Land-sea communication synchronization method and system of super-converged architecture

Abstract

The application relates to the technical field of data transmission, in particular to a land-sea communication synchronization method and system of a hyper-converged architecture. The method comprises the following steps: acquiring time delay distribution data sets of a submarine optical cable and generating a mapping diagram, extracting time delay features and determining a non-standardized feature cluster group; acquiring temperature gradient data based on the cluster group, predicting the dynamic change trend of the time delay gradient, forming a prediction sequence, and adjusting compensation parameters to generate a time delay compensation model; extracting compensation coefficients from the model, injecting a time synchronization protocol, determining a synchronization error correction value, calibrating a clock signal of data exchange, and generating a high-precision time synchronization sequence; generating a feedback signal according to the synchronization sequence, dynamically adjusting the collection frequency of an environmental sensor, and constructing a stable time delay management framework. The application realizes accurate modeling and compensation of the nonlinear time delay distribution under a complex submarine environment, improves the time synchronization precision and system stability of land-sea communication, and guarantees the reliability and real-time performance of the land-sea communication network.

Description

Technical Field

[0001] This invention relates to the field of data transmission technology, and in particular to a land-sea communication synchronization method and system based on a hyper-converged architecture. Background Technology

[0002] In land-sea communication networks, submarine optical cables serve as the primary information transmission channel, boasting long transmission distances and wide coverage, but operating in extremely complex environments. The propagation of optical signals within these cables is not only limited by their physical structure but also significantly affected by environmental factors such as seawater temperature, pressure, and ocean currents, resulting in nonlinear and dynamically fluctuating signal delays.

[0003] Existing technologies typically rely on static compensation or fixed parameters for time delay correction. However, this approach struggles to adapt to the rapid changes in the marine environment, easily leading to time delay compensation failure and the continuous accumulation of clock synchronization errors. Particularly in applications with extremely high time synchronization requirements, such as financial transactions, satellite navigation, and distributed computing, even microsecond-level time delay deviations can cause serious consequences such as incorrect transaction ordering, decreased positioning accuracy, or even system instability. Furthermore, traditional methods have limited capabilities in handling nonlinear time delay distributions, making it difficult to comprehensively capture the regularities in complex environments. While high-frequency acquisition can enhance sensitivity to time delay changes, it increases system load and energy consumption; low-frequency acquisition may miss critical changes, weakening the effectiveness of the compensation model.

[0004] Therefore, how to accurately model and predict nonlinear time delay distribution in a dynamic and complex seabed environment, and achieve high-precision time synchronization of communication under a hyper-converged architecture integrating land and sea, has become an urgent technical challenge to be solved. Summary of the Invention

[0005] This invention proposes a hyperconverged architecture-based land-sea communication synchronization method and system for achieving high-precision time synchronization and stable latency management in complex environments. Firstly, this invention provides a hyperconverged architecture-based land-sea communication synchronization method, mainly comprising:

[0006] Step S1: Obtain the time delay distribution dataset of submarine optical cables and generate a time delay distribution map; extract time delay features based on the time delay distribution map and determine the cluster groups of non-normalized time delay features;

[0007] Step S2: Obtain temperature gradient data based on the cluster group, predict the dynamic change trend of the time delay gradient, and obtain a time delay gradient prediction sequence; adjust the time delay compensation parameters based on the time delay gradient prediction sequence to generate a time delay compensation model;

[0008] Step S3: Extract compensation coefficients from the time delay compensation model, inject them into the time synchronization protocol, and determine the synchronization error correction value; calibrate the clock signal for data exchange according to the synchronization error correction value, and generate a time synchronization sequence;

[0009] Step S4: Generate a feedback signal based on the time synchronization sequence, adjust the data acquisition frequency, generate a stable time delay management framework, and perform land-sea synchronous communication based on the time delay management framework.

[0010] As a preferred embodiment of the present invention, step S1 involves obtaining a time delay distribution dataset of the submarine optical cable and generating a time delay distribution mapping map, including:

[0011] By collecting optical signal propagation data and environmental sensor data from each segment of the submarine optical cable, an initial time delay distribution dataset including repeater amplifier spacing parameters and seawater temperature gradient values ​​was obtained.

[0012] Based on the initial time delay distribution dataset, a data processing algorithm is used to analyze the optical signal propagation data and extract the time delay values ​​of each segment.

[0013] Based on the time delay value and the seawater temperature gradient value, a preliminary mapping map characterizing the time delay distribution is generated, wherein the preliminary mapping map reflects the correspondence between the time delay of each segment and environmental parameters.

[0014] As a preferred embodiment of the present invention, step S1, extracting time delay features based on the time delay distribution mapping map and determining cluster groups of non-standardized time delay features, includes:

[0015] Based on the time delay distribution mapping, a convolutional neural network is used to extract features from the time delay distribution to obtain a feature vector representing the nonlinear time delay gradient distribution.

[0016] For outliers in the feature vector, a clustering analysis algorithm is used to group them, generating multiple clusters of non-standardized time delay features;

[0017] Based on the cluster groups, the distribution pattern of time delay features within each group is determined, and classification results representing non-standardized time delay features are generated.

[0018] As a preferred embodiment of the present invention, step S2, obtaining the time delay gradient prediction sequence, includes:

[0019] Temperature gradient data related to non-standardized time delay features are extracted from the clusters; if the temperature gradient data exceeds a preset threshold, the temperature gradient data is analyzed through a long short-term memory network to predict the dynamic change trend of the time delay gradient; based on the dynamic change trend, a prediction sequence characterizing the future change of the time delay gradient is generated, wherein the prediction sequence includes time delay gradient values ​​at multiple time points.

