Forward error correction method based on link state self-adaption, medium and equipment

Abstract

The invention relates to the technical field of data coding, and particularly provides a forward error correction method based on link state self-adaption, a medium and equipment, and the method can comprise the steps: predicting a future packet loss rate based on collected performance index data of a WAN link; wherein the performance index data comprises a real-time packet loss rate, one-way delay data, jitter data and bandwidth data; obtaining a redundancy proportion of forward error correction coding through the future packet loss rate; wherein the redundancy proportion is a ratio of a redundant data packet to an original data packet; and encoding the data stream of the WAN link by using the encoding parameter corresponding to the redundancy ratio to obtain encoded data. According to the embodiment of the invention, the coding mode can be dynamically adjusted, and the data protection reliability is improved.

Description

Technical Field

[0001] This application relates to the field of data coding technology, and more specifically, to a method, medium, and device for forward error correction based on link state adaptation. Background Technology

[0002] Forward Error Correction (FEC) is a technique that adds redundant packets to the original data packets, enabling the receiving end to recover the original data without retransmission when encountering a certain degree of packet loss, thereby effectively reducing retransmission latency and jitter. In SD-WAN, data streams are typically transmitted through multiple physical or logical links, which can easily lead to problems such as packet loss and latency during transmission.

[0003] Currently, to avoid packet loss in SD-WAN, FEC (Framework Error Correction) technology is used for packet loss recovery. Traditional FEC technology typically pre-sets a fixed redundancy ratio, which is used to add redundant packets regardless of the actual packet loss situation. While this approach is relatively simple to implement and has low computational overhead, it lacks flexibility and provides insufficient data protection when the network environment quality is poor.

[0004] Therefore, how to provide a more flexible forward error correction method based on link state adaptation has become an urgent technical problem to be solved. Summary of the Invention

[0005] The purpose of some embodiments of this application is to provide a method, medium, and device for forward error correction based on link state adaptation. The technical solutions of the embodiments of this application can adaptively adjust the redundancy ratio of forward error correction according to the network environment, which is highly flexible, improves data protection, and has high reliability.

[0006] In a first aspect, some embodiments of this application provide a forward error correction method based on link state adaptation, comprising: predicting future packet loss rate based on collected WAN link performance index data; wherein the performance index data includes real-time packet loss rate, one-way latency data, jitter data, and bandwidth data; obtaining the redundancy ratio of forward error correction coding through the future packet loss rate; wherein the redundancy ratio is the ratio of redundant data packets to original data packets; and encoding the data stream of the WAN link using the coding parameters corresponding to the redundancy ratio to obtain encoded data.

[0007] Some embodiments of this application predict future packet loss rates using collected performance metrics data, adaptively adjust the redundancy ratio of the encoding based on the future packet loss rate, and encode the data stream using the encoding parameters corresponding to this redundancy ratio to obtain encoded data. These embodiments can adaptively adjust the redundancy ratio of forward error correction according to the network environment, offering high flexibility, improved data protection, and higher reliability.

[0008] In some embodiments, predicting the future packet loss rate based on the collected WAN link performance index data includes: preprocessing the performance index data to obtain time series data; and inputting the time series data into a pre-trained time series prediction model to obtain the future packet loss rate.

[0009] Some embodiments of this application obtain the future packet loss rate by preprocessing time series data of performance index data and a time series prediction model, thereby achieving effective and accurate prediction of the packet loss rate.

[0010] In some embodiments, the preprocessing of the performance index data to obtain time series data includes: smoothing the real-time packet loss rate and historical packet loss rate to obtain the average packet loss rate; obtaining the time series data through key features in the average packet loss rate and historical index data; wherein the key features include the mean and variance of the historical packet loss rate, the jitter peak value of the jitter data, and the delay trend data of the one-way delay data.

[0011] Some embodiments of this application determine time series data by smoothing performance index data and extracting features from historical performance data, thereby achieving data preprocessing operations.

[0012] In some embodiments, before inputting the time series data into a pre-trained time series prediction model, the method further includes: determining that the confidence level of the time series prediction model meets a preset confidence threshold.

