Mobile network traffic orientation system and method

Abstract

The invention relates to the technical field of mobile network flow management, and discloses a mobile network flow orientation system and method, and the method comprises the steps: obtaining the current network feature data of each original link, and predicting the congestion coefficient of each original link in a future time period according to the current network feature data; determining the link state type of each original link according to the congestion coefficient; when at least one congested link exists, obtaining the time delay sensitivity of each service module in the congested link, and screening out a service module to be redirected according to the time delay sensitivity; counting all non-congestion links, matching an optimal switching link from the non-congestion links, and marking the optimal switching link as a target link; in a future time period, switching the mobile network flow of the service module to be redirected from the congestion link to a target link; according to the invention, flow redirection can be executed in a self-adaptive manner, so that each service module can obtain stable and high-quality network service at any time.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile network traffic management, and more specifically, to a mobile network traffic orientation system and method. Background Art

[0002] In a mobile network environment with a microservice architecture, business modules usually rely on multiple physical or virtual communication links for data transmission to ensure efficient and stable business operations. However, existing traffic scheduling methods fail to fully utilize the network resources of multiple links, resulting in continuous high loads on some links, while the bandwidth and computing power of other links are not reasonably allocated, resulting in low utilization of network resources and affecting the service quality of the business.

[0003] Traditional traffic steering methods are usually based on fixed rules or real-time load balancing strategies, lacking advance awareness of network congestion, resulting in the inability to adjust link resources in a timely manner when the network status changes suddenly, which in turn affects the continuity and stability of the service. At the same time, different services have different requirements for network characteristics. For example, low-latency services are highly sensitive to network jitter and delay, while high-throughput services mainly focus on bandwidth utilization. Existing traffic steering solutions fail to fully consider service needs during the scheduling process, which may cause high-latency sensitive services to run on inappropriate links, affecting user experience. In addition, after network congestion occurs, existing methods lack accurate optimization strategies during service traffic migration. The selection of target links is usually based on static thresholds or fixed weights, and fails to be dynamically optimized in combination with future network status, resulting in load bottlenecks on some links after migration. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a mobile network traffic orientation system and method.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A mobile network traffic orientation method, the method comprising:

[0007] S101: in the current time period [TK, T], the original links where each service module is located and the current network characteristic data of each original link are obtained, and the current network characteristic data is input into a preconfigured prediction model to predict the congestion coefficient of each original link in the future time period, where T and K are integers greater than zero;

[0008] S102: Determine the link state type of each original link according to the congestion coefficient; the link state type includes a congested link and a non-congested link;

[0009] S103: When there is at least one congested link, obtain the delay sensitivity of each service module in the congested link, and select the service module to be redirected from each service module of the corresponding congested link according to the delay sensitivity;

[0010] S104: Count all non-congested links, and use a pre-configured genetic algorithm to match the optimal switching link from all non-congested links for the service module to be redirected, and mark it as a target link;

[0011] S105: In a future period, the mobile network traffic of the service module to be redirected is switched from the congested link to the target link, so as to achieve traffic redirection optimization for the service module to be redirected.

[0012] Preferably, the current network characteristic data includes bandwidth utilization, throughput, delay, packet loss rate and jitter value at each moment in the current time period.

[0013] Preferably, the method for generating the preconfigured prediction model is as follows:

[0014] Collecting historical congestion coefficient training data, dividing the historical congestion coefficient training data into a congestion coefficient training set and a congestion coefficient test set, wherein the historical congestion coefficient training data includes current network feature data in multiple time periods and their corresponding congestion coefficients;

[0015] The generation process of the congestion coefficient in the historical congestion coefficient training data is as follows:

[0016] Obtain network indicator data of the original link within the observation time window, wherein the network indicator data includes traffic arrival rate, link processing capacity, average delay and packet loss rate;

[0017] Substitute the network indicator data into the pre-built congestion coefficient calculation model to obtain the congestion coefficient;

[0018] The expression of the congestion coefficient calculation model is as follows:

[0019] ;