[0020] As a preferred embodiment of the present invention, step S2, adjusting the time delay compensation parameters according to the time delay gradient prediction sequence to generate a time delay compensation model, includes:

[0021] Based on the predicted time delay gradient sequence, an adaptive filter is used to adjust the time delay compensation parameters in real time.

[0022] By integrating the relay spacing influence factor in the time delay gradient prediction sequence, an adjusted set of compensation parameters is generated.

[0023] Based on the set of compensation parameters, the convergence of the compensation parameters is determined, and an optimization model characterizing the delay compensation effect is generated, wherein the optimization model includes dynamic compensation rules for delay fluctuations.

[0024] As a preferred embodiment of the present invention, step S3, which involves extracting compensation coefficients from the time delay compensation model, injecting them into the time synchronization protocol, and determining the synchronization error correction value, includes:

[0025] Extract compensation coefficients representing the delay compensation effect from the delay compensation model; based on the compensation coefficients, inject parameters into the time synchronization protocol in the hyperconverged architecture and adjust the clock parameters of the time synchronization protocol; based on the adjusted clock parameters, calculate the synchronization error in data exchange and generate a correction value representing the magnitude of the error, wherein the correction value is used to optimize the time synchronization accuracy.

[0026] As a preferred embodiment of the present invention, step S3, calibrating the clock signal for data exchange according to the synchronization error correction value, and generating a time synchronization sequence, includes:

[0027] Based on the synchronization error correction value, the clock signal for data exchange is calibrated to generate a calibrated clock signal sequence. If the synchronization accuracy of the calibrated clock signal sequence is lower than a preset target threshold, the delay compensation model is iteratively updated and the compensation coefficients are regenerated. Based on the updated compensation coefficients, the clock signal sequence is adjusted to generate the final time synchronization sequence.

[0028] As a preferred embodiment of the present invention, step S4 includes:

[0029] Based on the time synchronization sequence, a feedback signal characterizing the delay management effect is generated; the feedback signal is transmitted back to the environmental sensor to adjust the data acquisition frequency of the environmental sensor; based on the adjusted data acquisition frequency, the changing trend of the delay distribution dataset is analyzed to determine the stability of the overall system; based on the stability judgment result, a stable framework characterizing dynamic delay management is generated, and land-sea synchronous communication is performed based on the stable framework.

[0030] Secondly, the present invention also provides a land-sea communication synchronization system with a hyper-converged architecture for implementing the above-mentioned method, the system comprising:

[0031] The clustering determination unit is used to acquire the time delay distribution dataset of submarine optical cables, generate a time delay distribution map, extract time delay features based on the time delay distribution map, and determine the cluster groups of non-normalized time delay features.

[0032] The model generation unit is used to obtain temperature gradient data based on the cluster group, predict the dynamic change trend of the time delay gradient, and obtain a time delay gradient prediction sequence; adjust the time delay compensation parameters based on the time delay gradient prediction sequence, and generate a time delay compensation model.

[0033] The sequence generation unit is used to extract compensation coefficients from the delay compensation model, inject them into the time synchronization protocol, determine the synchronization error correction value, calibrate the clock signal for data exchange according to the synchronization error correction value, and generate a time synchronization sequence.

[0034] The framework generation unit is used to generate feedback signals based on the time synchronization sequence, adjust the data acquisition frequency, generate a stable time delay management framework, and perform land-sea synchronous communication based on the time delay management framework.

[0035] Thirdly, the present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the above-described method.

[0036] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0037] This invention constructs a multi-source fusion time delay distribution dataset by collecting optical signal propagation data from submarine optical cables and information such as temperature and pressure from environmental sensors. A mapping map is generated using Fourier transform and interpolation algorithms, intuitively reflecting the correspondence between time delay and environmental parameters, thus providing a reliable basis for anomaly identification and nonlinear feature modeling. Furthermore, a convolutional neural network is used to extract non-standardized time delay features and combine this with cluster analysis to form clusters. Temperature gradient data is then extracted based on these clusters and dynamically predicted using a long short-term memory network, effectively solving the problem of traditional methods failing to capture nonlinear and dynamic change patterns and significantly improving prediction accuracy. Subsequently, an adaptive filter combined with a relay spacing factor is introduced to optimize compensation parameters, generating a time delay compensation model. The extracted compensation coefficients are then injected into a time synchronization protocol to obtain synchronization error correction values, which can eliminate clock deviations in data exchange in real time, thereby generating a high-precision time synchronization sequence and ensuring the communication stability of the cross-sea link. Finally, a feedback signal is generated from the time synchronization sequence and transmitted back to the environmental sensors to dynamically adjust the data acquisition frequency. Combined with trend analysis and stability judgment, a stable time delay management framework is formed, realizing a closed-loop control of synchronization-feedback-optimization. By combining the above technical solutions, the problem of insufficient synchronization accuracy caused by the nonlinearity and dynamism of submarine optical cable time delay distribution is not only solved, but also the stability and energy efficiency of the system are improved through intelligent prediction and adaptive compensation. It has wide application value in high-precision communication applications such as financial transactions, satellite navigation and distributed computing. Attached Figure Description

[0038] Figure 1 This is a flowchart of a land-sea communication synchronization method based on a hyperconverged architecture, as described in an embodiment of the present invention.