[0013] Some embodiments of this application perform subsequent operations when the confidence level of the time series prediction model meets a preset confidence threshold, thereby improving the accuracy of packet loss rate prediction.

[0014] In some embodiments, the method further includes: if the confidence level of the time series prediction model does not meet the preset confidence threshold, or if the performance index data is abnormal, then the average packet loss rate of the performance index data is used as the future packet loss rate.

[0015] In some embodiments of this application, when the confidence level of the time series prediction model does not meet the preset confidence threshold or the data is abnormal, the average packet loss rate is selected as the future packet loss rate to provide data support for the subsequent calculation of the redundancy ratio.

[0016] In some embodiments, predicting the future packet loss rate based on the collected WAN link performance index data includes: when the future packet loss rate is the predicted packet loss rate at time i+j, comparing the predicted packet loss rate with the current packet loss rate at time i to determine the future packet loss rate; where i and j are both positive integers.

[0017] Some embodiments of this application determine the future packet loss rate by comparing the packet loss rate at different times, thereby avoiding frequent adjustments to the redundancy ratio adjustment strategy and ensuring the stability of the adjustment strategy.

[0018] In some embodiments, comparing the predicted packet loss rate with the current packet loss rate at time i to determine the future packet loss rate includes: if it is determined that the time interval between time i+j and time i is greater than a preset time interval, and the rate of change between the predicted packet loss rate and the current packet loss rate is greater than a preset threshold, then the predicted packet loss rate is used as the future packet loss rate; otherwise, the current packet loss rate is used as the future packet loss rate.

[0019] Some embodiments of this application determine the future packet loss rate by comparing the packet loss rate at different times with a preset time interval and a rate of change, thereby preventing instability caused by frequent adjustments to the redundancy ratio adjustment strategy.

[0020] In some embodiments, obtaining the redundancy ratio of forward error correction coding through the future packet loss rate includes: calculating the redundancy ratio by combining the future packet loss rate and preset parameters with a redundancy ratio range; wherein the redundancy ratio range includes an upper limit and a lower limit value that the redundancy ratio satisfies, and the preset parameters include a scaling factor and a constant bias.

[0021] Some embodiments of this application determine the redundancy ratio by using a redundancy ratio range, future packet loss rate, and preset parameters, thereby achieving effective and accurate acquisition of the redundancy ratio.

[0022] In some embodiments, before predicting future packet loss rate based on collected WAN link performance metrics data, the method further includes: monitoring the performance metrics data using active packet probing and deep packet analysis.

[0023] Some embodiments of this application achieve effective monitoring and acquisition of performance index data through active and passive detection technologies.

[0024] Secondly, some embodiments of this application provide a forward error correction device based on link state adaptive analysis, comprising: a prediction module, configured to predict future packet loss rate based on collected WAN link performance index data; wherein the performance index data includes real-time packet loss rate, one-way latency data, jitter data, and bandwidth data; an acquisition module, configured to acquire the redundancy ratio of forward error correction coding based on the future packet loss rate; wherein the redundancy ratio is the ratio of redundant data packets to original data packets; and an encoding module, configured to encode the data stream of the WAN link using encoding parameters corresponding to the redundancy ratio to obtain encoded data.

[0025] Thirdly, some embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, can implement the method described in any embodiment of the first aspect.

[0026] Fourthly, some embodiments of this application provide an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, can implement the method as described in any embodiment of the first aspect.

[0027] Fifthly, some embodiments of this application provide a computer program product, the computer program product including a computer program, wherein the computer program, when executed by a processor, can implement the method described in any embodiment of the first aspect. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of some embodiments of this application, the accompanying drawings used in some embodiments of this application will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 System diagrams for forward error correction based on link state adaptation provided for some embodiments of this application; Figure 2 One of the flowcharts for a forward error correction method based on link state adaptation provided for some embodiments of this application; Figure 3 The second flowchart of a forward error correction method based on link state adaptation provided for some embodiments of this application; Figure 4 A block diagram of a device for forward error correction based on link state adaptation provided for some embodiments of this application; Figure 5A schematic diagram of an electronic device provided for some embodiments of this application. Detailed Implementation

[0030] The technical solutions of some embodiments of this application will now be described with reference to the accompanying drawings.

