviation is large. Once the symmetry deviation of a quadrilateral is calculated in real time, the previously established correspondence can be used to accurately determine the current network congestion level. For example, if the calculated symmetry deviation falls within the deviation range corresponding to the mild congestion level, it can be determined that the current network is in a mild congestion state.

[0125] Based on the determined congestion intensity, a final symmetry attenuation coefficient can be generated. Since the attenuation patterns of quadrilateral vertex symmetry under different congestion intensities have been summarized—for example, as congestion intensity increases, the symmetry of quadrilateral vertices gradually deteriorates—a symmetry attenuation coefficient that more accurately reflects the impact of the current network congestion state on quadrilateral symmetry can be generated based on the current congestion intensity and these attenuation patterns. This coefficient effectively reflects the degree of change in network topology.

[0126] Step 406 involves normalizing the ratio of side length changes, the diagonal angle offset, and the symmetry attenuation coefficient. Their values ​​are mapped to a uniform interval (e.g., [0, 1]), eliminating differences in units and numerical ranges between different parameters and making them comparable. For example, for the ratio of side length changes... If its original value range is [ , The normalized value is... = - - In this way, the three parameters can be compared on the same scale.

[0127] Real-time transmission priority information of network topology paths is acquired. Different service or data transmission paths may have different priorities; for example, critical service data transmission paths have higher priority, while ordinary data transmission paths have lower priority. The priority information of each path is recorded through the network management system. Based on the priority of each path, different correction ratios are set for three parameters. For higher-priority paths, the corresponding parameters have greater weight during correction and have a greater impact on the final result. For example, for critical service data transmission paths, the correction ratio for the side length ratio change rate is 0.4, the correction ratio for the diagonal angle offset is 0.4, and the correction ratio for the symmetry attenuation coefficient is 0.2; for ordinary data transmission paths, the correction ratios may be 0.2, 0.3, and 0.5, respectively. Through such dynamic adjustment, the corrected side length ratio change rate, diagonal angle offset, and symmetry attenuation coefficient are obtained, making these parameters more accurately reflect the changes in network topology under different priority paths.

[0128] Step 407 involves fusing the corrected side length ratio change rate, diagonal angle offset, and symmetry attenuation coefficient according to certain rules. A weighted summation method is used, assigning different weights to each parameter based on its importance in reflecting the propagation path of network topology anomalies. For example, based on the actual network situation and the degree of influence on topology changes, the side length ratio change rate has a weight of 0.3, the diagonal angle offset has a weight of 0.4, and the symmetry attenuation coefficient has a weight of 0.3.

[0129] Calculate the dynamic geometric correction factor:

[0130] The three parameters are multiplied by their respective weights and then summed, resulting in the dynamic geometric correction factor F = 0.3 × corrected side length ratio change rate + 0.4 × corrected diagonal angle offset + 0.3 × corrected symmetry attenuation coefficient. This weighted summation yields a comprehensive value, which is the dynamic geometric correction factor. It effectively reflects changes in the propagation paths of network topology anomalies, providing a comprehensive geometric correction basis for evaluating the timekeeping accuracy of network time servers, making the evaluation results more accurately reflect the impact of network topology changes on timekeeping accuracy.

[0131] By initializing historical baseline data and dynamically updating it based on network traffic, the baseline reference values ​​can be made to closely match the actual network operating status, ensuring the accuracy and reliability of the assessment. Calculating the side length ratio change rate quantifies network topology changes from a side length perspective, intuitively reflecting changes in network connectivity and helping to detect anomalies related to network transmission path length. Combining packet loss rate and latency fluctuations to adjust the diagonal angle offset comprehensively considers the impact of various network factors on the topology shape angle, enabling more accurate capture of angular changes in the network topology and identifying potential network transmission problems. By analyzing the symmetry of quadrilateral vertices and establishing a mapping relationship with congestion intensity to generate a symmetry attenuation coefficient, the network topology can be assessed from a structural symmetry perspective, effectively reflecting the degree of damage to the topology structure caused by network congestion and providing early warning of network failures. Normalizing and dynamically adjusting parameters eliminates dimensional differences and incorporates transmission priority, allowing each parameter to participate reasonably in the correction, making the assessment results more consistent with actual network conditions. The generated dynamic geometric correction factor integrates multiple parameters that reflect changes in network topology, effectively characterizing the propagation path of network topology anomalies. It provides a strong geometric dimension correction basis for evaluating the timing accuracy of network time servers, improves the accuracy of evaluation results, and helps operations and maintenance personnel quickly locate abnormal areas in network topology.

[0132] In a preferred embodiment of the present invention, step 5 above, which fuses the dynamic geometric correction factor with the comprehensive time synchronization accuracy evaluation value, and adjusts the confidence interval of the evaluation value by the coverage shrinkage rate of the quadrilateral and the vertex offset direction, to generate an optimized comprehensive time synchronization accuracy evaluation value and a dynamic confidence range, may include:

[0133] Step 500: Based on the correlation between the coverage shrinkage rate of the quadrilateral and the sudden change in network traffic, analyze the impact of the dynamic geometric correction factor on the confidence interval boundary expansion of the comprehensive time synchronization accuracy evaluation value, and generate the boundary expansion coefficient.