[0020] Where: represents the congestion factor, is the length of the observation time window, represents the traffic arrival rate at time h, represents the link processing capacity at time h, is the maximum value function, is a very small positive number, represents the end-to-end delay at time h, is the reference delay threshold, represents the packet loss rate at time h, and is a weight coefficient greater than zero, where is the Softplus function, and the expression of the Softplus function is , where: is the input variable, e is a constant, is the natural logarithm function;

[0021] The LSTM neural network is used as the training model, and the current network feature data in the congestion coefficient training set is used as the input of the training model, and the congestion coefficient is used as the output of the training model. The training model is trained to obtain the first regression network to be verified;

[0022] The congestion coefficient test set is used to perform model verification on the first regression network to be verified until the test error of the first regression network to be verified is less than a preset error threshold, and the corresponding first regression network to be verified is output as a trained prediction model.

[0023] Preferably, determining the link state type of each original link according to the congestion coefficient includes:

[0024] Set the congestion factor threshold;

[0025] Compare the congestion coefficient of each original link with the congestion coefficient threshold to determine whether the congestion coefficient is greater than or equal to the congestion coefficient threshold;

[0026] If so, the corresponding original link is marked as a congested link and used as the link state type of the corresponding original link; if not, the corresponding original link is marked as a non-congested link and used as the link state type of the corresponding original link.

[0027] Preferably, the acquiring the delay sensitivity of each service module in the congested link and selecting the service module to be redirected from each service module of the corresponding congested link according to the delay sensitivity includes:

[0028] Retrieve the delay sensitivity comparison table, traverse the delay sensitivity comparison table according to the identifier of each business module, and extract the delay sensitivity of each business module;

[0029] The delay sensitivity of each service module is sorted from high to low according to the value, and the service module corresponding to the first ranking is marked as the service module to be redirected.

[0030] Preferably, when there is no congested link, K=K+r is set, and the process returns to step S101, where r is an integer greater than zero.

[0031] Preferably, the method of using a preconfigured genetic algorithm to match an optimal switching link from all non-congested links for the service module to be redirected includes:

[0032] a1: Initialize the population: randomly generate an original population based on all non-congested links, the original population contains M individuals, each individual represents a non-congested link, and M is an integer greater than zero;

[0033] a2: Fitness evaluation: For each individual, predict the remaining bandwidth utilization of the corresponding non-congested link after migrating the traffic of the service module to be redirected to the corresponding non-congested link; and input the remaining bandwidth utilization of the corresponding non-congested link into the pre-built fitness function to calculate the fitness of each individual;

[0034] The calculation formula of the pre-constructed fitness function is: ; Where: For fitness, is the remaining bandwidth utilization;

[0035] a3: Selection: Use the roulette method to select two individuals with high fitness in the original population as the father and mother;

[0036] a4: Crossover: Perform crossover operation on the father and mother to produce new individuals;

[0037] a5: Mutation: Perform mutation operation on the new individuals to obtain N new individuals, combine the N new individuals into a new population, replace the original population with the new population, and return to step a2;

[0038] a6: Repeat steps a2 to a5 above until the fitness of the individuals in the original population or the new population is greater than or equal to the preset fitness threshold, or the number of iterations is greater than or equal to the preset maximum number of iterations threshold, and output the non-congested link represented by the corresponding individual as the optimal switching link.

[0039] Preferably, the prediction of the remaining bandwidth utilization of the corresponding non-congested link after the traffic of the service module to be redirected is migrated to the corresponding non-congested link includes:

[0040] Acquire traffic migration characteristic data in the current time period [TK, T], wherein the traffic migration characteristic data includes the bandwidth capacity of the congested link where the service module to be redirected is located, the remaining bandwidth utilization of the congested link, the congestion coefficient of the congested link, the bandwidth capacity of the non-congested link, the remaining bandwidth utilization of the non-congested link, and the congestion coefficient of the non-congested link;

[0041] The traffic migration feature data is input into the pre-built LSTM neural network model to obtain the remaining bandwidth utilization of the corresponding non-congested link in the future period after the traffic of the service module to be redirected is migrated to the corresponding non-congested link;

[0042] The generation process of the LSTM neural network model is as follows:

[0043] Collecting historical bandwidth utilization training data, dividing the historical bandwidth utilization training data into a bandwidth utilization training set and a bandwidth utilization test set, wherein the historical bandwidth utilization training data includes traffic migration feature data in multiple time periods and their corresponding remaining bandwidth utilization;

[0044] The LSTM neural network is used as a training model, and the traffic migration feature data in the bandwidth utilization training set is used as the input of the LSTM neural network, and the remaining bandwidth utilization is used as the output of the LSTM neural network. The LSTM neural network is trained to obtain a second regression network to be verified.

