Network load-aware communication traffic distribution method and system

Abstract

The application discloses a communication traffic distribution method and system based on network load perception, relates to the technical field of network communication, and comprises the following steps: collecting multi-dimensional load indexes of each target node in a network, performing weighted fusion processing through a sliding time window, and generating a comprehensive load score sequence of each target node; based on historical cumulative data of the comprehensive load scores of each target node, a time series decomposition algorithm is adopted to extract short-term load change trend characteristics of each node; a traffic distribution decision model is established in combination with the comprehensive load score sequence and the trend characteristics, and optimal traffic distribution weights of each target node are obtained; according to the optimal weights and type characteristics of to-be-distributed traffic, traffic differentiation distribution is realized; and meanwhile, the distribution execution state is monitored, and when an exception is detected, the distribution process is automatically re-executed. The application realizes full-process intelligent optimization from load perception to differentiation distribution, and improves the system load balancing effect and running stability.

Description

Technical Field

[0001] This invention relates to the field of network communication technology, and in particular to a communication traffic distribution method and system based on network load awareness. Background Technology

[0002] With the rapid growth of internet applications, network traffic distribution technology has become a key technology for ensuring system performance. Existing traffic distribution methods mainly include round-robin algorithms, weighted round-robin algorithms, and load-based dynamic allocation algorithms. Round-robin algorithms distribute requests to nodes sequentially according to a preset order, offering simplicity but failing to consider differences in node performance. Weighted round-robin algorithms adjust traffic allocation ratios by setting static weight values, which can balance the load on nodes with different performance levels to some extent. Load-based dynamic allocation algorithms dynamically adjust traffic allocation strategies based on the real-time load status of nodes. Regarding load assessment, existing methods typically use single or a few indicators such as CPU utilization, memory utilization, and the number of connections to measure node load levels.

[0003] However, existing traffic distribution methods typically employ a uniform distribution strategy when handling different types of traffic, failing to fully consider the characteristics and priority requirements of different business traffic, thus unable to meet differentiated service needs. Furthermore, these methods also have limitations in terms of the comprehensiveness of load assessment and dynamic response capabilities. These technical issues lead to low resource utilization and decreased service quality in existing traffic distribution systems under high concurrency and multi-business scenarios. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a communication traffic distribution method and system based on network load awareness to solve the technical problems of unbalanced load distribution, delayed system response, and degraded service quality.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a communication traffic distribution method based on network load awareness, comprising:

[0008] Collect multi-dimensional load metrics for each target node in the network;

[0009] A sliding time window is used to perform weighted fusion processing on multi-dimensional load indicators to generate a comprehensive load score sequence for each target node;

[0010] Based on the historical cumulative data of the comprehensive load score of each target node, the short-term load change trend characteristics of each target node are extracted using the time series decomposition algorithm;

[0011] A traffic distribution decision model is established based on the comprehensive load score sequence and short-term load change trend characteristics to obtain the optimal traffic allocation weight for each target node.

[0012] Based on the optimal traffic allocation weight and combined with the type characteristics of the traffic to be distributed, the traffic is distributed to the corresponding target nodes in a differentiated manner.

[0013] Monitor the traffic distribution execution status of each target node, and re-execute the traffic distribution process when an abnormality in distribution execution or abnormal node response is detected.

[0014] As a preferred embodiment of the network load-aware communication traffic distribution method of the present invention, the method includes: weighted fusion processing of multi-dimensional load indicators using a sliding time window, which includes:

[0015] Standardize the multidimensional load indicators;

[0016] Based on the changing characteristics of multidimensional load indicators, an adaptive weighting calculation method is used to determine the weight coefficients of each load indicator.

[0017] A two-layer sliding time window, including short-time and long-time windows, is used to weight and fuse the standardized multi-dimensional load indicators to generate a comprehensive load score sequence for each target node.

[0018] Based on the fluctuation characteristics of the comprehensive load score sequence, the length parameter of the double-layer sliding time window is dynamically adjusted.

[0019] As a preferred embodiment of the network load-aware communication traffic distribution method of the present invention, the method includes: extracting short-term load change trend features of each target node using a time series decomposition algorithm, including:

[0020] Obtain historical cumulative data of the comprehensive load score of each target node and generate a historical load score sequence;

[0021] Based on the load change cycle characteristics of historical load score sequences, the trend smoothing window size for STL decomposition is determined, and STL time series decomposition is performed to obtain trend components and residual components.

[0022] Linear regression fitting is performed on the trend component, and the regression slope is calculated as the load growth rate.

[0023] Statistical characteristic parameters are calculated based on residual components, and the load fluctuation amplitude is obtained by combining autocorrelation analysis.

[0024] Based on the load growth rate and load fluctuation amplitude, short-term load change trend characteristics of each target node are generated.