[0134] Step 501: Based on the boundary expansion coefficient, extract the mapping relationship between the vertex offset direction and the network topology anomaly propagation path, identify the priority of the anomaly propagation direction, and determine the dynamic distribution correction ratio of the comprehensive timing accuracy evaluation value within the confidence interval according to the priority of the anomaly propagation direction.

[0135] Step 502: Based on the boundary expansion coefficient, the confidence interval of the comprehensive time synchronization accuracy evaluation value is initially scaled to obtain the initially scaled confidence interval. Then, combined with the dynamic distribution correction ratio, the initially scaled confidence interval is dynamically calibrated again to generate the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range.

[0136] In this embodiment of the invention, during network operation, a high-precision coordinate monitoring device is used to track the coordinates of the four vertices of a quadrilateral in real time. The vertex coordinates are recorded at fixed time intervals (e.g., every second). After recording the coordinates for a certain period (e.g., 1 minute), the area of ​​the quadrilateral is calculated. Assume that the initial coordinates of the four vertices of the quadrilateral are A(…). , ), B ( , ), C( ′, ′), D( ′, The initial area is obtained by applying the quadrilateral area formula (e.g., by dividing the quadrilateral into triangles to calculate the area). After a period of time (e.g., 1 minute), the vertex coordinates are recorded again and the area is calculated. Shrinkage rate using the formula ×1 calculates the shrinkage rate of the quadrilateral's coverage area.

[0137] Use network traffic monitoring tools (such as traffic sensors, network traffic analysis software, etc.) to monitor network traffic in real time. This includes not only recording the volume of traffic (e.g., the amount of data transmitted per second, in bytes), but also monitoring for sudden traffic spikes. When network traffic increases significantly within a short period (e.g., within 10 seconds) (e.g., an increase exceeding a certain percentage, such as 50% of the original traffic), it is considered a sudden traffic event.

[0138] The calculated quadrilateral coverage shrinkage rate was compared and analyzed with network traffic mutations. By plotting a time series graph of shrinkage rate versus traffic changes, the trend of shrinkage rate during traffic mutations (such as sudden traffic spikes) was observed. For example, if the shrinkage rate also increases when network traffic suddenly increases, and this occurs in multiple traffic mutations, it indicates a positive correlation between the shrinkage rate and network traffic mutations. Conversely, if the shrinkage rate does not change significantly or even decreases when traffic increases, it indicates that the relationship between the two is not significant or that a negative correlation exists.

[0139] The dynamic geometric correction factor includes parameters such as the side length ratio change rate, diagonal angle offset, and symmetry attenuation coefficient. When analyzing its impact on the expansion of the confidence interval boundary of the comprehensive time synchronization accuracy evaluation value, it is essential to first clarify the changes in each parameter. For example, the side length ratio change rate is obtained by comparing the length ratio of the quadrilateral's sides at different times. If the side length ratio change rate increases, it indicates a significant change in the quadrilateral's side length, which may affect the confidence interval boundary of the time synchronization accuracy evaluation value. The diagonal angle offset reflects the difference between the diagonal angle and the angle under ideal symmetry; an increase in the offset also affects the confidence interval boundary. The symmetry attenuation coefficient comprehensively considers the changes in the symmetry of the quadrilateral's vertices; an increase in the attenuation coefficient means a deterioration in symmetry, which also affects the confidence interval boundary.

[0140] Based on the correlation analysis between the quadrilateral coverage shrinkage rate and network traffic mutations, and combined with the parameter changes of the dynamic geometric correction factor, their impact on the expansion of the confidence interval boundary is determined. For example, if the quadrilateral coverage shrinkage rate is positively correlated with network traffic mutations, and the rate of change of the side length ratio in the dynamic geometric correction factor increases, then the confidence interval boundary may expand outward; conversely, if the shrinkage rate is negatively correlated with traffic mutations, and the rate of change of the side length ratio decreases, the confidence interval boundary may shrink inward.

[0141] By comprehensively considering the correlation between the shrinkage rate and network traffic mutations, as well as the changes in the dynamic geometric correction factor parameter, a coefficient reflecting the impact of the dynamic geometric correction factor on the expansion of the confidence interval boundary is determined; this coefficient is called the boundary expansion coefficient. For example, when the correlation between the shrinkage rate and network traffic mutations is high, and the dynamic geometric correction factor parameter changes significantly, the boundary expansion coefficient may be relatively large, such as 1.5; conversely, when the correlation is low and the parameter changes are small, the boundary expansion coefficient may be relatively small, such as 0.5. In this way, the boundary expansion coefficient can accurately reflect the impact of network state changes on the confidence interval boundary.