[0045] The bandwidth utilization test set is used to perform model verification on the second regression network to be verified until the test error of the second regression network to be verified is less than the preset error threshold, and the corresponding second regression network to be verified is output as the trained LSTM neural network model.

[0046] A mobile network traffic orientation system is implemented based on the mobile network traffic orientation method described above, and the system comprises:

[0047] The data acquisition module is used to obtain the original links where each service module is located and the current network characteristic data of each original link in the current time period [TK, T], and input the current network characteristic data into the preconfigured prediction model to predict the congestion coefficient of each original link in the future time period, where T and K are integers greater than zero;

[0048] A link analysis module, used to determine the link state type of each original link according to the size of the congestion coefficient; the link state type includes a congested link and a non-congested link;

[0049] A link screening module, used for obtaining the delay sensitivity of each service module in the congested link when there is at least one congested link, and screening the service modules to be redirected from the service modules of the corresponding congested link according to the delay sensitivity;

[0050] The link optimization module is used to count all non-congested links and use a pre-configured genetic algorithm to match the optimal switching link from all non-congested links for the service module to be redirected, and mark it as the target link;

[0051] The directional optimization module is used to switch the mobile network traffic of the service module to be redirected from the congested link to the target link in the future period, so as to realize the traffic redirection optimization of the service module to be redirected.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The present application discloses a mobile network traffic orientation system and method, comprising: obtaining current network feature data of each original link, and predicting the congestion coefficient of each original link in a future period according to the current network feature data; determining the link state type of each original link according to the size of the congestion coefficient; when there is at least one congested link, obtaining the delay sensitivity of each service module in the congested link, and filtering out the service module to be redirected according to the delay sensitivity; counting all non-congested links, and matching the optimal switching link therefrom, and marking it as a target link; in the future period, switching the mobile network traffic of the service module to be redirected from the congested link to the target link. ; Based on the above characteristics, the present invention realizes efficient allocation of network resources and optimization of business stability through intelligent traffic orientation based on network status prediction. Compared with existing methods, the present invention can perceive network congestion trends in advance, reduce the lag of traffic scheduling, and ensure that the business module completes the optimal path adjustment before the network status changes; by combining the business requirements for network characteristics, the traffic distribution of different types of services is optimized, so that key businesses can run in a suitable link environment, improve the overall service quality, and help ensure that each business module can obtain stable and high-quality network services at any time, thereby ensuring the stability and efficient operation of the entire business system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A flow chart of a mobile network traffic orientation method provided by the present invention;

[0055] Figure 2 A schematic diagram of the module structure of a mobile network traffic orientation system provided by the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Example 1

[0058] See also Figure 1 As shown, this embodiment discloses a mobile network traffic orientation method, the method comprising:

[0059] S101: In the current time period [TK, T], the original links where each service module is located and the current network characteristic data of each original link are obtained, and the current network characteristic data is input into a preconfigured prediction model to predict the congestion coefficient of each original link in the future time period, where T and K are integers greater than zero;

[0060] It should be understood that: under the microservice architecture, the business module refers to an application instance based on the microservice architecture (MSA). Each business module can run independently and the service registration center schedules traffic. For example, a business system includes multiple business modules such as VoIP calls, online videos, and file downloads; and the original link refers to the physical link or virtual link (such as 5G, Wi-Fi, MEC edge computing, and SD-WAN) that connects each business module to the external network. Usually, there will be multiple communication links under the microservice architecture to ensure the stable operation of the business module, but the existing methods fail to fully utilize the resources of multiple communication links, resulting in some links still bearing too much traffic from business modules under high load, while the resources of other links are not fully utilized, resulting in an uneven distribution of network resources, affecting the stability of the business and the overall network performance;