[0025] As a preferred embodiment of the network load-aware communication traffic distribution method of the present invention, the process of constructing the traffic distribution decision model includes:

[0026] Based on the comprehensive load score sequence of each target node, the mean stability and fluctuation severity of the comprehensive load score sequence are analyzed, and a load stability evaluation index is constructed.

[0027] By combining the load growth rate and load fluctuation amplitude in the short-term load change trend characteristics, a load prediction reliability index for each target node is constructed through trend weighting.

[0028] Obtain the hardware resource configuration information of each target node, and calculate the node resource carrying capacity score based on the resource capacity limit and the current load level;

[0029] A multi-factor decision matrix is ​​established, with load stability evaluation index, load prediction reliability index and node resource carrying capacity score as decision factors. The comprehensive performance evaluation value of each target node is obtained through dynamic weighted fusion algorithm.

[0030] Based on the comprehensive performance evaluation value and the current system load distribution status, the optimal traffic allocation weight for each target node is determined using a dual-mode weight allocation algorithm based on prediction reliability.

[0031] As a preferred embodiment of the network load-aware communication traffic distribution method of the present invention, the method includes: distributing traffic differentially to the corresponding target nodes.

[0032] Extract features from the traffic to be distributed to obtain traffic type and priority identifiers;

[0033] The service compatibility of each target node with the current traffic is calculated based on the traffic type identifier, and a set of candidate nodes that meet the service compatibility standard is selected according to the preset compatibility threshold.

[0034] The optimal traffic allocation weights for candidate nodes are adjusted based on traffic priority identifiers to generate the final allocation weights.

[0035] Based on the final allocation weights, a weighted round-robin algorithm is used to distribute traffic to the corresponding target nodes and record distribution statistics.

[0036] As a preferred embodiment of the network load-aware communication traffic distribution method of the present invention, the method includes: determining the optimal traffic allocation weight for each target node using a dual-mode weight allocation algorithm based on prediction reliability, which includes:

[0037] Obtain the load prediction reliability index of each target node, and calculate the dual-mode fusion ratio based on the load prediction reliability index of each target node.

[0038] Based on the comprehensive performance evaluation value of each target node and the short-term load change trend characteristics, calculate the forward-looking weight allocation coefficient;

[0039] Based on the comprehensive performance evaluation value of each target node and the current system load distribution status, calculate the responsive weight allocation coefficient;

[0040] The forward-looking weight allocation coefficient and the responsive weight allocation coefficient are weighted and merged according to the dual-mode fusion ratio to generate the optimal traffic allocation weight for each target node.

[0041] As a preferred embodiment of the network load-aware communication traffic distribution method of the present invention, the multi-dimensional load indicators include CPU utilization, memory usage, network bandwidth utilization, I / O response latency, and number of connections.

[0042] Secondly, the present invention provides a network load-aware communication traffic distribution system, comprising:

[0043] The data acquisition module is used to collect multi-dimensional load metrics of each target node in the network;

[0044] The load assessment module is used to perform weighted fusion processing on multi-dimensional load indicators using a sliding time window to generate a comprehensive load score sequence for each target node.

[0045] The trend analysis module is used to extract the short-term load change trend characteristics of each target node based on the historical cumulative data of the comprehensive load score of each target node and the time series decomposition algorithm.

[0046] The decision modeling module is used to establish a traffic distribution decision model based on the comprehensive load score sequence and short-term load change trend characteristics, and obtain the optimal traffic allocation weight for each target node.

[0047] The traffic distribution module is used to distribute traffic to the corresponding target nodes in a differentiated manner based on the optimal traffic allocation weight and the type characteristics of the traffic to be distributed;

[0048] The monitoring and feedback module is used to monitor the traffic distribution execution status of each target node. When an abnormality in the distribution execution or an abnormality in the node response is detected, the traffic distribution process is re-executed.

[0049] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the network load-aware communication traffic distribution method of the first aspect of the present invention.

[0050] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the network load-aware communication traffic distribution method of the first aspect of the present invention.

[0051] The beneficial effects of this invention are as follows: This invention extracts trend and residual components using the STL time series decomposition algorithm, achieving precise quantification of load growth rate and fluctuation amplitude. This provides a reliable basis for load trend prediction for traffic distribution decisions, improving the accuracy of load prediction and the foresight of traffic allocation. By constructing a multi-dimensional comprehensive evaluation system and a dual-mode weight allocation mechanism, adaptive intelligent traffic allocation is achieved, improving the system's load balancing effect and robustness. Through traffic feature extraction and business adaptability evaluation, combined with a priority adjustment mechanism, a differentiated traffic distribution strategy is implemented, improving the accuracy of traffic distribution and business matching. In summary, this invention achieves intelligent optimization of the entire process from load perception and collection to differentiated distribution, comprehensively improving the system's load balancing effect and operational stability. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart of a network load-aware communication traffic distribution method.

[0054] Figure 2 This is a flowchart illustrating the short-term load change trend feature extraction process for a network load-aware communication traffic distribution method.