[0031] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0032] In related technologies, Software-Defined Wide Area Network (SD-WAN) separates the network control plane from the forwarding plane and uses a centralized controller for unified management, enabling intelligent and flexible management of hybrid links (such as MPLS, broadband internet, 4G / 5G). In SD-WAN, data flows are typically transmitted through multiple physical or logical links to improve reliability and performance. However, the inherent instability and unpredictability of public network links (such as the internet) often lead to packet loss, latency, and jitter, severely impacting the user experience of critical applications such as VoIP, video conferencing, and virtual desktops.

[0033] Forward Error Correction (FEC) is a technique that adds redundant packets to the original data packets, enabling the receiver to recover the original data without retransmission when encountering a certain degree of packet loss, thereby effectively reducing retransmission latency and jitter. Traditional FEC schemes typically use statically configured redundancy ratios or adjustment mechanisms based on simple feedback, which have significant drawbacks: Static FEC is inflexible: This scheme pre-sets a fixed redundancy (e.g., 20%), continuously adding redundant packets according to this ratio regardless of actual network packet loss. However, a fixed redundancy ratio can easily lead to bandwidth waste when link quality is good, and insufficient protection when link quality is poor, failing to adapt to dynamically changing network environments. In adaptive FEC based on real-time feedback, the receiver (decoding side) monitors packet loss and feeds the information back to the sender. The sender dynamically adjusts the number of redundant packets in the next sending window based on the feedback packet loss information. However, the adjustment mechanism based on receiver feedback has feedback delays, cannot cope with rapid deterioration of network conditions, and requires receiver support, limiting deployment. Some schemes pre-establish multiple tunnels with different redundancies and switch between them. After real-time detection of network quality, the most suitable FEC tunnel is selected for data transmission based on the packet loss rate. This scheme requires additional tunnel resources and maintenance overhead. Most existing schemes only react to historical or current network conditions, lacking the ability to predict short-term changes in link quality and making it difficult to achieve forward-looking adjustments.

[0034] In view of this, some embodiments of this application provide a forward error correction method based on link state adaptation. This method dynamically adjusts the forward error correction coding strategy based on real-time perception and prediction of network link states to optimize network transmission quality, improve bandwidth utilization, and adapt to dynamically changing network environments. Specifically, the method predicts future packet loss rates using collected WAN link performance data, determines the redundancy ratio of forward error correction coding based on the future packet loss rate, and finally encodes the data stream using the coding parameters corresponding to the redundancy ratio to obtain encoded data. The embodiments of this application can adapt to dynamically changing network environments, realizing a shift from passive response to proactive prevention. While ensuring network transmission reliability, it significantly improves bandwidth utilization, and is particularly suitable for services sensitive to latency and packet loss, such as video conferencing and VoIP.

[0035] The following is in conjunction with the appendix Figure 1 The overall structure of a forward error correction system based on link state adaptation provided by some embodiments of this application is illustrated by way of example.

[0036] like Figure 1 As shown, some embodiments of this application provide a system diagram for forward error correction based on link state adaptation. This forward error correction system based on link state adaptation (also referred to as an SD-WAN system) may include an SD-WAN central controller and an SD-WAN edge device (CPE). The SD-WAN system includes a link state monitoring module, a prediction and decision engine, an FEC codec module, and a policy configuration and management interface.

[0037] Specifically, the link status monitoring module is located on the SD-WAN edge device and is responsible for continuously collecting performance data (i.e., performance metric data) of each WAN link. It mainly includes an active probing submodule and a passive analysis submodule. Active probing includes probing link quality through conventional physical link quality detection methods, while the passive analysis submodule mainly analyzes application service data packets through the DPI analysis engine to evaluate application quality.

[0038] The prediction and decision engine can be located in the SD-WAN centralized controller (for global optimization) or distributed on edge devices (for low latency and fast response). This engine includes a data smoothing unit, a machine learning prediction unit, and a redundancy ratio calculation unit. It receives monitoring data, performs predictions and calculations, and generates FEC policy instructions for each WAN link.