[0039] Figure 2 This is a structural diagram of a land-sea communication synchronization system with a hyperconverged architecture, according to an embodiment of the present invention. Detailed Implementation

[0040] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0041] like Figure 1 This embodiment provides a land-sea communication synchronization method based on a hyperconverged architecture, which may specifically include:

[0042] Step S1: Obtain the time delay distribution dataset of submarine optical cables and generate a time delay distribution map; extract time delay features based on the time delay distribution map and determine the cluster groups of non-normalized time delay features;

[0043] In step S1, the time delay distribution dataset of the submarine optical cable is obtained, and a time delay distribution mapping map is generated, including:

[0044] By collecting optical signal propagation data and environmental sensor data from each segment of the submarine optical cable, an initial time delay distribution dataset including repeater amplifier spacing parameters and seawater temperature gradient values ​​was obtained.

[0045] Based on the initial time delay distribution dataset, a data processing algorithm is used to analyze the optical signal propagation data and extract the time delay values ​​of each segment.

[0046] Based on the time delay value and the seawater temperature gradient value, a preliminary mapping map characterizing the time delay distribution is generated, wherein the preliminary mapping map reflects the correspondence between the time delay of each segment and environmental parameters.

[0047] Specifically, due to the extremely complex transmission environment of submarine optical cables, the optical signal is not only constrained by the physical characteristics of the cable itself during propagation, but also affected by environmental factors such as seawater temperature, pressure, and ocean current changes. This results in communication delays exhibiting nonlinear and dynamically changing characteristics. Traditional static calibration methods that rely on fixed compensation parameters cannot adapt to this complexity, leading to problems such as insufficient clock synchronization accuracy and error accumulation in data exchange. Therefore, in a hyperconverged architecture land-sea communication synchronization method, the initial delay distribution dataset, including relay amplifier spacing parameters and seawater temperature gradient values, is obtained by first collecting optical signal propagation data from each segment of the submarine optical cable and real-time readings from environmental sensors deployed at the cable nodes. The environmental sensors include temperature sensors and pressure sensors. An optical time-domain reflectometer is used to collect the propagation characteristics of the optical signal segment by segment. Event points corresponding to relay nodes are formed in the reflector echo traces, which are used to locate the spatial positions of adjacent relays and calculate the relay amplifier spacing. The temperature and pressure sensors provide external parameters directly related to the environment in which the optical cable is located, thereby ensuring that the dataset can cover the entire length of the optical cable in both spatial and temporal dimensions.

[0048] Subsequently, based on the initial time delay distribution dataset, a preset data processing algorithm is used to analyze the optical signal propagation data to extract the time delay value of each segment. This preset data processing algorithm includes converting the time-domain signal to a frequency-domain signal using a Fourier transform algorithm, identifying the propagation delay of the optical signal and separating noise interference using spectral characteristics, and then obtaining the average time delay value of each segment by integrating the signal peak position. For example, the time delay value obtained in a 100-kilometer-long optical cable is 5 milliseconds. The extracted time delay value accurately reflects the signal propagation time in a real physical environment. Further, the time delay value is paired with the seawater temperature gradient value according to their corresponding positions, and an interpolation algorithm is used to generate a two-dimensional time delay distribution map based on the pairing result. The horizontal axis represents the optical cable. The plot uses the vertical axis to represent the time delay value, and the color gradient visually reflects the impact of different temperature gradients on propagation time delay, thus modeling the correspondence between environmental parameters and time delay characteristics. In scenarios with significant differences in environmental conditions, such as shallow and deep sea areas, the mapping map can reflect the non-uniform distribution of temperature gradients. For example, when the temperature gradient in shallow sea areas is 0.2 degrees Celsius per meter, the time delay value shows an increase of about 10%. This change is highlighted in the form of a heat map, thereby optimizing the synchronization accuracy of long-distance optical cables. When the temperature gradient value exceeds a preset threshold, the mapping map can automatically mark and highlight abnormal areas, and reveal the non-linear relationship between time delay and temperature through curve fitting. The fitting results show that the time delay has a quadratic growth trend with the temperature gradient, thus providing a predictive basis for subsequent synchronization compensation algorithms.

[0049] The above technical solution enables hierarchical extraction and mapping of multi-source environmental and signal data to time delay distribution characteristics. This ensures the integrity of data acquisition, clarifies the logical correspondence between different parameters, and, through Fourier transform and interpolation fitting algorithms, accurately characterizes the dynamic and nonlinear laws of time delay distribution. This provides reliable data support and model foundation for further high-precision synchronization of land and sea communications under a hyper-converged architecture.

[0050] Further, in step S1, delay features are extracted based on the delay distribution map, and clusters of non-standardized delay features are determined, including:

[0051] Based on the time delay distribution map, a convolutional neural network is used to extract features from the time delay distribution to obtain a feature vector representing the nonlinear time delay gradient distribution. For outliers in the feature vector, a clustering analysis algorithm is used to group them to generate multiple clusters of non-standardized time delay features. Based on the clusters, the distribution pattern of time delay features within each group is determined to generate a classification result representing the non-standardized time delay features.

[0052] Specifically, the acquired time delay distribution map is used as input data and fed into the convolutional layer of a convolutional neural network. The convolutional neural network includes multiple convolutional and pooling layers to capture the spatial features reflecting the time delay distribution in the map. In the convolutional layer, the convolutional kernel performs local convolution operations on the input map to extract local time delay gradient patterns. The data is then processed by a nonlinear activation function such as ReLU to maintain the nonlinear relationship, thereby generating a preliminary feature map. Subsequently, the pooling layer downsamples the preliminary feature map to reduce the data dimensionality and retain key information, forming a compressed feature representation. This feature representation is then fed into a fully connected layer for fusion, enabling it to combine global time delay distribution information and output a feature vector representing the nonlinear time delay gradient distribution. This feature vector contains both the magnitude and direction components of the time delay change.