[0142] Step 501: Continuously monitor the coordinate changes of the quadrilateral vertices, specifically by calculating the difference in coordinates of each vertex at different times. Taking vertex A as an example, assume that at time... The coordinates are ( , ), at time The coordinates are ( , If ), then the difference is ( - , - The offset direction of the vertex is determined by the sign and magnitude of the difference. - >0 and - If the value is greater than 0, then the offset direction of vertex A is to the upper right; if - <0 and - If the value is less than 0, then the offset direction of vertex A is to the lower left, and so on.

[0143] By monitoring changes in data transmission paths within the network, network topology monitoring tools (such as network topology discovery software) are used to obtain the data transmission paths between various nodes in the network. Simultaneously, paths with high packet loss rates are monitored; packet loss rate refers to the ratio of the number of data packets lost within a certain period to the total number of data packets transmitted. When the packet loss rate of a path exceeds a certain threshold (e.g., 5%), that path is identified as an abnormal propagation path. By analyzing changes in data transmission paths and packet loss rates within the network topology, abnormal propagation paths in the network topology are determined.

[0144] The analysis compares the vertex offset direction with abnormal propagation paths in the network topology. For example, when a vertex offsets in a specific direction, it checks if there are any related abnormal propagation paths in the network topology. If the vertex offset direction is consistent with the direction of abnormal traffic increase on a certain path in the network, and that path has a high packet loss rate, then a correlation is considered to exist between them. For example, if vertex A offsets in a certain direction, and at the same time, the traffic on a related path suddenly increases and the packet loss rate exceeds a threshold, it can be determined that there is a mapping relationship between the offset direction of vertex A and the abnormal propagation path of that path.

[0145] Based on the mapping relationship between vertex offset directions and network topology anomaly propagation paths, and considering factors such as path traffic volume, packet loss rate, and impact on time synchronization accuracy, anomaly propagation directions are prioritized. For example, if data transmission on a network topology path corresponding to a vertex offset direction has a significant impact on time synchronization accuracy (e.g., large time deviation variations along the path), and the path has a high packet loss rate and high traffic volume, then this anomaly propagation direction has a higher priority. Conversely, if a path corresponding to a vertex offset direction has a smaller impact on time synchronization accuracy, a lower packet loss rate, and lower traffic volume, then this anomaly propagation direction has a lower priority. This method accurately identifies the priority of different anomaly propagation directions.

[0146] Based on the priority of anomaly propagation directions and combined with the boundary expansion coefficient, the dynamic distribution correction ratio of the comprehensive timing accuracy evaluation value within the confidence interval is determined. For anomaly propagation directions with high priority, their influence ratio on the distribution of the comprehensive timing accuracy evaluation value within the confidence interval is increased accordingly. For example, assuming a correction ratio of 0.6 for high-priority anomaly propagation directions and 0.4 for low-priority anomaly propagation directions. By setting different correction ratios, the distribution of the comprehensive timing accuracy evaluation value within the confidence interval can more accurately reflect the impact of network topology anomaly propagation paths and vertex offset directions, thereby more accurately assessing timing accuracy.

[0147] Step 502: Based on the boundary expansion coefficient, initially scale the confidence interval of the comprehensive time synchronization accuracy evaluation value. If the boundary expansion coefficient is large, it indicates that the network state has changed significantly, and the confidence interval needs to be expanded outward to increase the range of possible evaluation values. For example, the original confidence interval is [ , The boundary expansion factor is 1.5. Therefore, the upper and lower limits are adjusted to [...]. - , + ],in It is determined based on the boundary expansion coefficient and the range of the original confidence interval, for example... =( - × 0.5 (here, 0.5 is an adjustment based on a boundary expansion coefficient of 1.5). Conversely, if the boundary expansion coefficient is small, it indicates that the network state is relatively stable, and the confidence interval may need to be narrowed inward, reducing the range of evaluation values. In this way, the initially scaled confidence interval can initially reflect the impact of network state changes on the range of evaluation values.

[0148] The process of performing secondary dynamic calibration by combining the dynamic distribution correction ratio is as follows:

[0149] Based on the dynamic distribution correction ratio determined in step 501, a second dynamic calibration is performed on the initially scaled confidence interval. For anomaly propagation directions with higher priority, the correction ratio is larger, and the distribution of evaluation values ​​within the confidence interval is adjusted according to this ratio. For example, if an anomaly propagation direction has high priority and its correction ratio is 0.6, the distribution of evaluation values ​​within the confidence interval is redistributed, making the evaluation values ​​more likely to be taken in regions related to that anomaly propagation direction.

[0150] Specifically, the confidence interval is divided into several sub-intervals, and the probability distribution of the evaluation value within these sub-intervals is adjusted according to the correction ratio. If the correction ratio for the sub-interval corresponding to the anomaly propagation direction is 0.6, then a higher weight is given when calculating the probability that the evaluation value falls within that sub-interval, making the evaluation value more likely to take a value within that sub-interval. Through this secondary dynamic calibration, an optimized comprehensive time synchronization accuracy evaluation value and a dynamic confidence range are generated. This allows the evaluation value to more accurately reflect the impact of network topology changes on time synchronization accuracy, and the dynamic confidence range to more reasonably represent the confidence level of the evaluation value, providing more reliable results for the time synchronization accuracy assessment of the network time server.