[0061] It should be noted that in the actual process, the original links of each business module can be obtained through API gateway log analysis, Service Mesh traffic monitoring or DPI (deep packet inspection) and other methods;

[0062] The current network characteristic data includes bandwidth utilization, throughput, latency, packet loss rate and jitter value at each moment in the current period;

[0063] Specifically, the method for generating the preconfigured prediction model is as follows:

[0064] Collecting historical congestion coefficient training data, dividing the historical congestion coefficient training data into a congestion coefficient training set and a congestion coefficient test set, wherein the historical congestion coefficient training data includes current network feature data in multiple time periods and their corresponding congestion coefficients;

[0065] It should be noted that the current network characteristic data and congestion coefficient in the historical congestion coefficient training data are recorded and calculated by technicians based on experimental data;

[0066] The generation process of the congestion coefficient in the historical congestion coefficient training data is as follows:

[0067] Obtain network indicator data of the original link within the observation time window, wherein the network indicator data includes traffic arrival rate, link processing capacity, average delay and packet loss rate;

[0068] Substitute the network indicator data into the pre-built congestion coefficient calculation model to obtain the congestion coefficient;

[0069] The expression of the congestion coefficient calculation model is as follows:

[0070] ;

[0071] Where: represents the congestion factor, is the length of the observation time window, which is set in advance by the technician. represents the traffic arrival rate at time h (unit: bps), Indicates the link processing capacity at time h (unit: bps), is the maximum value function, is a very small positive number used to avoid division by zero errors or extreme cases in mathematical calculations. represents the end-to-end delay at time h (unit: ms), is the reference delay threshold (unit: ms), which is used to normalize the impact of delay on congestion. represents the packet loss rate at time h, and are weight coefficients greater than zero, which are used to adjust the contribution of delay and packet loss rate to the congestion coefficient. They are set by actual personnel based on experimental data. is the Softplus function, and the expression of the Softplus function is , where: is the input variable, e is a constant, is the natural logarithm function;

[0072] The LSTM neural network is used as the training model, and the current network feature data in the congestion coefficient training set is used as the input of the training model, and the congestion coefficient is used as the output of the training model. The training model is trained to obtain the first regression network to be verified;

[0073] The congestion coefficient test set is used to perform model verification on the first regression network to be verified until the test error of the first regression network to be verified is less than a preset error threshold, and the corresponding first regression network to be verified is output as a trained prediction model.

[0074] S102: Determine the link state type of each original link according to the congestion coefficient; the link state type includes a congested link and a non-congested link;

[0075] In implementation, determining the link state type of each original link according to the congestion coefficient includes:

[0076] Set the congestion factor threshold;

[0077] Compare the congestion coefficient of each original link with the congestion coefficient threshold to determine whether the congestion coefficient is greater than or equal to the congestion coefficient threshold;

[0078] If yes, the corresponding original link is marked as a congested link and used as the link state type of the corresponding original link; if no, the corresponding original link is marked as a non-congested link and used as the link state type of the corresponding original link;

[0079] It can be understood that: when the link status type of the original link is a congested link, it means that the communication task undertaken is overloaded, that is, a certain original link bears too much mobile network traffic of business modules, and part or all of the traffic needs to be redirected to alleviate the network congestion of the corresponding congested link to ensure the stable operation of each business module in the congested link. Otherwise, no processing is required.

[0080] S103: When there is at least one congested link, obtain the delay sensitivity of each service module in the congested link, and select the service module to be redirected from each service module of the corresponding congested link according to the delay sensitivity;

[0081] In implementation, the obtaining of the delay sensitivity of each service module in the congested link and selecting the service module to be redirected from each service module of the corresponding congested link according to the delay sensitivity includes:

[0082] Retrieve the delay sensitivity comparison table, traverse the delay sensitivity comparison table according to the identifier of each business module, and extract the delay sensitivity of each business module;

[0083] It should be noted that: the delay sensitivity comparison table is pre-stored in the system database, and is manually formulated by technical personnel based on the actual situation of each business module;

[0084] It should be understood that latency sensitivity refers to the degree of requirement of a certain service module or application for network latency (i.e., the time required for a data packet to travel from the source to the destination). In network performance optimization and traffic scheduling, different service modules have different tolerances for latency. For example, VoIP calls have very strict latency requirements, and any higher latency may cause the call to be unsmooth. However, service modules such as file downloads have relatively loose latency requirements, and even a slightly higher latency will not have a significant impact on the user experience.