[0053] In practical implementation, the convolutional neural network can adopt a variant of the LeNet-5 architecture, using historical latency data as training labels to generate multi-dimensional feature vectors. These vectors can characterize latency peaks caused by sudden temperature changes, thereby enabling effective identification of nonlinear latency patterns in complex seabed environments. For example, when the acquired latency distribution map shows that the latency gradient of a certain cable segment rises sharply from 0.5 ms / km to 2.0 ms / km, the convolutional operation can capture this nonlinear change and generate vector components that reflect a gradient slope of approximately 1.5, thus supporting subsequent anomaly detection.

[0054] For outliers in the aforementioned feature vectors—that is, time delay feature points that significantly deviate from the normal pattern, and also for time delay mutation points in nonlinear and non-uniform environments—anomalies are identified by judging deviations from the average value and using a standard deviation threshold for screening. Then, the K-means clustering algorithm is applied to group the outliers. Variance is minimized by minimizing the Euclidean distance from each point to the cluster center, and the elbow rule is used to determine the appropriate number of clusters, ultimately generating multiple cluster groups. Each cluster group corresponds to a non-standardized time delay feature, such as high gradient anomalies or uneven relay spacing anomalies. In specific implementation, the K value can be set to a positive integer greater than or equal to 3, thus obtaining high temperature gradient groups, medium spacing influence groups, and mixed anomaly groups, making the classification more accurate and facilitating targeted time delay compensation. Based on the feature vectors within each cluster group, statistical indicators of the time delay features, such as mean and variance, are calculated, and their distribution patterns, such as normal or skewed distributions, are analyzed. This generates classification results, and feature labels are assigned to each group. These results will serve as an important reference for subsequent model optimization.

[0055] The above technical solution, through the joint processing of convolutional neural networks and cluster analysis, not only realizes the correspondence between the original data and feature vectors, cluster groups and classification results, but also establishes the mapping between latency features and environmental factors, thereby providing more accurate and interpretable synchronization support for land and sea communication under the hyperconverged architecture.

[0056] Step S2: Obtain temperature gradient data based on the cluster group, predict the dynamic change trend of the time delay gradient, and obtain a time delay gradient prediction sequence; adjust the time delay compensation parameters based on the time delay gradient prediction sequence to generate a time delay compensation model;

[0057] In step S2, obtaining the time-delay gradient prediction sequence includes:

[0058] Temperature gradient data related to non-standardized time delay features are extracted from the clusters; if the temperature gradient data exceeds a preset threshold, the temperature gradient data is analyzed through a long short-term memory network to predict the dynamic change trend of the time delay gradient; based on the dynamic change trend, a prediction sequence characterizing the future change of the time delay gradient is generated, wherein the prediction sequence includes time delay gradient values ​​at multiple time points.

[0059] Specifically, temperature gradient data directly related to non-standardized time delay features are extracted from the aforementioned clusters. This temperature gradient data consists of numerical sequences formed by the seawater temperature change values ​​collected by environmental sensors and the corresponding optical cable segment locations. By sorting the non-standardized time delay features within the clusters according to their degree of anomaly and then matching the corresponding temperature gradient values, it is ensured that the extracted data are concentrated in high-gradient regions associated with anomalous time delay points. For example, only data points with temperature changes greater than 0.5 degrees Celsius per kilometer are retained, while irrelevant data with changes less than 0.1 degrees Celsius are removed, thereby improving the relevance of the input data and the prediction accuracy of the long short-term memory network.

[0060] Subsequently, when the temperature gradient data exceeds a preset threshold, it is analyzed and predicted using a Long Short-Term Memory (LSTM) network. As a recurrent neural network, LSTM can process time-series data and retain long-term dependencies. In this embodiment, the extracted temperature gradient numerical sequence is first normalized to the range of 0 to 1, and then input into the hidden layer of the network. After adjusting the weight parameters using the backpropagation algorithm, the network outputs the corresponding time-delay gradient change pattern. For example, when the temperature gradient is 1.5 degrees Celsius per kilometer, which is greater than the preset threshold of 1.0 degrees Celsius per kilometer, the predicted trend shows that the time-delay gradient increases at a rate of 0.1 milliseconds per hour, thus providing advance adjustment for subsequent compensation. Furthermore, to improve the robustness of the prediction, the LSTM network can be extended to a multi-layer structure, such as a two-layer LSTM, to enhance its ability to remember long sequences and introduce historical time-delay data into the input, thereby improving the prediction dynamics. The trend analysis comprehensively considers the seasonal variations in ocean temperature. For example, under high-temperature conditions in summer, the prediction results tend to show a sharp increase in time delay, thereby improving the model's adaptability to the actual environment. This prediction result is more consistent with the actual change patterns than single-parameter prediction. Finally, a prediction sequence representing future changes is generated based on the obtained dynamic trend. This prediction sequence consists of time delay gradient values ​​at multiple time points and can be generated by linear interpolation combined with the trend slope to ensure that the sequence covers at least N time points and provides continuity. N is a positive integer greater than or equal to 10, thus realizing a complete logical chain from temperature gradient anomaly identification to dynamic trend prediction and then to the generation of future time delay sequences. This prediction sequence can directly provide input data for the adaptive filter to adjust the compensation parameters, thereby achieving a close connection between time delay prediction and compensation calibration, ultimately reducing the error accumulation in the clock signal synchronization process of submarine cables.

[0061] Further, in step S2, adjusting the time delay compensation parameters according to the time delay gradient prediction sequence to generate a time delay compensation model includes:

[0062] Based on the time delay gradient prediction sequence, the time delay compensation parameters are adjusted in real time using an adaptive filter; the relay spacing influence factor in the time delay gradient prediction sequence is fused to generate an adjusted set of compensation parameters; based on the set of compensation parameters, the convergence of the compensation parameters is determined, and an optimization model characterizing the time delay compensation effect is generated, wherein the optimization model includes dynamic compensation rules for time delay fluctuations.