[0055] Figure 3 A flowchart for constructing a traffic distribution decision model for a network load-aware communication traffic distribution method.

[0056] Figure 4 This is a schematic diagram of a network load-aware communication traffic distribution system. Detailed Implementation

[0057] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0058] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0059] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0060] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a communication traffic distribution method based on network load awareness, the flowchart of which is shown below. Figure 1 As shown, it includes the following steps:

[0061] S1: Collect multi-dimensional load metrics for each target node in the network.

[0062] The multidimensional load metrics include CPU utilization, memory usage, network bandwidth utilization, I / O response latency, and number of connections.

[0063] It should be noted that the selection of multi-dimensional load metrics comprehensively considers the key dimensions of node performance. Among them, CPU utilization and memory usage reflect the basic computing resource status of the node, network bandwidth utilization reflects data transmission capability, I / O response latency characterizes storage system performance, and the number of connections reflects the concurrent processing load of the node. These five metrics can comprehensively and accurately evaluate the real-time load level of the target node.

[0064] S2: A sliding time window is used to perform weighted fusion processing on multi-dimensional load indicators to generate a comprehensive load score sequence for each target node.

[0065] Specifically, step S2 includes:

[0066] S2.1: Standardize the multidimensional load indicators.

[0067] The standardization process employs a segmented standardization method, using linear standardization for percentage indicators and robust standardization for numerical indicators to eliminate dimensional differences between different indicators.

[0068] S2.2: Based on the changing characteristics of multidimensional load indicators, an adaptive weight calculation method is used to determine the weight coefficients of each load indicator.

[0069] In one specific embodiment, the variation characteristics of each load indicator are calculated, and these characteristics are quantified by calculating the coefficient of variation of each load indicator. Based on the variation characteristics, the objective weight of each load indicator is calculated using the entropy weight method. Based on a preset indicator importance level, the subjective weight of each load indicator is determined using the analytic hierarchy process (AHP). The objective weight and subjective weight are then weighted and fused to obtain the weight coefficient of each load indicator. The indicator importance level is set by analyzing the correlation between historical load data and system performance bottlenecks.

[0070] S2.3: A two-layer sliding time window, including short-term and long-term windows, is used to weight and fuse the standardized multi-dimensional load indicators to generate a comprehensive load score sequence for each target node.

[0071] It should be noted that the short-term window is used to capture the real-time change characteristics of the load, while the long-term window is used to reflect the historical trend characteristics of the load. The comprehensive load score is calculated by weighted average of the short-term score and the long-term score.

[0072] Furthermore, the short-term window length is preferably 5-15 time units, the long-term window length is preferably 30-90 time units, and the weight ratio of the short-term score and the long-term score is preferably 0.6:0.4 to 0.8:0.2. The time unit can be seconds, minutes, or other preset sampling intervals. The dual-layer sliding window mechanism can ensure the real-time performance of load assessment while taking into account the stability of historical trends.

[0073] S2.4: Dynamically adjust the length parameter of the double-layer sliding time window based on the fluctuation characteristics of the comprehensive load score sequence.

[0074] The volatility characteristic includes the rating variance. Specifically, the dynamic adjustment process is as follows: when the rating variance is below the stability threshold, it indicates that the load change is relatively stable, and the time window length is appropriately extended to improve the stability of the rating; when the rating variance is above the volatility threshold, it indicates that the load change is relatively drastic, and the time window length is appropriately shortened to improve the response speed to load changes; when the rating variance is between the two thresholds, the current window length is kept unchanged.

[0075] In a preferred implementation, the stability threshold is determined based on the lower quartile of the historical score variance distribution, and the volatility threshold is determined based on the upper quartile of the historical score variance distribution. The adjustment range of the time window length is preferably 10%-30% of the original length. This dynamic adjustment mechanism enables the system to adaptively balance the real-time performance and stability of load assessment, thereby improving the load balancing effect.

[0076] S3: Based on the historical cumulative data of the comprehensive load score of each target node, the short-term load change trend characteristics of each target node are extracted using the time series decomposition algorithm.

[0077] Specifically, the flowchart for extracting short-term load change trend features is as follows: Figure 2 As shown, it includes:

[0078] S3.1: Obtain historical cumulative data of the comprehensive load score of each target node and generate a historical load score sequence.

[0079] It should be noted that historical accumulated data refers to the cumulative record of the comprehensive load score sequence of each target node over a period of time.

[0080] S3.2: Determine the trend smoothing window size for STL decomposition based on the load change cycle characteristics of the historical load score sequence, perform STL time series decomposition, and obtain the trend component and residual component.

[0081] This invention employs the STL time series decomposition algorithm because it can decompose historical load rating sequences into trend components and residual components. The trend component reflects the long-term direction of load change, while the residual component reflects the random fluctuation characteristics of the load, providing a data foundation for subsequent load growth rate calculation and fluctuation amplitude analysis.