[0039] The FEC encoding / decoding module, located on the SD-WAN edge device, is responsible for real-time FEC encoding and encapsulation of the transmitted data stream according to the FEC policy issued by the prediction and decision engine, as well as decoding and reconstructing the received data stream. This module needs to support rapid dynamic configuration of the (n, k) parameters (i.e., encoding parameters).

[0040] The policy configuration and management interface, as part of the SD-WAN centralized controller, provides administrators with a graphical interface for configuring algorithm parameters and visually displaying link status, prediction results, and the application effect of FEC policies.

[0041] The following is in conjunction with the appendix Figure 2 This application provides an exemplary implementation of a forward error correction process based on link state adaptation performed by an SD-WAN edge device, as illustrated in some embodiments of this application.

[0042] Please see the appendix Figure 2 , Figure 2 A flowchart of a link-state adaptive forward error correction method is provided for some embodiments of this application. This link-state adaptive forward error correction method may include: S210, based on the collected WAN link performance index data, predict the future packet loss rate; wherein, the performance index data includes real-time packet loss rate, one-way latency data, jitter data, and bandwidth data.

[0043] For example, in a specific embodiment of this application, after the link status monitoring module collects the performance data of the WAN link, the prediction and decision engine predicts and determines the future packet loss rate. The performance data includes, but is not limited to: real-time packet loss rate, one-way delay data, jitter data, and available bandwidth (as a specific example of bandwidth data).

[0044] In some embodiments of this application, before executing S210, the forward error correction method based on link state adaptation may further include: monitoring the performance index data using active probing data packets and deep packet analysis.

[0045] For example, in a specific embodiment of this application, the link status monitoring module obtains the data through a combination of two methods: active probing, such as sending specially crafted UDP probe packets with timestamps (as a specific example of active probe data packets), and passive deep packet analysis (such as deep packet detection DPI statistics of service flow characteristics). The collected performance data will be stamped with high-precision timestamps and tagged with link identifiers.

[0046] In some embodiments of this application, S210 may include: S211, preprocess the performance index data to obtain time series data.

[0047] For example, in a specific embodiment of this application, to avoid frequent policy oscillations caused by instantaneous network jitter, the collected real-time raw indicator data (as a specific example of performance indicator data) is preprocessed. After the link state monitoring module collects performance data, the preprocessed time series data is provided to the prediction and decision engine.

[0048] In some embodiments of this application, S211 may include: smoothing the real-time packet loss rate and the historical packet loss rate to obtain an average packet loss rate; obtaining the time series data through key features in the average packet loss rate and historical index data; wherein, the key features include the mean and variance of the historical packet loss rate, the jitter peak value of the jitter data, and the delay trend data of the one-way delay data.

[0049] For example, in a specific embodiment of this application, for highly volatile indicators such as the instantaneous packet loss rate (as a specific example of the real-time packet loss rate), an exponentially weighted moving average (EWMA) method is used for smoothing to obtain a stable and more representative average packet loss rate L_avg, thereby eliminating short-term fluctuations. The formula for EWMA is: L_avg(t) = α L_current(t) + (1 - α) L_avg(t-1). Here, L_current(t) is the instantaneous packet loss rate measured at time t, L_avg(t-1) is the average value at the previous time step, and α is a smoothing factor (0 < α < 1), which can be dynamically adjusted according to network stability. Then, key features are extracted from historical indicator data, such as the mean and variance of the packet loss rate, jitter peaks, and latency trends. These features are then combined with the results of the above smoothing process to generate time series data, providing input for subsequent time series prediction models.

[0050] S212, The time series data is input into a pre-trained time series prediction model to obtain the future packet loss rate.

[0051] For example, in a specific embodiment of this application, the time series data obtained above is input into a time series prediction model to predict the link packet loss rate L_pred within a short-term time window (such as the next 1-3 RTTs) (as a specific example of the future packet loss rate). Here, RTT stands for Round-Trip Time.