[0063] Specifically, an adaptive filter is introduced for the predicted time delay gradient sequence to achieve real-time adjustment of the time delay compensation parameters. In this process, the influence factor of relay spacing is integrated. Finally, by judging the convergence of parameters, an optimized model representing the time delay compensation effect is generated, thereby achieving stable synchronization under the submarine optical cable link. In the specific implementation process, firstly, based on the time delay gradient prediction sequence output by the long short-term memory network, the current time delay value and historical time delay data are extracted and used as the input signal of the adaptive filter. The adaptive filter iteratively updates the filter coefficients based on the minimum mean square error algorithm, so that the error between the predicted time delay and the actual time delay is gradually reduced. The weights are corrected in real time by the recursive least squares method to achieve dynamic correction of the compensation parameters. For example, when the predicted sequence shows an average delay of 150 milliseconds, the initial coefficient of the adaptive filter is set to 0.01. After ten iterations, the compensation parameter is optimized to 0.005 milliseconds / ℃, thereby reducing the overall delay fluctuation. Furthermore, an influence factor on relay spacing is introduced. The weight value is obtained by quantifying the relay amplifier spacing parameter. This weight is multiplied by the compensation parameter output by the adaptive filter to obtain the fused parameter set. Normalization is then used to ensure that the parameter values ​​are limited to between 0 and 1. For instance, when the relay spacing is 50 kilometers, the influence factor is calculated to be 1.0. If the spacing increases to 60 kilometers, the influence factor increases to 1.2. After fusion, the parameter value in the compensation parameter set is adjusted from 0.03 to 0.036, thereby improving the delay compensation accuracy. In scenarios with drastic temperature changes, the fused parameter set can cover a range of 0.01 to 0.05 to support dynamic delay adjustment and reduce synchronization errors.

[0064] Subsequently, the convergence of the compensation parameter set is determined based on its variance. When the parameter change rate is lower than a preset threshold and the variance is less than the set variance, the compensation parameters are confirmed to have converged, and an optimization model is constructed based on the convergence result. This model adjusts the latency fluctuation by setting dynamic compensation rules. For example, it defines that when the latency fluctuation exceeds five milliseconds, the compensation coefficient is automatically increased. This rule is embedded into the real-time monitoring process to ensure that the output compensation effect index can reflect the system stability, thereby effectively reducing the clock deviation in data exchange. This allows for the construction of a stable latency management framework under the hyperconverged architecture and ensures high-precision synchronization of the global communication network.

[0065] Step S3: Extract compensation coefficients from the time delay compensation model, inject them into the time synchronization protocol, and determine the synchronization error correction value; calibrate the clock signal for data exchange according to the synchronization error correction value, and generate a time synchronization sequence;

[0066] In step S3, the compensation coefficients are extracted from the delay compensation model, injected into the time synchronization protocol, and the synchronization error correction value is determined, including:

[0067] Extract compensation coefficients representing the delay compensation effect from the delay compensation model; based on the compensation coefficients, inject parameters into the time synchronization protocol in the hyperconverged architecture and adjust the clock parameters of the time synchronization protocol; based on the adjusted clock parameters, calculate the synchronization error in data exchange and generate a correction value representing the magnitude of the error, wherein the correction value is used to optimize the time synchronization accuracy.

[0068] Specifically, compensation coefficients characterizing the delay compensation effect are extracted from the delay compensation model and used for subsequent adjustment of time synchronization protocol parameters, thereby achieving precise calibration of the submarine communication link clock signal in the overall scheme. Specifically, the compensation coefficients are optimized parameters generated by an adaptive filter during iterative calculation, combining the influence factors of relay spacing and temperature gradient data. These parameters are used to quantify the adjustment magnitude of the nonlinear delay gradient. The extracted compensation coefficients are first mapped to the clock offset parameters of the time synchronization protocol. The time synchronization protocol in the hyperconverged architecture adopts a standard implementation of a precision time protocol, and the injection process is executed through a software-defined network interface to ensure that frequency and phase corrections are updated in real time. For example, when the compensation coefficient is 1.1, the clock parameter offset value increases accordingly. An additional 10% is added to offset the error caused by time delay fluctuations in the submarine communication link, thereby enhancing the reliability of communication. During the parameter injection process, the compensation coefficient is decomposed into frequency adjustment components and phase adjustment components. The frequency component is obtained by multiplying the coefficient by the reference clock frequency, while the phase component is calculated by linearly combining the coefficient with the historical phase deviation. The above-mentioned injection vector is applied to the clock model of the time synchronization protocol. This model is based on a master-slave synchronization structure and ensures that the deviation between the master clock and the slave clock is minimized after the parameter update. After the above injection is completed, the clock signal in the communication data exchange process is simulated and calculated according to the adjusted clock parameters. The clock deviation of each exchange node is extracted and the deviation is quantified as a synchronization error. The calculation process of the synchronization error is achieved by summing the basic differences or using the simplified root mean square formula.

[0069] Through the above technical solutions, the correction value can be continuously updated in the form of weighted average, forming an iterative optimization closed loop, thereby effectively reducing the synchronization error in extreme environments such as high temperature gradient submarine optical cable sections. The accurate extraction and dynamic injection of the above compensation coefficients not only establishes a logical mapping relationship between environmental parameters and clock correction parameters, but also enables the overall system to maintain high-precision communication synchronization in complex marine environments, thereby ensuring the stability and real-time performance of communication in the land-sea integrated hyperconverged architecture.