[0082] In one specific embodiment, step S3.2 includes: performing periodic analysis on the historical load score series to identify the dominant period length and periodic change pattern of the historical load score series; calculating the trend smoothing window size of STL decomposition using an empirical formula based on the identified dominant period length, and setting the seasonal smoothing parameter; executing the STL time series decomposition algorithm to complete the time series decomposition through inner and outer loop iterations; extracting the trend component and residual component from the STL decomposition results, and verifying the completeness of the decomposition results.

[0083] Through the STL time series decomposition process described above, the extraction of trend components provides smooth data input for linear regression fitting, eliminating the influence of periodic fluctuations and random noise in the original sequence on slope calculation. Simultaneously, the acquisition of residual components provides pure fluctuation data for statistical feature calculation, making the quantification of load fluctuation amplitude more accurate. Furthermore, the trend smoothing window size, determined based on the dominant period length, can adaptively adjust the decomposition precision, ensuring that the trend components retain long-term variation characteristics.

[0084] S3.3: Perform linear regression fitting on the trend component and calculate the regression slope as the load growth rate.

[0085] Furthermore, a trend regression dataset is constructed with time sequence number as the independent variable and trend component value as the dependent variable. The least squares method is used to perform linear regression fitting on the trend regression dataset to establish a linear regression model between the trend component and the time sequence number. This includes: initializing the regression parameters, solving for the optimal parameters by minimizing the sum of squared residuals, and establishing a linear relationship; extracting the slope coefficient of the regression line from the linear regression model, which is the load growth rate of the target node; and performing a significance test on the regression fitting results to ensure the effectiveness of the linear regression model.

[0086] S3.4: Calculate statistical characteristic parameters based on residual components, and obtain the load fluctuation amplitude by combining autocorrelation analysis.

[0087] Furthermore, the standard deviation of the residual components is calculated as the basic fluctuation characteristic parameter; autocorrelation function analysis is performed on the residual components to calculate the autocorrelation coefficients under multiple lag orders; the correlation index of the residual components is calculated based on the autocorrelation coefficients; and the standard deviation of the residual components and the correlation index are weighted and combined to obtain the load fluctuation amplitude of the target node.

[0088] S3.5: Based on the load growth rate and load fluctuation amplitude, generate short-term load change trend characteristics for each target node.

[0089] Preferably, through the integrated processing flow of time series decomposition and statistical analysis described above, the historical load data of each target node is transformed into two quantitative indicators: load growth rate and load fluctuation amplitude. The generated short-term load change trend characteristics can simultaneously reflect the development direction and stability level of node load, providing data support for the load prediction reliability assessment in the subsequent traffic distribution decision model, and enabling the calculation of traffic allocation weights to be dynamically adjusted based on load trend prediction.

[0090] S4: Establish a traffic distribution decision model based on the comprehensive load score sequence and short-term load change trend characteristics to obtain the optimal traffic allocation weight for each target node.

[0091] Specifically, the flowchart for building the traffic distribution decision model is as follows: Figure 3 As shown, it includes:

[0092] S4.1: Based on the comprehensive load score sequence of each target node, analyze the mean stability and fluctuation severity of the comprehensive load score sequence, and construct a load stability evaluation index.

[0093] In one embodiment, the construction of the load stability evaluation index includes: calculating the arithmetic mean and weighted mean of the comprehensive load score sequence, and assessing the mean stability by the relative deviation between the two; calculating the coefficient of variation using moving variance analysis to quantify the severity of load fluctuations; identifying load mutation points through difference analysis and statistically analyzing the frequency and magnitude of mutations; calculating the mean absolute deviation between the comprehensive load score and the moving average, and obtaining the dispersion by combining the interquartile range; and weighting and fusing the above indicators using the information entropy weighting method to obtain the load stability evaluation index.

[0094] S4.2: Combining the load growth rate and load fluctuation amplitude in the short-term load change trend characteristics, construct the load prediction reliability index for each target node through trend weighting processing.

[0095] Furthermore, the construction of the load forecast reliability index includes: calculating the trend stability coefficient based on the consistency of the absolute value and direction of load growth rate; calculating the fluctuation controllability coefficient based on the relative magnitude and frequency of load fluctuations; calculating the historical forecast performance coefficient through historical forecast accuracy statistics; giving higher weight to recent forecast performance using a time decay weighting method; and obtaining the load forecast reliability index for each target node by normalizing the trend stability coefficient, fluctuation controllability coefficient, and historical forecast performance coefficient, and then using a weighted average.

[0096] S4.3: Obtain the hardware resource configuration information of each target node, and calculate the node resource carrying capacity score based on the resource capacity limit and the current load level.

[0097] The hardware resource configuration information includes the number of CPU cores, memory capacity, network bandwidth limit, storage I / O performance, and maximum concurrent connections.