[0151] By analyzing the correlation between the quadrilateral coverage shrinkage rate and network traffic mutations, as well as the impact of the dynamic geometric correction factor on the confidence interval boundary expansion, a boundary expansion coefficient is generated. This allows the confidence interval of the comprehensive time synchronization accuracy evaluation value to be reasonably adjusted according to the actual network situation, improving the confidence of the evaluation value and more accurately reflecting the impact of network topology changes on time synchronization accuracy, avoiding evaluation value errors caused by network changes. The mapping relationship between vertex offset direction and network topology anomaly propagation path is extracted, the priority of anomaly propagation direction is identified, and the dynamic distribution correction ratio of the comprehensive time synchronization accuracy evaluation value within the confidence interval is determined. This helps to understand the impact mechanism of network topology anomalies on time synchronization accuracy. Adjusting the distribution of the evaluation value within the confidence interval according to the priority of anomaly propagation direction allows the evaluation value to more accurately reflect the impact of network topology anomaly propagation path. The confidence interval is initially scaled, and a second dynamic calibration is performed in conjunction with the dynamic distribution correction ratio to generate an optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range. This allows the comprehensive time synchronization accuracy evaluation value to be dynamically adjusted according to changes in network topology, and the dynamic confidence range can more reasonably represent the credibility of the evaluation value, thereby improving the reliability and effectiveness of time synchronization accuracy assessment, providing more accurate evaluation results for the operation of the network time server, and helping to promptly identify and resolve time synchronization accuracy issues.

[0152] In a preferred embodiment of the present invention, step 6 above, based on the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range, combined with a preset congestion state classification threshold, outputs the network time server's time synchronization accuracy stability level, abnormal event warning information, and a network topology abnormal area location report based on quadrilateral vertex offset direction identification, which may include:

[0153] Step 600: The dynamic confidence range of the optimized comprehensive timing accuracy evaluation value is matched with the preset congestion state classification threshold in multiple dimensions to determine the timing accuracy stability level.

[0154] Step 601: Generate a level confidence parameter based on the degree of deviation between the time synchronization accuracy stability level and the dynamic confidence range;

[0155] Step 602: Based on the level confidence parameter, set the dynamic anomaly event detection threshold, identify anomaly events with evaluation values ​​≥ the dynamic anomaly event detection threshold, and analyze the propagation direction and diffusion rate of the anomaly event on the topological path by combining the spatiotemporal correlation of the quadrilateral vertex offset direction.

[0156] Step 603: Based on the propagation direction and diffusion rate, and combined with the coupling relationship between the vertex offset direction and the network topology path, locate the starting node and relay node of the abnormal propagation path, and associate them with the node performance baseline data.

[0157] Step 604: Based on the node performance baseline data and diffusion rate, match the feature library of historical network congestion events to generate a repair priority list for core switch nodes and border router nodes.

[0158] Step 605: Based on the repair priority list and the level confidence parameter, generate a location report containing node identifiers, anomaly propagation paths, and repair suggestions.

[0159] In this embodiment of the invention, preset congestion state classification thresholds are defined. These thresholds are set based on the network service's requirements for timing accuracy, combined with historical network data and experience. For example, they are divided into different threshold ranges corresponding to four levels: "excellent", "good", "medium", and "poor".

[0160] The optimized comprehensive time synchronization accuracy evaluation value's dynamic confidence range is matched against these thresholds in a multi-dimensional manner. This involves not only comparing whether the evaluation value itself falls within a certain threshold range, but also considering the relationship between the upper and lower limits of the dynamic confidence range and the threshold boundaries. For example, if the lower limit of the dynamic confidence range is higher than the lower threshold limit corresponding to the "Good" level, and the upper limit is lower than the upper threshold limit corresponding to the "Good" level, then the initial determination of the time synchronization accuracy stability level is "Good." If the lower limit of the dynamic confidence range is lower than the lower threshold limit corresponding to the "Poor" level, even if the evaluation value itself falls within other ranges, the level may still be determined as "Poor" according to the rules. Through this multi-dimensional matching, comprehensively considering the evaluation value and its fluctuation range, the final time synchronization accuracy stability level is determined.

[0161] Step 601: Calculate the deviation between the timing accuracy stability level and the dynamic confidence range. For example, if the timing accuracy stability level is determined to be "Good," but the lower limit of the dynamic confidence range is close to the lower threshold of the "Medium" level and the upper limit is close to the upper threshold of the "Excellent" level, it indicates that there is some uncertainty in the level determination. A level reliability parameter is generated by quantifying the degree of deviation. A rule can be set, such as lowering the reliability parameter when the distance between the boundary of the dynamic confidence range and the boundary of the corresponding level threshold is less than a certain percentage (e.g., 10%); the greater the distance, the higher the reliability parameter. Assuming the reliability is represented by a numerical range of 0-1, if the deviation is small, the reliability parameter can be set to 0.8-1; if the deviation is large, the reliability parameter can be set to 0-0.2. The final level reliability parameter reflects the reliability of the current timing accuracy stability level determination.