[0070] Further, in step S3, the clock signal for data exchange is calibrated according to the synchronization error correction value to generate a time synchronization sequence, including:

[0071] Based on the synchronization error correction value, the clock signal for data exchange is calibrated to generate a calibrated clock signal sequence. If the synchronization accuracy of the calibrated clock signal sequence is lower than a preset target threshold, the delay compensation model is iteratively updated and the compensation coefficients are regenerated. Based on the updated compensation coefficients, the clock signal sequence is adjusted to generate the final time synchronization sequence.

[0072] Specifically, in this embodiment, the clock signal for data exchange is calibrated according to the synchronization error correction value to generate a calibrated clock signal sequence. Specifically, a compensation coefficient is first extracted from the optimized time delay compensation model and combined with the synchronization error correction value to apply an offset adjustment to the clock signal for data exchange. The compensation coefficient is calculated based on the time delay value and temperature gradient data extracted from the preliminary time delay distribution map to ensure that the influence of nonlinear time delay gradient is reduced during the propagation of the clock signal in each segment of the submarine optical cable.

[0073] Next, the clock signal is calibrated segment by segment using the adjusted offset value, generating a calibrated clock signal sequence containing multiple timestamps. The generated clock signal sequence accurately reflects the correction effect of the repeater amplifier spacing parameter on signal propagation, thus providing accurate calibration data for delay compensation in communication. If the synchronization accuracy of the calibrated clock signal sequence is lower than a preset target threshold, such as less than 0.1 milliseconds, the process of iteratively updating the delay compensation model will be triggered. At this time, the deviation value between each timestamp in the calibrated clock signal sequence is first calculated and compared with the preset target threshold to determine whether the compensation coefficient needs to be updated. If the deviation value exceeds the threshold, the prediction sequence of the delay gradient is retrained through a long short-term memory network, and the latest seawater temperature gradient value is incorporated for dynamic optimization. The delay change trend is analyzed through historical delay datasets to update the prediction sequence and adjust the compensation coefficient. The introduction of the long short-term memory network, with its advantage of remembering long-term dependencies, effectively captures the changes in the nonlinear delay gradient distribution, providing a reliable prediction basis for the optimization of the compensation model.

[0074] Subsequently, the adaptive filter is adjusted in real time based on the relay amplifier spacing parameter, and the compensation parameters are dynamically optimized using the minimum mean square error algorithm to ensure the convergence and accuracy of the compensation model in the actual environment, thereby effectively reducing synchronization errors and improving system stability. During the iterative update process, for special cases with high temperature gradients or long relay spacing, the number of training rounds or multiple iterations are increased to further optimize the compensation coefficients, ensuring clock synchronization accuracy and stability in complex environments. Finally, based on the updated compensation coefficients, the clock signal sequence is adjusted, and the final time synchronization sequence is generated. Specifically, the regenerated compensation coefficients are injected into the time synchronization protocol, and the calibrated clock signal sequence is subjected to a secondary offset adjustment to generate the final clock sequence containing optimized timestamps, ensuring that the final clock signal is consistent with the actual delay and can efficiently support subsequent real-time data exchange. At the same time, this final time synchronization sequence will also be fed back to the environmental sensor through feedback signals to further adjust the data acquisition frequency and ensure the stability of the overall delay management system.

[0075] The above-mentioned technical solution, through continuous iteration and optimization, ensures high-precision and stable clock synchronization in complex land and sea communication environments, improves the reliability of data transmission over long distances via optical cables, and reduces communication errors and packet loss caused by time delay fluctuations, thereby meeting the high requirements of communication for synchronization accuracy and real-time performance.

[0076] Step S4: Generate a feedback signal based on the time synchronization sequence, adjust the data acquisition frequency, generate a stable latency management framework, and perform land-sea synchronous communication based on the latency management framework; specifically including:

[0077] Based on the time synchronization sequence, a feedback signal characterizing the delay management effect is generated; the feedback signal is transmitted back to the environmental sensor to adjust the data acquisition frequency of the environmental sensor; based on the adjusted data acquisition frequency, the changing trend of the delay distribution dataset is analyzed to determine the stability of the overall system; based on the stability judgment result, a stable framework characterizing dynamic delay management is generated, and land-sea synchronous communication is performed based on the stable framework.

[0078] Specifically, after generating the time synchronization sequence, the delay compensation value and synchronization error correction value contained in the time synchronization sequence are further extracted and used as the core parameters of the feedback signal. This feedback signal directly characterizes the optimization degree of the overall delay management effect. Based on the extracted core parameters, the intensity value of the feedback signal is calculated and used for subsequent transmission and processing. The feedback signal is encoded as a digital pulse sequence and transmitted back to the environmental sensor deployed on the seabed via an optical fiber transmission link. The intensity value of the feedback signal determines the amplitude of the pulse sequence. When the environmental sensor receives the pulse sequence, it analyzes it and determines the dynamic adjustment of the data acquisition frequency based on the comparison result of the intensity value and a preset threshold. If the intensity value is higher than the threshold, it indicates that the current delay management effect is good, and the environmental sensor reduces the acquisition frequency to reduce energy consumption; conversely, the acquisition frequency is increased to obtain more environmental data, thereby ensuring the capture of dynamic delay changes. For example, when the feedback signal intensity value is 0.8, the acquisition frequency is adjusted from once per minute to once every two minutes, effectively reducing unnecessary sampling operations and improving overall efficiency.