[0098] In addition, the calculation process of the node resource carrying capacity score is as follows: Based on the multi-dimensional load indicators and hardware resource configuration information collected by S1, the resource utilization rate of each target node in each dimension is calculated; the resource utilization rate of each dimension is converted into the percentage of remaining available capacity; the remaining available capacity of each dimension is weighted and averaged to obtain the overall resource carrying capacity score of the node.

[0099] It should be noted that the load stability evaluation index reflects the historical performance and stability of a node's load. Nodes with higher stability are less prone to performance fluctuations when handling new traffic, which helps ensure consistent service quality. The load prediction reliability index assesses the accuracy of predicting the future load status of a node based on its load change trend characteristics. Nodes with higher reliability facilitate accurate traffic planning and capacity estimation. The resource carrying capacity score is based on the node's hardware configuration and current resource usage, directly reflecting the node's physical carrying capacity for receiving new traffic. These three dimensions comprehensively evaluate the node's overall service capabilities from the perspectives of historical stability, future predictability, and current carrying capacity, achieving accurate and efficient traffic distribution through multi-dimensional comprehensive evaluation.

[0100] S4.4: Establish a multi-factor decision matrix, using load stability evaluation index, load prediction reliability index and node resource carrying capacity score as decision factors, and obtain the comprehensive performance evaluation value of each target node through dynamic weighted fusion algorithm.

[0101] Furthermore, the calculation of the comprehensive performance evaluation value includes: constructing a multi-factor decision matrix from the three decision factors of each target node, where rows represent nodes and columns represent decision factors; normalizing each decision factor in the multi-factor decision matrix to eliminate the influence of different indicator dimensions; dynamically adjusting the weight coefficients of the three decision factors according to the current system load distribution status and the traffic distribution strategy objectives, for example, increasing the weight of the resource carrying capacity score when the system load is high, increasing the weight of the load stability evaluation index when the load fluctuation is large, and increasing the weight of the load prediction reliability index when the prediction accuracy is low; and using a weighted summation method to fuse the normalized decision factors of each node according to the dynamic weights to obtain the comprehensive performance evaluation value of each target node.

[0102] The traffic distribution strategy aims to achieve load balancing, system performance optimization, and service stability assurance. Under the load balancing strategy, load stability is given priority; under the performance optimization strategy, resource carrying capacity is given priority; and under the stability assurance strategy, prediction reliability is given priority.

[0103] S4.5: Based on the comprehensive performance evaluation value and combined with the current system load distribution status, the optimal traffic allocation weight for each target node is determined using a dual-mode weight allocation algorithm based on prediction reliability.

[0104] Specifically, step S4.5 includes:

[0105] S4.5.1: Obtain the load prediction reliability index of each target node, and calculate the dual-mode fusion ratio based on the load prediction reliability index of each target node.

[0106] Specifically, high and low confidence thresholds are set. Based on the comparison between the load prediction confidence index of each target node and the threshold, the target nodes are divided into high-confidence, medium-confidence, and low-confidence node sets. The comprehensive performance evaluation value of each target node is obtained, and the sum of the comprehensive performance evaluation values ​​of all nodes within each node set is calculated to obtain the weight coefficient of each node set. Different forward-looking preference coefficients are set for each node set, with the highest forward-looking preference coefficient for the high-confidence node set and the lowest for the low-confidence node set. The weight coefficients of each node set are weighted and calculated with the corresponding forward-looking preference coefficients to obtain the forward-looking mode fusion ratio. Based on the constraint that the sum of the dual-mode fusion ratios is 1, the responsive mode fusion ratio is calculated. Through node grouping and weighted calculation, the abstract "confidence" is transformed into a concrete "fusion ratio".

[0107] S4.5.2: Calculate the forward-looking weight allocation coefficient based on the comprehensive performance evaluation value of each target node and the short-term load change trend characteristics.

[0108] Furthermore, the comprehensive performance evaluation value of each target node is obtained as the basic weight; the load growth rate and load fluctuation amplitude are extracted from the short-term load change trend characteristics; a trend adjustment factor is calculated based on the load growth rate, and nodes with lower load growth rates are assigned higher trend adjustment factors, while nodes with higher load growth rates are assigned lower trend adjustment factors; a stability adjustment factor is calculated based on the load fluctuation amplitude, and nodes with smaller load fluctuation amplitudes are assigned higher stability adjustment factors; the comprehensive performance evaluation value of each target node is multiplied by the corresponding trend adjustment factor and stability adjustment factor to obtain the initial forward-looking weight of each node; the initial forward-looking weight of each node is normalized so that the sum of the forward-looking weight allocation coefficients of all nodes is 1.

[0109] S4.5.3: Calculate the response weight allocation coefficient based on the comprehensive performance evaluation value of each target node and the current system load distribution status.