[0162] Step 602: Set a dynamic anomaly detection threshold based on the level confidence parameter. A higher confidence parameter results in a more stringent threshold; a lower confidence parameter results in a more lenient threshold. For example, when the level confidence parameter is 0.9, the dynamic anomaly detection threshold is set to a value close to the upper limit of the corresponding level threshold; when the level confidence parameter is 0.3, the threshold can be appropriately lowered. In the comprehensive time synchronization accuracy evaluation value, look for cases where the evaluation value is greater than or equal to the dynamic anomaly detection threshold; once found, these are identified as anomalies.

[0163] This study analyzes anomalous events by combining the spatiotemporal correlation of quadrilateral vertex offset directions. It records the offset direction and time point of each quadrilateral vertex when an anomalous event occurs, and analyzes the changes in vertex offset directions over subsequent time periods. If multiple vertices continuously offset in the same direction within a certain time period, and the offset angle gradually increases, it indicates that the anomalous event has a specific propagation direction along the topological path. By calculating the change in distance or angle of vertex offset per unit time, the propagation rate of the anomalous event is assessed.

[0164] Step 603: Based on the analyzed propagation direction and diffusion rate of the abnormal event, and combined with the coupling relationship between the vertex offset direction and the network topology path, node localization is performed. The consistency between the vertex offset direction and the data transmission path in the network is observed. If the offset direction of a vertex is consistent with the abnormal increase in data traffic on a certain network topology path, and this path conforms to the propagation direction of the abnormal event, then the nodes on this path are likely related nodes of the abnormal propagation path. By tracking the starting direction and changes of vertex offset during the propagation of the abnormal event, the starting node of the abnormal propagation path is determined; then, based on the changes in the vertex offset direction and the propagation direction during the diffusion process, the nodes that act as relays on the path are identified. The performance baseline data of these nodes is correlated, including data indicators such as the node's processing capacity, bandwidth utilization, and latency under historical normal operating conditions.

[0165] Step 604 involves combining the performance baseline data of the located abnormal propagation path nodes with the propagation rate of the abnormal event, and then matching this data with a feature database of historical network congestion events. The feature database stores information such as node performance change characteristics and propagation rates for different types of network congestion events in the past. Through comparative analysis, it determines which historical events in the feature database the current abnormal event is similar to, thereby determining the priority of critical nodes such as core switch nodes and border router nodes during the repair process. For example, if the performance baseline data of a node shows a decrease in its processing capacity after an abnormal event occurs, and the abnormal event propagates rapidly, and has a high degree of matching with the event characteristics in the feature database that lead to serious network failures, then the repair priority of this node is high; conversely, the repair priority is low. Finally, a list containing the repair priorities of each critical node is generated.

[0166] Step 605: Based on the repair priority list and the level of reliability parameters, generate a location report. The report should clearly list the identifiers of the abnormal nodes (such as node IP address, device name, etc.), and describe in detail the abnormal propagation path, including the starting node, relay nodes, and data transmission direction. Specific repair recommendations should be given for nodes of different priorities. For nodes with high repair priority, immediate device inspection, restart, or replacement is recommended; for nodes with low repair priority, continuous monitoring or maintenance during periods of low network load is suggested. The report should also indicate the reliability of the time synchronization accuracy stability level, allowing maintenance personnel to fully understand the current status and potential risks of the network time server.

[0167] The timing accuracy stability level is determined through multi-dimensional matching, comprehensively considering evaluation values ​​and their fluctuation range. This makes the level determination more closely reflect the actual network condition, providing operations and maintenance personnel with an intuitive and accurate reference for timing accuracy stability. A level credibility parameter is generated to quantify the reliability of the level determination, allowing operations and maintenance personnel to understand the credibility of the timing accuracy stability level and avoid misjudgments due to inaccurate level determinations. Dynamically setting abnormal event detection thresholds and analyzing the propagation characteristics of abnormal events enables timely and accurate identification of abnormal events and understanding their propagation patterns, facilitating proactive measures to prevent anomalies from escalating. Accurately locating the starting and relay nodes of anomaly propagation paths and correlating them with performance baseline data provides clear targets and data support for quickly troubleshooting and resolving network problems, improving fault location efficiency. A repair priority list is generated, rationally arranging the repair order based on historical event characteristics and current anomaly conditions, ensuring that critical nodes are repaired first, optimizing the network repair process, and reducing the impact time of network failures. Generate detailed location reports, including information such as nodes, paths, and repair suggestions, providing clear troubleshooting guidelines for operations and maintenance personnel. At the same time, indicate the level of trust to assist operations and maintenance personnel in making scientific decisions and ensure the stable operation of the network time server in dynamic network environments.