[0052] During the training phase of the time series prediction model, lightweight machine learning models are used, such as online learning time series prediction models, like ARIMA or small LSTM networks. ARIMA stands for Autoregressive Integrated Moving Average, and its core function is time series prediction; LSTM stands for Long Short-Term Memory, and its core function is also time series prediction.

[0053] During the training phase, the model's input consists of historical time-series data, including smoothed packet loss rate sequences, latency trends, and jitter values. The model output is the predicted packet loss rate L_pred for a future period. Initially, the model can be pre-trained using historical data, and during operation, it continuously learns and fine-tunes parameters online using the latest monitored performance data (e.g., online updates using stochastic gradient descent) to adapt to changes in network characteristics. SGD, or Stochastic Gradient Descent, is designed to minimize the model's prediction error.

[0054] In some embodiments of this application, before executing S212, the forward error correction method based on link state adaptation may further include: determining that the confidence level of the time series prediction model meets a preset confidence threshold.

[0055] For example, in a specific embodiment of this application, the data source is intelligently selected based on the confidence level of the time series prediction model. If the confidence level of the time series prediction model is greater than a preset confidence threshold, then the predicted L_pred is used to continue executing S212. The preset confidence threshold can be flexibly set according to the actual application scenario.

[0056] In some embodiments of this application, the forward error correction method based on link state adaptation may further include: if the confidence level of the time series prediction model does not meet the preset confidence threshold, or if the performance index data is abnormal, then the average packet loss rate of the performance index data shall be used as the future packet loss rate.

[0057] For example, in a specific embodiment of this application, if the confidence level of the time series prediction model is not greater than a preset confidence threshold, it indicates that the time series prediction model is unusable. Alternatively, if an anomaly is detected in the performance data, it also indicates that the time series prediction model is unusable. If unusable, the current average packet loss rate L_avg (degradation mode) is used as L_pred to execute S212 to ensure system robustness and basic functional implementation.

[0058] In some embodiments of this application, S210 may further include: when the future packet loss rate is the predicted packet loss rate at time i+j, comparing the predicted packet loss rate with the current packet loss rate at time i to determine the future packet loss rate; wherein i and j are both positive integers.

[0059] For example, in a specific embodiment of this application, to avoid frequent switching of the FEC strategy when the link state fluctuates near a certain critical value, a minimum time interval (as a specific example of a preset time interval) and a hysteresis threshold (as a specific example of a preset threshold) are introduced to determine whether the FEC adjustment strategy (i.e., the redundancy ratio) is recalculated and adjusted at time i+j. By comparing the packet loss rates at these two times, it is determined whether the future packet loss rate of the redundancy ratio at time i+j has changed compared to time i.

[0060] In some embodiments of this application, S210 may further include: if it is determined that the time interval between the (i+j)th time and the ith time is greater than a preset time interval, and the rate of change between the predicted packet loss rate and the current packet loss rate is greater than a preset threshold, then the predicted packet loss rate is used as the future packet loss rate; otherwise, the current packet loss rate is used as the future packet loss rate.

[0061] For example, in a specific embodiment of this application, when it is determined that the time interval between time i+j and time i is greater than the minimum time interval, and the rate of change of the predicted packet loss rate L_pred relative to the current packet loss rate exceeds the lag threshold (e.g., the lag threshold is 2%), the predicted packet loss rate is used as the future packet loss rate for subsequent calculations in S220; otherwise, the current packet loss rate is still used for subsequent calculations, or no calculation is needed, and the redundancy ratio calculated historically corresponding to the current packet loss rate is used directly for subsequent encoding processes. By setting the minimum time interval and the lag threshold as described above, it is possible to determine whether to adjust the current FEC parameter (i.e., the redundancy ratio), which can effectively prevent policy oscillation. After applying the policy, the system enters a waiting state for the next cycle, and then restarts monitoring, forming a closed loop.

[0062] S220, obtain the redundancy ratio of the forward error correction coding based on the future packet loss rate; wherein, the redundancy ratio is the ratio of redundant data packets to original data packets.