[0085] Sort the delay sensitivity of each service module from high to low according to the value, and mark the service module that corresponds to the first one in the ranking as the service module to be redirected;

[0086] It is understandable that the greater the delay sensitivity, the more sensitive the service module is to network delay. Any increase in delay will significantly affect its performance and user experience. In traffic scheduling and network optimization, such services usually need to give priority to low-latency links to ensure service quality and real-time response. For example, the service system includes service modules such as VoIP calls and file downloads. Compared with file downloads, VoIP calls require lower delays, so they need to be treated with higher priority.

[0087] For example, it is assumed that there are multiple service modules such as VoIP calls, online videos and file downloads in the service system, and there are two original links, namely Z1 and Z2. VoIP calls and online videos are located in the original link Z1, while file downloads are located in the original link Z2. However, it is predicted that Z1 will be congested, so Z1 is marked as a congested path, and after querying the delay sensitivity comparison table, it is found that the delay sensitivity of VoIP calls is 0.8, while the delay sensitivity of online videos is 0.6. Therefore, after sorting according to the delay sensitivity, VoIP calls are marked as service modules to be redirected for subsequent traffic redirection optimization processing;

[0088] In a specific implementation, when there is no congested link, K=K+r is set, and the process returns to step S101, where r is an integer greater than zero;

[0089] It can be understood that: when there is no congested link, it means that the original links will communicate smoothly in the future and there is no network congestion. Therefore, no optimization is required, and the congestion detection phase of the next time period is entered.

[0090] S104: Count all non-congested links, and use a pre-configured genetic algorithm to match the optimal switching link from all non-congested links for the service module to be redirected, and mark it as a target link;

[0091] In implementation, the method of using a preconfigured genetic algorithm to match the optimal switching link from all non-congested links for the service module to be redirected includes:

[0092] a1: Initialize the population: randomly generate an original population based on all non-congested links, the original population contains M individuals, each individual represents a non-congested link, and M is an integer greater than zero;

[0093] a2: Fitness evaluation: For each individual (i.e., each non-congested link), predict the remaining bandwidth utilization of the corresponding non-congested link after the traffic of the service module to be redirected is migrated to the corresponding non-congested link; and input the remaining bandwidth utilization of the corresponding non-congested link into the pre-built fitness function to calculate the fitness of each individual;

[0094] Specifically, the prediction of the remaining bandwidth utilization of the corresponding non-congested link after the traffic of the service module to be redirected is migrated to the corresponding non-congested link includes:

[0095] Acquire traffic migration characteristic data in the current time period [TK, T], wherein the traffic migration characteristic data includes the bandwidth capacity of the congested link where the service module to be redirected is located, the remaining bandwidth utilization of the congested link, the congestion coefficient of the congested link, the bandwidth capacity of the non-congested link, the remaining bandwidth utilization of the non-congested link, and the congestion coefficient of the non-congested link;

[0096] The traffic migration feature data is input into the pre-built LSTM neural network model to obtain the remaining bandwidth utilization of the corresponding non-congested link in the future period after the traffic of the service module to be redirected is migrated to the corresponding non-congested link;

[0097] Specifically, the generation process of the LSTM neural network model is as follows:

[0098] Collecting historical bandwidth utilization training data, dividing the historical bandwidth utilization training data into a bandwidth utilization training set and a bandwidth utilization test set, wherein the historical bandwidth utilization training data includes traffic migration feature data in multiple time periods and their corresponding remaining bandwidth utilization;

[0099] It should be noted that the traffic migration feature data of the historical bandwidth utilization training data and its corresponding remaining bandwidth utilization are recorded by the technicians based on the experimental data;

[0100] The LSTM neural network is used as a training model, and the traffic migration feature data in the bandwidth utilization training set is used as the input of the LSTM neural network, and the remaining bandwidth utilization is used as the output of the LSTM neural network. The LSTM neural network is trained to obtain a second regression network to be verified.