[0168] like Figure 2 As shown, embodiments of the present invention also provide a network time server timing accuracy evaluation system, comprising:

[0169] The data acquisition module is used to collect core and auxiliary indicator data from the network time server in a dynamic network environment in real time.

[0170] The data processing module is used to normalize the core indicator data to generate a standardized core indicator dataset, and dynamically integrate them according to the fluctuation characteristics and correlation of each core indicator to generate a comprehensive core indicator representation value.

[0171] The dynamic compensation and correction module is used to combine the network delay distribution characteristics and clock frequency offset in the auxiliary indicator data to dynamically compensate and correct the comprehensive core indicator characterization value, and generate a comprehensive time synchronization accuracy evaluation value.

[0172] The dynamic geometric correction module is used to calculate in real time the side length ratio change rate, diagonal angle offset and symmetry attenuation coefficient of the quadrilateral vertex coordinates based on the position information of the four target detection points in the auxiliary index data, and generate dynamic geometric correction factors.

[0173] The evaluation and optimization module is used to integrate the dynamic geometric correction factor with the comprehensive time synchronization accuracy evaluation value. It adjusts the confidence interval of the evaluation value by adjusting the coverage shrinkage rate of the quadrilateral and the vertex offset direction, and generates the optimized comprehensive time synchronization accuracy evaluation value and dynamic confidence range.

[0174] The report generation module is used to output a report on the stability level of timing accuracy, early warning information of abnormal events, and the location of network topology abnormal areas based on quadrilateral vertex offset direction identification, based on the optimized comprehensive timing accuracy evaluation value and dynamic confidence range, combined with the preset congestion state classification threshold.

[0175] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0176] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0177] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for evaluating the time service precision of a network time server, characterized in that, The method comprises: Step 1, collecting the running data of the network time server in the dynamic network environment in real time, including core index data and auxiliary index data, and normalizing the core index data to generate a standardized core index data set; Step 2, according to the fluctuation characteristics and correlation of each core index, dynamically integrating the standardized core index data set to generate a comprehensive core index representation value, including: performing sliding window segmentation on the time deviation sequence, analyzing the standard deviation of the short-term fluctuation amplitude and the positive or negative slope distribution of the trend direction in each window to generate a time deviation dynamic fluctuation feature; based on the time deviation dynamic fluctuation feature, extracting the peak distribution characteristics of the time jitter parameter, screening the jitter events through a preset threshold, and calculating the jitter event frequency and intensity cumulative value to generate a time jitter abnormal event quantization parameter set; according to the time jitter abnormal event quantization parameter set, converting the synchronization period stability coefficient from time domain to frequency domain, calculating the main oscillation frequency and the harmonic component energy proportion, and generating a synchronization period stability spectrum quantization index; combining the synchronization period stability spectrum quantization index, based on the historical baseline data of the maximum delay error value, calculating the extreme event triggering probability exceeding the historical threshold, and generating a maximum delay risk level parameter in association with the real-time network load state; inputting the time deviation dynamic fluctuation feature, the time jitter abnormal event quantization parameter set, the synchronization period stability spectrum quantization index and the maximum delay risk level parameter into a dynamic time window, calculating the time-varying correlation coefficient matrix between each parameter, and constructing a multi-dimensional dynamic correlation model; based on the multi-dimensional dynamic correlation model, projecting the standardized core index data set in the feature space to extract dynamic feature vectors of different time granularities; performing multi-time granularity superposition and normalization processing on the dynamic feature vectors to generate a comprehensive core index representation value reflecting the coupling effect of the dynamic network environment; Step 3, combining the auxiliary index data, dynamically compensating and correcting the comprehensive core index representation value to generate a comprehensive time service accuracy evaluation value, including: setting a dynamic abnormal jump threshold according to the burst fluctuation intensity and duration of the network delay distribution characteristics, and identifying the abnormal jump interval in the time deviation sequence that is greater than or equal to the dynamic abnormal jump threshold; segmenting and compensating the abnormal jump interval by using a sliding window smoothing algorithm to calculate a time deviation correction sequence after compensation; based on the time deviation correction sequence after compensation, inversely calibrating the cumulative drift error of the synchronization period stability coefficient to generate a drift error compensation coefficient; superimposing the time deviation correction sequence and the drift error compensation coefficient into the comprehensive core index representation value to generate a time service accuracy evaluation value after preliminary compensation, and calculating the real-time change rate of the network delay distribution characteristics based on the real-time fluctuation amplitude of the time service accuracy evaluation value; dynamically adjusting the window length and compensation strength of the sliding window smoothing algorithm according to the real-time change rate of the network delay distribution characteristics to generate a comprehensive time service accuracy evaluation value containing dynamic network delay correction and clock frequency offset suppression; Step 4, based on the four target detection points in the auxiliary index data, obtain the coordinates of the vertices of the formed quadrilateral region, and according to the real-time changes of the quadrilateral vertex coordinates, calculate the edge length proportion change rate, the diagonal line angle offset amount and the symmetry attenuation coefficient, generate the dynamic geometric correction factor, the four target detection points are respectively located in the core switch node, the border router node, the key service server node and the node where the time server is located in the network topology; Step 5, fuse the dynamic geometric correction factor with the comprehensive timing accuracy evaluation value, generate the optimized comprehensive timing accuracy evaluation value and dynamic confidence range through the credible interval of the vertex offset direction adjustment evaluation value and the coverage range contraction rate of the quadrilateral; Step 6, based on the optimized comprehensive timing accuracy evaluation value and dynamic confidence range, combined with the preset congestion state grading threshold, output the timing accuracy stability level of the network time server in the current dynamic network environment, the abnormal event early warning information and the network topology abnormal area positioning report based on the quadrilateral vertex offset direction identification.