[0063] For example, in a specific embodiment of this application, the optimal FEC redundancy ratio R under L_pred is calculated using a prediction and decision engine. This prediction and decision engine can receive parameter configurations from the SD-WAN central controller (as a specific example of preset parameters). The principle of forward error correction coding is: (n, k) FEC coding is used (i.e., k original data packets generate nk redundant packets, for a total of n packets). The receiving end only needs to successfully receive any k packets to recover the original data. Redundancy ratio R = (n - k) / k.

[0064] In some embodiments of this application, S220 may include: calculating the future packet loss rate and preset parameters by combining the redundancy ratio range to obtain the redundancy ratio; wherein, the redundancy ratio range includes an upper limit and a lower limit value that the redundancy ratio satisfies, and the preset parameters include a scaling factor and a constant bias.

[0065] For example, in a specific embodiment of this application, the policy configuration and management interface of the SD-WAN centralized controller can configure the redundancy ratio range, scaling factor γ, and constant bias. To ensure a high probability of successful recovery, the number of redundant packets should cover the predicted packet loss scenarios and have an appropriate margin. The formula for calculating the redundancy ratio is: R = min(R_max, max(R_min,γ)). L_pred + R_min and R_max are the lower and upper limits of the preset redundancy ratio range (e.g., R_min=0%, R_max=50%), used to prevent the algorithm from failing or consuming too much bandwidth in extreme cases. The scaling factor γ is usually >= 1.0 to provide additional protection margin to cope with prediction uncertainty; constant bias. Used to provide minimal protection even with extremely low packet loss rates.

[0066] It should be noted that the scaling factor γ and the constant bias... The value of γ is not completely fixed; it can be fine-tuned based on the historical learning performance of the network state. For example, if the recent prediction deviation is large, γ can be increased appropriately to increase the protection margin.

[0067] S230, the data stream of the WAN link is encoded using the encoding parameters corresponding to the redundancy ratio to obtain encoded data.

[0068] For example, in a specific embodiment of this application, the calculated new redundancy ratio R is sent to the FEC encoder / decoder module. The FEC encoder then encodes and encapsulates the application data stream according to the (n, k) parameters corresponding to the new redundancy ratio R. This process is continuously looped to ensure that the FEC strategy always matches the current and predicted link state.

[0069] The following is in conjunction with the appendix Figure 3 The present application provides an exemplary description of the specific process of forward error correction based on link state adaptation in some embodiments of this application.

[0070] Please see the appendix Figure 3 , Figure 3 A flowchart of a forward error correction method based on link state adaptation is provided for some embodiments of this application.

[0071] The above process is illustrated below by example.

[0072] S310 preprocesses the performance metrics data of the WAN link to obtain time series data.

[0073] S320: Determine that the time series prediction model is in a usable state, input the time series data into the pre-trained time series prediction model, and obtain the future packet loss rate.

[0074] In this embodiment, it is assumed that the adjustment mechanism of the FEC strategy has already been triggered, i.e., the minimum time interval and hysteresis threshold are met. The specific implementation process can be referred to the method embodiment provided above.

[0075] S330, in conjunction with the redundancy ratio range, calculates the future packet loss rate and preset parameters to obtain the redundancy ratio.

[0076] S340 uses the coding parameters corresponding to the redundancy ratio to encode the data stream of the WAN link, thus obtaining encoded data.

[0077] For example, suppose an internet link in an enterprise's SD-WAN network carries high-definition video conferencing traffic. The administrator configures the following parameters through the policy configuration and management interface: γ = 1.2, δ = 0.05, R_min = 5%, R_max = 40%, smoothing factor α = 0.2, prediction window of 2 RTTs (approximately 100ms), lag threshold of 2%, and minimum time interval of 1s.

[0078] In Scenario 1, the link status monitoring module reports a smoothed packet loss rate L_avg = 0.2%. The time series prediction model, based on stable historical data, predicts a future packet loss rate L_pred = 0.3%. Following the R calculation formula in the above embodiment, R = 5.36%. Since this value is very close to R_min and does not change significantly, maintaining a redundancy of 5% (i.e., adding approximately one redundant packet for every 19 packets) incurs minimal overhead.