[0101] The bandwidth utilization test set is used to perform model verification on the second regression network to be verified until the test error of the second regression network to be verified is less than the preset error threshold, and the corresponding second regression network to be verified is output as the trained LSTM neural network model;

[0102] In genetic algorithms, fitness evaluation is one of the core steps, which is used to determine the fitness of each individual, that is, to determine how well each individual performs in a given problem; the design of the fitness function directly affects the efficiency of the algorithm and the quality of the final solution;

[0103] The calculation formula of the pre-constructed fitness function is: ; Where: For fitness, is the remaining bandwidth utilization;

[0104] a3: Selection: Use the roulette method to select two individuals with high fitness in the original population as the father and mother;

[0105] The roulette wheel method is a commonly used selection method, which is used in genetic algorithms to select individuals with higher fitness to enter the next generation. It simulates the process of roulette, and each individual obtains a corresponding "roulette wheel" area according to its fitness. The higher the fitness of the individual, the larger the corresponding area, and the higher the probability of being selected.

[0106] a4: Crossover: Perform crossover operation on the father and mother to produce new individuals;

[0107] It should be noted that the crossover operation on the paternal parent and the maternal parent is implemented based on a crossover operation, and the crossover operation includes but is not limited to one of a single-point crossover, a uniform crossover or a sequential crossover, etc.;

[0108] In genetic algorithms, crossover is an important genetic operation used to generate a new generation of candidate solutions; the basic idea of ​​crossover is to simulate the sexual reproduction process in biological genetics, in which two parents produce offspring by combining their genetic information; this process helps to introduce diversity in the solution space and may produce new individuals that are more adaptable to the environment;

[0109] a5: Mutation: Perform mutation operation on the new individuals to obtain N new individuals, combine the N new individuals into a new population, replace the original population with the new population, and return to step a2;

[0110] In genetic algorithms, mutation operations are used to introduce genetic diversity and prevent the algorithm from falling into a local optimum. The mutation operation on new individuals is implemented by uniform mutation or Gaussian mutation.

[0111] a6: Repeat steps a2 to a5 above until the fitness of the individuals in the original population or the new population is greater than or equal to the preset fitness threshold, or the number of iterations is greater than or equal to the preset maximum number of iterations threshold, and output the non-congested link represented by the corresponding individual as the optimal switching link;

[0112] For example: Assume that the maximum number of iterations is 100 times, and record the individual with the highest fitness in the current population and its fitness value after each iteration; if it is found that the fitness value has not changed significantly in a certain generation, it is considered that the convergence condition is met, the iteration is stopped, and the non-congested link represented by the corresponding individual is output as the optimal switching link.

[0113] S105: In a future period, switching the mobile network traffic of the service module to be redirected from the congested link to the target link, so as to achieve traffic redirection optimization of the service module to be redirected;

[0114] It can be understood that after the redirected business module is switched from the congested link to the target link, the network switches of both the congested link and the target link are in the normal range, which is conducive to ensuring that each business module can obtain stable and high-quality network services at any time, thereby ensuring the stability and efficient operation of the entire business system.

[0115] Example 2

[0116] See also Figure 2 As shown, based on the same inventive concept, this embodiment discloses a mobile network traffic orientation system. For details not provided in this embodiment, please refer to the description of the relevant parts in Embodiment 1. The system includes:

[0117] The data acquisition module 210 is used to acquire the original links where each service module is located and the current network characteristic data of each original link in the current time period [TK, T], and input the current network characteristic data into the preconfigured prediction model to predict the congestion coefficient of each original link in the future time period, where T and K are integers greater than zero;

[0118] The link analysis module 220 is used to determine the link status type of each original link according to the size of the congestion coefficient; the link status type includes a congested link and a non-congested link;

[0119] The link screening module 230 is used to obtain the delay sensitivity of each service module in the congested link when there is at least one congested link, and screen the service modules to be redirected from the service modules of the corresponding congested link according to the delay sensitivity;

[0120] The link optimization module 240 is used to count all non-congested links, and use a pre-configured genetic algorithm to match the optimal switching link from all non-congested links for the service module to be redirected, and mark it as a target link;

[0121] The directional optimization module 250 is used to switch the mobile network traffic of the service module to be redirected from the congested link to the target link in the future period, so as to achieve traffic redirection optimization of the service module to be redirected.