[0079] Based on this, a new time delay distribution dataset is formed using the adjusted acquisition frequency, and time delay value sequences are extracted from it. The mean and variance of the sequences are calculated. The mean is used to characterize the overall average time delay level, and the variance is used to reflect the amplitude of time delay fluctuations. Subsequently, the time delay value sequences are smoothed using a moving average method to obtain a more continuous trend curve. A linear regression model is then used to fit the trend slope on this curve. If the absolute value of the slope is less than a set threshold, the trend is considered stable; if it is greater than this value, the trend is considered unstable. At the same time, a comprehensive judgment is made by combining seawater temperature gradient data. If the variance is lower than a preset threshold and the slope meets the stability condition, the overall system is considered to be in a stable state; otherwise, it is marked as unstable, thus achieving a quantitative evaluation of system stability. In longer optical cable sections... In application scenarios, this can be further expanded to include anomaly detection steps. That is, when the dataset formed under the adjusted acquisition frequency shows a decreasing trend in latency, even if the initial variance is large, the system can be judged to be stable by reducing the variance. This can reduce energy consumption while maintaining synchronization accuracy. For example, if the initial variance is 2.0, it can be reduced to 0.4 after adjustment, which is significantly lower than the threshold and is considered to be in a stable state. This extends the lifespan of the sensor and improves the long-term latency management effect. Finally, the system summarizes the above stability judgment results to form a core data structure containing trend curves and statistical indicators. Based on this, dynamic management rules are defined. For example, if the system is judged to be stable three times in a row, the current acquisition frequency is locked to the default setting. This forms a complete stability framework to guide subsequent latency compensation operations.

[0080] Through the above technical solutions, an organic closed loop is achieved between time synchronization sequence, feedback signal, environmental data acquisition and stability judgment, forming a logical chain from synchronization to feedback and then to dynamic adjustment, thereby ensuring the long-term effectiveness of latency management and high-precision synchronization capability in the land and sea communication scenario of hyper-converged architecture.

[0081] This invention also provides a hyperconverged architecture land-sea communication synchronization system for implementing the above-mentioned methods, such as... Figure 2 As shown, the system includes:

[0082] The clustering determination unit is used to acquire the time delay distribution dataset of submarine optical cables, generate a time delay distribution map, extract time delay features based on the time delay distribution map, and determine the cluster groups of non-normalized time delay features.

[0083] The model generation unit is used to obtain temperature gradient data based on the cluster group, predict the dynamic change trend of the time delay gradient, and obtain a time delay gradient prediction sequence; adjust the time delay compensation parameters based on the time delay gradient prediction sequence, and generate a time delay compensation model.

[0084] The sequence generation unit is used to extract compensation coefficients from the delay compensation model, inject them into the time synchronization protocol, determine the synchronization error correction value, calibrate the clock signal for data exchange according to the synchronization error correction value, and generate a time synchronization sequence.

[0085] The framework generation unit is used to generate feedback signals based on the time synchronization sequence, adjust the data acquisition frequency, generate a stable time delay management framework, and perform land-sea synchronous communication based on the time delay management framework.

[0086] The present invention also provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the above-described method.

[0087] In summary, this invention constructs an initial time delay distribution dataset containing relay amplifier spacing parameters and seawater temperature gradient values ​​by collecting submarine optical cable signal propagation data and environmental sensor data. Data processing algorithms are then used to extract time delay values ​​and generate a mapping map. This not only ensures the fusion of multi-source information from the signal and environmental layers but also provides a complete data foundation for subsequent nonlinear feature identification. Subsequently, a convolutional neural network is used to extract features from the mapping map, and cluster analysis is employed to form clusters of non-standardized time delay features. This effectively identifies abnormal time delay distribution patterns, achieving a hierarchical correspondence between the original data and feature classification results, providing targeted input for the dynamic prediction stage. Next, temperature gradient data is obtained using the clusters, and the dynamic trend of time delay gradient changes is predicted using a long short-term memory network to generate a time delay gradient prediction sequence. Finally, an adaptive filter combined with a relay spacing factor is used to optimize the compensation parameters, forming a time delay compensation model. This process, through the synergy of time series forecasting and adaptive parameter iteration, enables the compensation model to adapt to environmental fluctuations in real time, ensuring the forward-looking nature and robustness of the calibration. Compensation coefficients are extracted from the compensation model and injected into the time synchronization protocol. The clock signal is then calibrated based on the calculated synchronization error correction value, generating a time synchronization sequence. This process achieves a seamless mapping between environmental parameters, compensation parameters, and protocol parameters, ensuring that synchronization errors can be dynamically eliminated and significantly improving the consistency and accuracy of the clock signal. Finally, feedback signals are generated and sent back to environmental sensors to dynamically adjust the data acquisition frequency. Stability analysis is then performed using new datasets, forming a dynamic delay management framework. This step not only establishes a closed-loop mechanism of synchronization-feedback-adjustment but also achieves system energy consumption optimization and long-term stable operation through trend prediction and stability assessment. Through the synergy of these technical solutions, accurate modeling and compensation of nonlinear delays in complex seabed environments are ensured, while dynamic adjustment enables system self-adaptation and sustainable optimization, thereby significantly improving the real-time performance, reliability, and stability of land-sea communication within a hyper-converged architecture.

[0088] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements and additions without departing from the principle of the present invention, and these improvements and additions should also be considered within the scope of protection of the present invention.