[0079] In Scenario 2, a deterioration in link quality was predicted: slight network congestion began to appear, with a slight increase in latency. After analyzing the time series data, the time series prediction model predicted that the packet loss rate might rise to L_pred = 12% within the next 100ms. The prediction and decision engine calculated R = 19.4%. The system immediately adjusted the FEC strategy to add one redundant packet for approximately every five packets (n:k ≈ 6:5). When the packet loss rate actually increased, the video conference was adequately protected, preventing buffering.

[0080] In Scenario 3, the link quality recovers after a sharp deterioration: the link experiences a severe failure, and the predicted packet loss rate L_pred spikes to 30%. R is calculated to be 41%. This value exceeds R_max = 40%, so a redundancy cap of 40% is adopted (i.e., 2 redundant packets are added for every 5 packets, n:k = 7:5) to provide maximum protection while avoiding complete bandwidth exhaustion.

[0081] When the network recovers and the predicted packet loss rate drops to L_pred = 4%, R = 9.8% is calculated. However, due to the lag mechanism, unless the predicted packet loss rate remains below (4% - 2% = 2%) or the current smoothing value also decreases significantly, the system may temporarily maintain a high level of redundancy to prevent the strategy from oscillating during the recovery period.

[0082] It is understood that the specific implementation process of S310~S340 can refer to the method implementation examples provided above. To avoid repetition, detailed descriptions are omitted here.

[0083] As can be seen from the above embodiments of this application, this application uses machine learning to predict future packet loss trends, realizing a shift from "passive response" to "proactive prevention," enabling earlier adaptation to network changes and improving user experience; it can rapidly increase protection strength before or during link degradation, effectively ensuring the smoothness of critical applications and significantly reducing latency and stuttering caused by packet loss and retransmission; it can automatically and accurately reduce redundant overhead when good link quality is predicted, avoiding unnecessary bandwidth waste, and the saved bandwidth can be used to transmit more business data or reduce bandwidth costs; through smoothing processing, hysteresis thresholds, and oscillation suppression mechanisms, it effectively avoids instability caused by frequent policy switching, ensuring the stability of the overall system performance; the system supports flexible parameter configuration and can adapt to different network environments and business needs (such as setting different optimization targets for latency-sensitive or bandwidth-sensitive services).

[0084] Please refer to Figure 4 , Figure 4 The diagram illustrates a block diagram of a link-state adaptive forward error correction apparatus provided in some embodiments of this application. It should be understood that this link-state adaptive forward error correction apparatus corresponds to the method embodiments described above and is capable of performing the various steps involved in the method embodiments. The specific functions of this link-state adaptive forward error correction apparatus can be found in the description above; detailed descriptions are omitted here to avoid repetition.

[0085] Figure 4The link-state adaptive forward error correction device includes at least one software functional module that can be stored in a memory or embedded in the device in the form of software or firmware. The device includes: a prediction module 410, used to predict the future packet loss rate based on collected WAN link performance indicator data; wherein the performance indicator data includes real-time packet loss rate, one-way latency data, jitter data, and bandwidth data; an acquisition module 420, used to acquire the redundancy ratio of the forward error correction coding based on the future packet loss rate; wherein the redundancy ratio is the ratio of redundant data packets to original data packets; and an encoding module 430, used to encode the data stream of the WAN link using encoding parameters corresponding to the redundancy ratio to obtain encoded data.

[0086] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the device described above can be referred to the corresponding process in the aforementioned method, and will not be elaborated further here.

[0087] Some embodiments of this application also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, can perform the operation of any of the methods corresponding to the methods provided in the above embodiments.

[0088] Some embodiments of this application also provide a computer program product, which includes a computer program, wherein when the computer program is executed by a processor, it can implement the operation of any of the methods corresponding to the above embodiments provided in the above embodiments.

[0089] like Figure 5 As shown, some embodiments of this application provide an electronic device 500, which includes a memory 510, a processor 520, and a computer program stored in the memory 510 and executable on the processor 520. When the processor 520 reads the program from the memory 510 via a bus 530 and executes the program, it can implement the methods of any of the above embodiments.