[0122] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters, weights and thresholds in the formula are set by technicians in this field according to actual conditions.

[0123] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired network or a wireless network. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media sets. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD) or a semiconductor medium. The semiconductor medium may be a solid-state hard disk.

[0124] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0125] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0126] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0127] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0128] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0129] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0130] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A mobile network traffic orientation method, characterized in that: The method comprises: S101: In the current time period [TK, T], the original links where each service module is located and the current network characteristic data of each original link are obtained, and the current network characteristic data is input into a preconfigured prediction model to predict the congestion coefficient of each original link in the future time period, where T and K are integers greater than zero; S102: Determine the link state type of each original link according to the congestion coefficient; the link state type includes a congested link and a non-congested link; S103: When there is at least one congested link, obtain the delay sensitivity of each service module in the congested link, and select the service module to be redirected from each service module of the corresponding congested link according to the delay sensitivity; S104: Count all non-congested links, and use a pre-configured genetic algorithm to match the optimal switching link from all non-congested links for the service module to be redirected, and mark it as a target link; S105: In a future period, the mobile network traffic of the service module to be redirected is switched from the congested link to the target link, so as to achieve traffic redirection optimization for the service module to be redirected.

2. A mobile network traffic orientation method according to claim 1, characterized in that: The current network characteristic data includes bandwidth utilization, throughput, delay, packet loss rate and jitter value at each moment in the current period.

3. A mobile network traffic orientation method according to claim 2, characterized in that: The method for generating the preconfigured prediction model is as follows: Collecting historical congestion coefficient training data, dividing the historical congestion coefficient training data into a congestion coefficient training set and a congestion coefficient test set, wherein the historical congestion coefficient training data includes current network feature data in multiple time periods and their corresponding congestion coefficients; The generation process of the congestion coefficient in the historical congestion coefficient training data is as follows: Obtain network indicator data of the original link within the observation time window, wherein the network indicator data includes traffic arrival rate, link processing capacity, average delay and packet loss rate; Substitute the network indicator data into the pre-built congestion coefficient calculation model to obtain the congestion coefficient; The expression of the congestion coefficient calculation model is as follows: ; Where: represents the congestion factor, is the length of the observation time window, represents the traffic arrival rate at time h, represents the link processing capacity at time h, is the maximum value function, is a very small positive number, represents the end-to-end delay at time h, is the reference delay threshold, represents the packet loss rate at time h, and is a weight coefficient greater than zero, where is the Softplus function, and the expression of the Softplus function is , where: is the input variable, e is a constant, is the natural logarithm function; The LSTM neural network is used as the training model, and the current network feature data in the congestion coefficient training set is used as the input of the training model, and the congestion coefficient is used as the output of the training model. The training model is trained to obtain the first regression network to be verified; The congestion coefficient test set is used to perform model verification on the first regression network to be verified until the test error of the first regression network to be verified is less than a preset error threshold, and the corresponding first regression network to be verified is output as a trained prediction model.

4. A mobile network traffic orientation method according to claim 3, characterized in that: Determining the link state type of each original link according to the congestion coefficient includes: Set the congestion factor threshold; Compare the congestion coefficient of each original link with the congestion coefficient threshold to determine whether the congestion coefficient is greater than or equal to the congestion coefficient threshold; If so, the corresponding original link is marked as a congested link and used as the link state type of the corresponding original link; if not, the corresponding original link is marked as a non-congested link and used as the link state type of the corresponding original link.

5. A method for mobile network traffic orientation according to claim 4, characterized in that: The obtaining of the delay sensitivity of each service module in the congested link and selecting the service module to be redirected from each service module of the corresponding congested link according to the delay sensitivity includes: Retrieve the delay sensitivity comparison table, traverse the delay sensitivity comparison table according to the identifier of each business module, and extract the delay sensitivity of each business module; The delay sensitivity of each service module is sorted from high to low according to the value, and the service module corresponding to the first ranking is marked as the service module to be redirected.