Claims

1. A land-sea communication synchronization method with a hyperconverged architecture, characterized in that, include: Step S1: Obtain the time delay distribution dataset of submarine optical cables and generate a time delay distribution map; extract time delay features based on the time delay distribution map and determine the cluster groups of non-normalized time delay features; Step S2: Obtain temperature gradient data based on the cluster group, predict the dynamic change trend of the time delay gradient, and obtain a time delay gradient prediction sequence; adjust the time delay compensation parameters based on the time delay gradient prediction sequence to generate a time delay compensation model; Step S3: Extract compensation coefficients from the time delay compensation model, inject them into the time synchronization protocol, and determine the synchronization error correction value; calibrate the clock signal for data exchange according to the synchronization error correction value, and generate a time synchronization sequence; Step S4: Generate a feedback signal based on the time synchronization sequence, adjust the data acquisition frequency, generate a stable time delay management framework, and perform land-sea synchronous communication based on the time delay management framework; In the delay distribution map, the horizontal axis represents the position of the optical cable segment, the vertical axis represents the delay value, and the color gradient reflects the influence of different temperature gradients on the propagation delay.

2. The method as described in claim 1, characterized in that, In step S1, the time delay distribution dataset of the submarine optical cable is obtained, and a time delay distribution mapping map is generated, including: By collecting optical signal propagation data and environmental sensor data from each segment of the submarine optical cable, an initial time delay distribution dataset including repeater amplifier spacing parameters and seawater temperature gradient values ​​was obtained. Based on the initial time delay distribution dataset, a data processing algorithm is used to analyze the optical signal propagation data and extract the time delay values ​​of each segment. Based on the time delay value and the seawater temperature gradient value, a preliminary mapping map characterizing the time delay distribution is generated, wherein the preliminary mapping map reflects the correspondence between the time delay of each segment and environmental parameters.

3. The method as described in claim 1, characterized in that, In step S1, delay features are extracted based on the delay distribution map, and clusters of non-normalized delay features are determined, including: Based on the time delay distribution mapping, a convolutional neural network is used to extract features from the time delay distribution to obtain a feature vector representing the nonlinear time delay gradient distribution. For outliers in the feature vector, a clustering analysis algorithm is used to group them, generating multiple clusters of non-standardized time delay features; Based on the cluster groups, the distribution pattern of time delay features within each group is determined, and classification results representing non-standardized time delay features are generated.

4. The method as described in claim 1, characterized in that, Step S2, obtaining the time-delay gradient prediction sequence, includes: Temperature gradient data related to non-standardized time delay features are extracted from the clusters; if the temperature gradient data exceeds a preset threshold, the temperature gradient data is analyzed through a long short-term memory network to predict the dynamic change trend of the time delay gradient; based on the dynamic change trend, a prediction sequence characterizing the future change of the time delay gradient is generated, wherein the prediction sequence includes time delay gradient values ​​at multiple time points.

5. The method as described in claim 4, characterized in that, In step S2, the time delay compensation parameters are adjusted according to the time delay gradient prediction sequence to generate a time delay compensation model, including: Based on the predicted time delay gradient sequence, an adaptive filter is used to adjust the time delay compensation parameters in real time. By integrating the relay spacing influence factor in the time delay gradient prediction sequence, an adjusted set of compensation parameters is generated. Based on the set of compensation parameters, the convergence of the compensation parameters is determined, and an optimization model characterizing the delay compensation effect is generated, wherein the optimization model includes dynamic compensation rules for delay fluctuations.

6. The method as described in claim 1, characterized in that, In step S3, compensation coefficients are extracted from the delay compensation model, injected into the time synchronization protocol, and the synchronization error correction value is determined, including: Extract compensation coefficients representing the delay compensation effect from the delay compensation model; based on the compensation coefficients, inject parameters into the time synchronization protocol in the hyperconverged architecture and adjust the clock parameters of the time synchronization protocol; based on the adjusted clock parameters, calculate the synchronization error in data exchange and generate a correction value representing the magnitude of the error, wherein the correction value is used to optimize the time synchronization accuracy.

7. The method as described in claim 6, characterized in that, In step S3, the clock signal for data exchange is calibrated according to the synchronization error correction value, and a time synchronization sequence is generated, including: Based on the synchronization error correction value, the clock signal for data exchange is calibrated to generate a calibrated clock signal sequence. If the synchronization accuracy of the calibrated clock signal sequence is lower than a preset target threshold, the delay compensation model is iteratively updated and the compensation coefficients are regenerated. Based on the updated compensation coefficients, the clock signal sequence is adjusted to generate the final time synchronization sequence.

8. The method as described in claim 1, characterized in that, Step S4 includes: Based on the time synchronization sequence, a feedback signal characterizing the delay management effect is generated; the feedback signal is transmitted back to the environmental sensor to adjust the data acquisition frequency of the environmental sensor; based on the adjusted data acquisition frequency, the changing trend of the delay distribution dataset is analyzed to determine the stability of the overall system; based on the stability judgment result, a stable framework characterizing dynamic delay management is generated, and land-sea synchronous communication is performed based on the stable framework.

9. A hyperconverged architecture land-sea communication synchronization system for implementing the method as described in any one of claims 1-8, characterized in that, The system includes: The clustering determination unit is used to acquire the time delay distribution dataset of submarine optical cables, generate a time delay distribution map, extract time delay features based on the time delay distribution map, and determine the cluster groups of non-normalized time delay features. The model generation unit is used to obtain temperature gradient data based on the cluster group, predict the dynamic change trend of the time delay gradient, and obtain a time delay gradient prediction sequence; adjust the time delay compensation parameters based on the time delay gradient prediction sequence, and generate a time delay compensation model. The sequence generation unit is used to extract compensation coefficients from the delay compensation model, inject them into the time synchronization protocol, determine the synchronization error correction value, calibrate the clock signal for data exchange according to the synchronization error correction value, and generate a time synchronization sequence. The framework generation unit is used to generate feedback signals based on the time synchronization sequence, adjust the data acquisition frequency, generate a stable time delay management framework, and perform land-sea synchronous communication based on the time delay management framework.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the method as described in any one of claims 1-8.

Images