[0110] Furthermore, the comprehensive performance evaluation value of each target node is obtained as the basic weight; the current system load distribution status is obtained, including the current load level of each node and the overall system load balance; the load adjustment factor is calculated based on the current load level of each node, assigning a higher load adjustment factor to nodes with lower current load levels and a lower load adjustment factor to nodes with higher current load levels; the balance adjustment factor is calculated based on the overall system load balance, and when the system load distribution is uneven, additional weight compensation is given to nodes with lighter loads; the comprehensive performance evaluation value of each target node is multiplied by the corresponding load adjustment factor and balance adjustment factor to obtain the initial value of the responsive weight of each node; the initial value of the responsive weight of each node is normalized so that the sum of the responsive weight allocation coefficients of all nodes is 1.

[0111] S4.5.4: The forward-looking weight allocation coefficient and the responsive weight allocation coefficient are weighted and merged according to the dual-mode fusion ratio to generate the optimal traffic allocation weight for each target node.

[0112] Specifically, the forward-looking weight allocation coefficient of each target node is multiplied by the forward-looking mode fusion ratio to obtain the forward-looking weight component; the responsive weight allocation coefficient of each target node is multiplied by the responsive mode fusion ratio to obtain the responsive weight component; and the forward-looking weight component and the responsive weight component corresponding to each node are added together to obtain the optimal traffic allocation weight of each target node.

[0113] Preferably, this invention employs a dual-mode weight allocation algorithm based on prediction reliability to dynamically balance forward-looking decision-making (based on prediction) and reactive decision-making (based on the current state) according to system state and node characteristics, thereby achieving an adaptive traffic allocation strategy. Through a prediction reliability layering mechanism and differentiated weight adjustment, the accuracy of load prediction is improved; through a dual-mode fusion mechanism, intelligent traffic allocation in complex dynamic environments is achieved, improving system load balancing performance, exhibiting good adaptability and robustness, and reducing the risk of system overload.

[0114] S5: Based on the optimal traffic allocation weight and combined with the type characteristics of the traffic to be distributed, the traffic is distributed to the corresponding target nodes in a differentiated manner.

[0115] Specifically, step S5 includes:

[0116] S5.1: Extract features from the traffic to be distributed to obtain traffic type identifier and priority identifier.

[0117] It should be noted that traffic feature extraction identifies the business type of traffic (such as HTTP, database access, file transfer, etc.) and obtains the traffic type identifier by analyzing features such as packet header information, protocol type, port number, and payload content; at the same time, it determines the priority identifier based on factors such as business importance, latency sensitivity, and bandwidth requirements, providing basic data for subsequent differentiated traffic distribution.

[0118] S5.2: Calculate the service compatibility of each target node with the current traffic based on the traffic type identifier, and select a set of candidate nodes that meet the service compatibility standard according to the preset compatibility threshold.

[0119] Furthermore, the calculation of service adaptability includes: obtaining the service processing capability configuration information of each target node, including supported service types and processing performance parameters of each service type; calculating the matching degree between the traffic type identifier and the service types supported by the node, with a matching degree of 1 for fully matched service types and a matching degree calculated based on similarity for partially matched types; calculating the performance adaptability by combining the node's service processing performance parameters and the resource requirements of the current traffic; and weightedly fusing the matching degree and performance adaptability to obtain the service adaptability of each target node to the current traffic.

[0120] In addition, the preset adaptation threshold is dynamically determined based on the system load status and service quality requirements: when the system load is low, a higher adaptation threshold is set to ensure optimal matching; when the system load is high, the adaptation threshold is appropriately lowered to expand the range of candidate nodes and ensure that traffic can be distributed in a timely manner; at the same time, the setting strategy of the adaptation threshold is adjusted based on the statistics of historical distribution effects.

[0121] Ideally, the business compatibility assessment mechanism achieves precise matching of traffic and nodes, avoiding the distribution of specific business traffic to unsuitable nodes and improving business processing efficiency. At the same time, the dynamic compatibility threshold strategy ensures the quality of business matching while taking into account the system load balancing requirements, enhancing the flexibility of traffic distribution and the overall system performance.

[0122] S5.3: Adjust the optimal traffic allocation weight of candidate nodes according to the traffic priority identifier to generate the final allocation weight.

[0123] Furthermore, weight adjustment coefficients are set for different priorities, with the adjustment coefficient for high-priority traffic being greater than 1 and the adjustment coefficient for low-priority traffic being less than 1. The performance ranking of each node in the candidate node set is obtained, and nodes with better performance are marked as preferred nodes. For high-priority traffic, the weight adjustment coefficient of preferred nodes is increased, while the weight adjustment coefficients of other nodes are decreased. For low-priority traffic, a relatively balanced weight adjustment strategy is adopted. The optimal traffic allocation weight of each candidate node is multiplied by its corresponding weight adjustment coefficient to obtain the adjusted weight value. The adjusted weight value is then normalized to ensure that the sum of the final allocated weights of all candidate nodes is 1. Through priority-differentiated weight adjustment, critical businesses are ensured to receive high-quality resources first, achieving a precise match between business importance and resource allocation, and improving overall service quality and resource utilization efficiency.