6. A method for mobile network traffic orientation according to claim 5, characterized in that: When there is no congested link, K=K+r is set and the process returns to step S101, where r is an integer greater than zero.

7. A method for mobile network traffic orientation according to claim 6, characterized in that: The method of using a preconfigured genetic algorithm to match an optimal switching link from all non-congested links for the service module to be redirected includes: a1: Initialize the population: randomly generate an original population based on all non-congested links, the original population contains M individuals, each individual represents a non-congested link, and M is an integer greater than zero; a2: Fitness evaluation: For each individual, predict the remaining bandwidth utilization of the corresponding non-congested link after migrating the traffic of the service module to be redirected to the corresponding non-congested link; and input the remaining bandwidth utilization of the corresponding non-congested link into the pre-built fitness function to calculate the fitness of each individual; The calculation formula of the pre-constructed fitness function is: ; Where: For fitness, is the remaining bandwidth utilization; a3: Selection: Use the roulette method to select two individuals with high fitness in the original population as the father and mother; a4: Crossover: Perform crossover operation on the father and mother to produce new individuals; a5: Mutation: Perform mutation operation on the new individuals to obtain N new individuals, combine the N new individuals into a new population, replace the original population with the new population, and return to step a2; a6: Repeat steps a2 to a5 above until the fitness of the individuals in the original population or the new population is greater than or equal to the preset fitness threshold, or the number of iterations is greater than or equal to the preset maximum number of iterations threshold, and output the non-congested link represented by the corresponding individual as the optimal switching link.

8. A method for mobile network traffic orientation according to claim 7, characterized in that: The prediction of the remaining bandwidth utilization of the corresponding non-congested link after the traffic of the service module to be redirected is migrated to the corresponding non-congested link includes: Acquire traffic migration characteristic data in the current time period [TK, T], wherein the traffic migration characteristic data includes the bandwidth capacity of the congested link where the service module to be redirected is located, the remaining bandwidth utilization of the congested link, the congestion coefficient of the congested link, the bandwidth capacity of the non-congested link, the remaining bandwidth utilization of the non-congested link, and the congestion coefficient of the non-congested link; The traffic migration feature data is input into the pre-built LSTM neural network model to obtain the remaining bandwidth utilization of the corresponding non-congested link in the future period after the traffic of the service module to be redirected is migrated to the corresponding non-congested link; The generation process of the LSTM neural network model is as follows: Collecting historical bandwidth utilization training data, dividing the historical bandwidth utilization training data into a bandwidth utilization training set and a bandwidth utilization test set, wherein the historical bandwidth utilization training data includes traffic migration feature data in multiple time periods and their corresponding remaining bandwidth utilization; The LSTM neural network is used as a training model, and the traffic migration feature data in the bandwidth utilization training set is used as the input of the LSTM neural network, and the remaining bandwidth utilization is used as the output of the LSTM neural network. The LSTM neural network is trained to obtain a second regression network to be verified. The bandwidth utilization test set is used to perform model verification on the second regression network to be verified until the test error of the second regression network to be verified is less than the preset error threshold, and the corresponding second regression network to be verified is output as the trained LSTM neural network model.

9. A mobile network traffic orientation system, characterized in that: The method is implemented based on a mobile network traffic orientation method according to any one of claims 1 to 8, and the system comprises: The data acquisition module is used to obtain the original links where each service module is located and the current network characteristic data of each original link in the current time period [TK, T], and input the current network characteristic data into the preconfigured prediction model to predict the congestion coefficient of each original link in the future time period, where T and K are integers greater than zero; A link analysis module, used to determine the link state type of each original link according to the size of the congestion coefficient; the link state type includes a congested link and a non-congested link; A link screening module, used for obtaining the delay sensitivity of each service module in the congested link when there is at least one congested link, and screening the service modules to be redirected from the service modules of the corresponding congested link according to the delay sensitivity; The link optimization module is used to count all non-congested links and use a pre-configured genetic algorithm to match the optimal switching link from all non-congested links for the service module to be redirected, and mark it as the target link; The directional optimization module is used to switch the mobile network traffic of the service module to be redirected from the congested link to the target link in the future period, so as to realize the traffic redirection optimization of the service module to be redirected.

Images