2. The network time server timekeeping precision evaluation method according to claim 1, characterized by, Based on the four target detection points in the auxiliary index data, obtain the coordinates of the vertices of the formed quadrilateral region, and according to the real-time changes of the quadrilateral vertex coordinates, calculate the edge length proportion change rate, the diagonal line angle offset amount and the symmetry attenuation coefficient, generate the dynamic geometric correction factor, including: Initialize the historical baseline data as the average of the coordinates of the vertices of the quadrilateral in the network non-congestion state, and dynamically update the historical baseline data according to the current network traffic state to generate an adaptive baseline reference value; Calculate the difference rate of the real-time edge length and the adaptive baseline reference value to obtain the edge length proportion change rate; Based on the edge length proportion change rate, combined with the packet loss rate sudden increase event of the network topology transmission path, calculate the baseline deviation of the diagonal line angle, and according to the real-time delay fluctuation amplitude, dynamically amplify or suppress the baseline deviation to generate the diagonal line angle offset amount; Analyze the attenuation law of the symmetry of the quadrilateral vertices with the network congestion state, and generate the symmetry attenuation coefficient according to the deviation degree of the vertex coordinates from the ideal symmetric position; Define the ideal symmetric position as the geometric center symmetric coordinates of the vertices of the quadrilateral, and combine the diagonal line angle offset amount to calculate the Euclidean distance between each vertex coordinate and the ideal symmetric position in real time to generate the symmetry deviation degree; According to the historical data of network congestion events, establish the mapping relationship between the symmetry deviation degree and the congestion intensity to generate the symmetry attenuation coefficient; Normalize the edge length proportion change rate, the diagonal line angle offset amount and the symmetry attenuation coefficient, and dynamically adjust the correction proportion of each parameter based on the real-time transmission priority of the network topology path to obtain the corrected parameters; Fuse the corrected parameters to generate the dynamic geometric correction factor reflecting the abnormal propagation path of the network topology.

3. The network time server timekeeping precision evaluation method according to claim 2, characterized by, Fuse the dynamic geometric correction factor with the comprehensive timing accuracy evaluation value, generate the optimized comprehensive timing accuracy evaluation value and dynamic confidence range through the credible interval of the vertex offset direction adjustment evaluation value and the coverage range contraction rate of the quadrilateral, including: According to the correlation between the coverage shrinkage of the quadrilateral and the network traffic mutation, the influence of the dynamic geometric correction factor on the boundary expansion of the confidence interval of the comprehensive timing accuracy evaluation value is analyzed, and a boundary expansion coefficient is generated; Based on the boundary expansion coefficient, the mapping relationship between the vertex offset direction and the abnormal propagation path of the network topology is extracted, the priority of the abnormal propagation direction is identified, and the dynamic distribution correction proportion of the comprehensive timing accuracy evaluation value in the confidence interval is determined according to the priority of the abnormal propagation direction; Based on the boundary expansion coefficient, the confidence interval of the comprehensive timing accuracy evaluation value is initially scaled to obtain an initially scaled confidence interval, and the initially scaled confidence interval is dynamically calibrated again to generate an optimized comprehensive timing accuracy evaluation value and a dynamic confidence range.

4. The network time server timekeeping accuracy evaluation method according to claim 3, characterized by, Based on the optimized comprehensive timing accuracy evaluation value and the dynamic confidence range, combined with the preset congestion state classification threshold, the timing accuracy stability level of the network time server in the current dynamic network environment, the abnormal event warning information and the network topology abnormal area positioning report based on the quadrilateral vertex offset direction identification are output, including: Multi-dimensional matching of the dynamic confidence range of the optimized comprehensive timing accuracy evaluation value and the preset congestion state classification threshold is performed to determine the timing accuracy stability level; According to the deviation degree of the timing accuracy stability level and the dynamic confidence range, a level confidence parameter is generated; Based on the level confidence parameter, a dynamic abnormal event detection threshold is set to identify abnormal events with an evaluation value greater than or equal to the dynamic abnormal event detection threshold, and the propagation direction and diffusion rate of the abnormal event on the topology path are analyzed based on the spatio-temporal correlation of the quadrilateral vertex offset direction; According to the propagation direction and diffusion rate, combined with the coupling relationship between the vertex offset direction and the network topology path, the starting node and relay node of the abnormal propagation path are located, and the node performance baseline data is associated; Based on the node performance baseline data and the diffusion rate, the feature library of historical network congestion events is matched to generate a repair priority list of core switch nodes and boundary router nodes; According to the repair priority list, combined with the level confidence parameter, a positioning report containing node identification, abnormal propagation path and repair suggestion is generated.