[0124] S5.4: Based on the final allocation weight, a weighted round-robin algorithm is used to distribute traffic to the corresponding target nodes and record distribution statistics.

[0125] Furthermore, traffic distribution based on the final allocation weight includes: converting the final allocation weight of each candidate node into an integer weight value, and ensuring weight accuracy through proportional scaling; establishing a weighted round-robin queue, and allocating a corresponding number of scheduling positions in the queue according to the weight value of each node; selecting target nodes sequentially according to the queue order for traffic distribution, and entering the next round of loop when the weight of a node is exhausted; recording statistical information such as the amount of traffic distributed, the number of distributions, and the response time of each node in real time; and updating the load status of nodes based on the distribution statistics to provide data support for the next traffic distribution decision.

[0126] S6: Monitor the traffic distribution execution status of each target node. When an abnormality in the distribution execution or an abnormality in the node response is detected, re-execute the traffic distribution process.

[0127] Specifically, the system collects traffic reception status, response time, and load change indicators of each target node in real time; it performs statistical analysis on the collected status indicators and calculates anomaly scores; when the anomaly score exceeds the anomaly judgment threshold for a preset number of consecutive times, it is judged as a continuous anomaly state; it selects a processing strategy according to the severity of the anomaly: for mild anomalies, only the weight of the node is adjusted; for severe anomalies, traffic distribution to the abnormal node is suspended; based on the anomaly type and historical recovery time, it dynamically determines the recovery interval and then re-executes the traffic distribution process.

[0128] The anomaly detection threshold is determined through statistical analysis of historical normal operation data, with a preset number of occurrences of 3.

[0129] This embodiment also provides a network load-aware communication traffic distribution system, as shown in the schematic diagram. Figure 4 As shown, the system includes:

[0130] The data acquisition module is used to collect multi-dimensional load metrics of each target node in the network;

[0131] The load assessment module is used to perform weighted fusion processing on multi-dimensional load indicators using a sliding time window to generate a comprehensive load score sequence for each target node.

[0132] The trend analysis module is used to extract the short-term load change trend characteristics of each target node based on the historical cumulative data of the comprehensive load score of each target node and the time series decomposition algorithm.

[0133] The decision modeling module is used to establish a traffic distribution decision model based on the comprehensive load score sequence and short-term load change trend characteristics, and obtain the optimal traffic allocation weight for each target node.

[0134] The traffic distribution module is used to distribute traffic to the corresponding target nodes in a differentiated manner based on the optimal traffic allocation weight and the type characteristics of the traffic to be distributed;

[0135] The monitoring and feedback module is used to monitor the traffic distribution execution status of each target node. When an abnormality in the distribution execution or an abnormality in the node response is detected, the traffic distribution process is re-executed.

[0136] This embodiment also provides a computer device applicable to the network load-aware communication traffic distribution method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the network load-aware communication traffic distribution method proposed in the above embodiment.

[0137] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0138] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the network load-aware communication traffic distribution method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0139] In summary, this invention extracts trend and residual components using the STL time series decomposition algorithm, achieving precise quantification of load growth rate and fluctuation amplitude. This provides a reliable basis for load trend prediction in traffic distribution decisions, improving the accuracy of load prediction and the foresight of traffic allocation. By constructing a multi-dimensional comprehensive evaluation system and a dual-mode weight allocation mechanism, adaptive intelligent traffic allocation is achieved, improving the system's load balancing effect and robustness. Through traffic feature extraction and business adaptability evaluation, combined with a priority adjustment mechanism, a differentiated traffic distribution strategy is implemented, improving the accuracy of traffic distribution and business matching. In summary, this invention achieves intelligent optimization of the entire process from load perception and collection to differentiated distribution, comprehensively improving the system's load balancing effect and operational stability.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A communication traffic distribution method based on network load awareness, characterized in that, Includes the following steps: Collect multi-dimensional load metrics for each target node in the network; The multidimensional load indicators are weighted and fused using a sliding time window to generate a comprehensive load score sequence for each target node. Based on the historical cumulative data of the comprehensive load score of each target node, the short-term load change trend characteristics of each target node are extracted using the time series decomposition algorithm; A traffic distribution decision model is established based on the comprehensive load score sequence and the short-term load change trend characteristics to obtain the optimal traffic allocation weight for each target node. Based on the optimal traffic allocation weight and combined with the type characteristics of the traffic to be distributed, the traffic is distributed to the corresponding target nodes in a differentiated manner. Monitor the traffic distribution execution status of each target node, and re-execute the traffic distribution process when an abnormality in distribution execution or abnormal node response is detected. The process of constructing the traffic distribution decision model includes: Based on the comprehensive load score sequence of each target node, the mean stability and fluctuation severity of the comprehensive load score sequence are analyzed, and a load stability evaluation index is constructed. By combining the load growth rate and load fluctuation amplitude in the short-term load change trend characteristics, a load prediction reliability index for each target node is constructed through trend weighting. Obtain the hardware resource configuration information of each target node, and calculate the node resource carrying capacity score based on the resource capacity limit and the current load level; A multi-factor decision matrix is ​​established, and the load stability evaluation index, the load prediction reliability index, and the node resource carrying capacity score are used as decision factors. The comprehensive performance evaluation value of each target node is obtained through a dynamic weighted fusion algorithm. Based on the comprehensive performance evaluation value and the current system load distribution, the optimal traffic allocation weight for each target node is determined using a dual-mode weight allocation algorithm based on prediction reliability.