[0090] Processor 520 can process digital signals and can include various computing architectures. For example, it can be a complex instruction set computer architecture, a reduced instruction set computer architecture, or an architecture that implements multiple instruction set combinations. In some examples, processor 520 can be a microprocessor.

[0091] The memory 510 can be used to store instructions executed by the processor 520 or data related to the execution of instructions. These instructions and / or data may include code for implementing some or all of the functions of one or more modules described in the embodiments of this application. The processor 520 of this disclosure embodiment can be used to execute the instructions in the memory 510 to implement the methods shown above. The memory 510 includes dynamic random access memory, static random access memory, flash memory, optical memory, or other memories well known to those skilled in the art.

[0092] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application. It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0093] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0094] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A forward error correction method based on link state adaptation, characterized in that, include: Based on the collected WAN link performance metrics data, the future packet loss rate is predicted; wherein, the performance metrics data includes real-time packet loss rate, one-way latency data, jitter data, and bandwidth data; The redundancy ratio of the forward error correction coding is obtained by using the future packet loss rate; wherein the redundancy ratio is the ratio of redundant data packets to original data packets. The data stream of the WAN link is encoded using the encoding parameters corresponding to the redundancy ratio to obtain encoded data.

2. The method as described in claim 1, characterized in that, The prediction of future packet loss rate based on the collected WAN link performance metrics data includes: The performance index data is preprocessed to obtain time series data; The time series data is input into a pre-trained time series prediction model to obtain the future packet loss rate.

3. The method as described in claim 2, characterized in that, The preprocessing of the performance index data to obtain time series data includes: The real-time packet loss rate and historical packet loss rate are smoothed to obtain the average packet loss rate; The time series data is obtained by using the key features in the average packet loss rate and historical indicator data; wherein, the key features include the mean and variance of the historical packet loss rate, the jitter peak value of the jitter data, and the delay trend data of the one-way delay data.

4. The method as described in claim 2 or 3, characterized in that, Before inputting the time series data into the pre-trained time series prediction model, the method further includes: The confidence level of the time series prediction model is determined to meet the preset confidence threshold.

5. The method as described in claim 4, characterized in that, The method further includes: If the confidence level of the time series prediction model does not meet the preset confidence threshold, or if the performance index data is abnormal, then the average packet loss rate of the performance index data will be used as the future packet loss rate.

6. The method according to any one of claims 1-3, characterized in that, The prediction of future packet loss rate based on the collected WAN link performance metrics data includes: When the future packet loss rate is the predicted packet loss rate at time i+j... The predicted packet loss rate is compared with the current packet loss rate at time i to determine the future packet loss rate; where i and j are both positive integers.

7. The method as described in claim 6, characterized in that, The step of comparing the predicted packet loss rate with the current packet loss rate at time i to determine the future packet loss rate includes: If it is determined that the time interval between the (i+j)th time and the ith time is greater than a preset time interval, and the rate of change between the predicted packet loss rate and the current packet loss rate is greater than a preset threshold, then the predicted packet loss rate is taken as the future packet loss rate; otherwise, the current packet loss rate is taken as the future packet loss rate.

8. The method according to any one of claims 1-3, characterized in that, The step of obtaining the redundancy ratio of the forward error correction coding through the future packet loss rate includes: The redundancy ratio is obtained by calculating the future packet loss rate and preset parameters by combining the redundancy ratio range; wherein, the redundancy ratio range includes the upper limit and lower limit values ​​that the redundancy ratio satisfies, and the preset parameters include a scaling factor and a constant bias.

9. The method according to any one of claims 1-3, characterized in that, Before predicting future packet loss rate based on collected WAN link performance metrics data, the method further includes: The performance metrics data are monitored using active packet probing and deep packet analysis.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein the computer program is executed by a processor to perform the method as described in any one of claims 1-9.

11. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the computer program is executed by the processor to perform the method as described in any one of claims 1-9.

12. A computer program product, characterized in that, The computer program product includes a computer program, wherein the computer program is executed by a processor to perform the method as described in any one of claims 1-9.

Images