5. The network time server timekeeping precision evaluation method according to claim 4, characterized by, The core index data includes time deviation sequence, time jitter parameter, synchronization period stability coefficient and maximum delay error value; the auxiliary index data includes network delay distribution characteristics, clock frequency offset and position information of the four preset target detection points.

6. A system for evaluating the timekeeping accuracy of a network time server, the system implementing the method of any one of claims 1 to 5, characterized in that, Including: The data acquisition module is used for real-time acquisition of the core index data and auxiliary index data of the network time server in the dynamic network environment; The data processing module is used for normalizing core index data to generate a standardized core index data set, and dynamically integrating according to fluctuation characteristics and correlation of each core index to generate a comprehensive core index representation value, including: performing sliding window segmentation on a time deviation sequence, analyzing standard deviation of short-term fluctuation amplitude and positive or negative slope distribution of trend direction in each window to generate time deviation dynamic fluctuation characteristics; based on the time deviation dynamic fluctuation characteristics, extracting peak value distribution characteristics of time jitter parameters, screening jitter events through a preset threshold, counting jitter event frequency and intensity cumulative value to generate a time jitter abnormal event quantization parameter set; according to the time jitter abnormal event quantization parameter set, converting a synchronization period stability coefficient from a time domain to a frequency domain, calculating a main oscillation frequency and a harmonic component energy proportion, and generating a synchronization period stability frequency spectrum quantization index; combining the synchronization period stability frequency spectrum quantization index, based on historical baseline data of a maximum delay error value, counting an extreme event trigger probability exceeding a historical threshold, and generating a maximum delay risk level parameter in association with a real-time network load state; inputting the time deviation dynamic fluctuation characteristics, the time jitter abnormal event quantization parameter set, the synchronization period stability frequency spectrum quantization index, and the maximum delay risk level parameter into a dynamic time window, calculating a time-varying correlation coefficient matrix among the parameters, and constructing a multi-dimensional dynamic correlation model; projecting the standardized core index data set based on the multi-dimensional dynamic correlation model to extract dynamic feature vectors of different time granularities; performing multi-time granularity superposition and normalization processing on the dynamic feature vectors to generate a comprehensive core index representation value reflecting a dynamic network environment coupling effect; The dynamic compensation correction module is used for dynamically compensating and correcting the comprehensive core index representation value in combination with network delay distribution characteristics and clock frequency offset in auxiliary index data to generate a comprehensive time service accuracy evaluation value, including: setting a dynamic abnormal jump threshold according to burst fluctuation intensity and duration of the network delay distribution characteristics to identify an abnormal jump interval in the time deviation sequence that is greater than or equal to the dynamic abnormal jump threshold; performing segmented compensation on the abnormal jump interval by using a sliding window smoothing algorithm to calculate a time deviation correction sequence after compensation; based on the time deviation correction sequence after compensation, performing reverse calibration on cumulative drift error of the synchronization period stability coefficient to generate a drift error compensation coefficient; superimposing the time deviation correction sequence and the drift error compensation coefficient into the comprehensive core index representation value to generate a time service accuracy evaluation value after preliminary compensation, and calculating a real-time change rate of the network delay distribution characteristics based on real-time fluctuation amplitude of the time service accuracy evaluation value; dynamically adjusting a window length and compensation strength of the sliding window smoothing algorithm according to the real-time change rate of the network delay distribution characteristics to generate a comprehensive time service accuracy evaluation value containing dynamic network delay correction and clock frequency offset suppression. The dynamic geometry correction module is configured to calculate a side length ratio change rate, a diagonal angle offset amount and a symmetry decay coefficient of the quadrilateral vertex coordinates based on position information of four target detection points in the auxiliary index data, the four target detection points being respectively located at a core switch node, a border router node, a key service server node and a time server node in the network topology, and to generate a dynamic geometry correction factor. The evaluation optimization module is configured to fuse the dynamic geometry correction factor with the comprehensive timing accuracy evaluation value, to generate an optimized comprehensive timing accuracy evaluation value and a dynamic confidence range through a credible interval of a quadrilateral coverage range shrinkage rate and a vertex offset direction adjustment evaluation value. The report generation module is configured to output a timing accuracy stability level, an abnormal event early warning information and a network topology abnormal area positioning report based on the optimized comprehensive timing accuracy evaluation value and the dynamic confidence range, and in combination with a preset congestion state grading threshold.

7. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a program, and the program is executed by the processor to implement the method in any one of claims 1 to 5.

Images