2. The communication traffic distribution method based on network load awareness as described in claim 1, characterized in that, The weighted fusion processing of the multidimensional load indicators using a sliding time window includes: Standardize the multidimensional load indicators; Based on the changing characteristics of the multidimensional load indicators, an adaptive weight calculation method is used to determine the weight coefficients of each load indicator. A two-layer sliding time window, including short-time and long-time windows, is used to weight and fuse the standardized multi-dimensional load indicators to generate a comprehensive load score sequence for each target node. The length parameter of the double-layer sliding time window is dynamically adjusted based on the fluctuation characteristics of the comprehensive load score sequence.

3. The communication traffic distribution method based on network load awareness as described in claim 1, characterized in that, The extraction of short-term load change trend features for each target node using the time series decomposition algorithm includes: Obtain historical cumulative data of the comprehensive load score of each target node and generate a historical load score sequence; Based on the load change cycle characteristics of the historical load score sequence, the trend smoothing window size of STL decomposition is determined, and STL time series decomposition is performed to obtain the trend component and residual component. Linear regression fitting is performed on the trend components, and the regression slope is calculated as the load growth rate. Statistical characteristic parameters are calculated based on the residual components, and the load fluctuation amplitude is obtained by combining autocorrelation analysis. Based on the load growth rate and the load fluctuation amplitude, short-term load change trend characteristics of each target node are generated.

4. The communication traffic distribution method based on network load awareness as described in claim 1, characterized in that, The step of distributing traffic differentially to the corresponding target nodes includes: Extract features from the traffic to be distributed to obtain traffic type and priority identifiers; Based on the traffic type identifier, calculate the service compatibility of each target node with the current traffic, and select a set of candidate nodes that meet the service compatibility standard according to the preset compatibility threshold. The optimal traffic allocation weights for candidate nodes are adjusted based on traffic priority identifiers to generate the final allocation weights. Based on the final allocation weights, a weighted round-robin algorithm is used to distribute traffic to the corresponding target nodes, and distribution statistics are recorded.

5. The communication traffic distribution method based on network load awareness as described in claim 1, characterized in that, The method of using a dual-mode weight allocation algorithm based on prediction confidence to determine the optimal traffic allocation weights for each target node includes: Obtain the load prediction reliability index of each target node, and calculate the dual-mode fusion ratio based on the load prediction reliability index of each target node. Based on the comprehensive performance evaluation value of each target node and the short-term load change trend characteristics, calculate the forward-looking weight allocation coefficient; Based on the comprehensive performance evaluation value of each target node and the current system load distribution status, calculate the responsive weight allocation coefficient; The forward-looking weight allocation coefficient and the responsive weight allocation coefficient are weighted and merged according to the dual-mode fusion ratio to generate the optimal traffic allocation weight for each target node.

6. The communication traffic distribution method based on network load awareness as described in claim 1, characterized in that, The multidimensional load metrics include CPU utilization, memory usage, network bandwidth utilization, I / O response latency, and number of connections.

7. A network load-aware communication traffic distribution system, based on the network load-aware communication traffic distribution method according to any one of claims 1 to 6, characterized in that, include: The data acquisition module is used to collect multi-dimensional load metrics of each target node in the network; The load assessment module is used to perform weighted fusion processing on the multi-dimensional load indicators using a sliding time window to generate a comprehensive load score sequence for each target node. The trend analysis module is used to extract the short-term load change trend characteristics of each target node based on the historical cumulative data of the comprehensive load score of each target node and the time series decomposition algorithm. The decision modeling module is used to establish a traffic distribution decision model based on the comprehensive load score sequence and the short-term load change trend characteristics, so as to obtain the optimal traffic allocation weight for each target node. The traffic distribution module is used to distribute traffic to the corresponding target nodes in a differentiated manner based on the optimal traffic allocation weight and the type characteristics of the traffic to be distributed; The monitoring and feedback module is used to monitor the traffic distribution execution status of each target node. When an abnormality in distribution execution or abnormal node response is detected, the traffic distribution process is re-executed.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the communication traffic distribution method based on network load awareness as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the network load-aware communication traffic distribution method according to any one of claims 1 to 6.

Images