Business feature-based directional shunting method, device and equipment and storage medium

Abstract

The application provides a service feature-based directional shunting method, device, equipment and storage medium, relates to the technical field of data directional shunting, and the method comprises the following steps: obtaining acceleration service information and acceleration platform information; obtaining corresponding attribute feature information according to the acceleration service information; grouping various attribute feature information, adding corresponding service feature labels to the attribute feature information in the same group; inputting the attribute feature information with the service feature labels and the acceleration platform information into a preset prediction model to obtain acceleration prediction information corresponding to the attribute feature information; determining shunting strategy information according to the acceleration prediction information; and distributing service traffic corresponding to the acceleration service information to a corresponding acceleration platform according to the shunting strategy information. By predicting the power consumption and acceleration time of the service in different acceleration platforms, the shunting strategy information is determined based on the principle of minimizing the power consumption and acceleration time, so that the comprehensive load of the power consumption and acceleration time of the acceleration platform is more balanced.

Description

Technical Field

[0001] This invention relates to the field of data targeted offloading technology, and in particular to methods, apparatus, equipment and storage media for targeted offloading based on business characteristics. Background Technology

[0002] Inter-network traffic refers to data traffic transmitted between different network or Internet Service Providers (ISPs). Some user services require access to resources or services outside the same network, such as browsing websites hosted by different ISPs. Because inter-network traffic crosses networks, it is prone to service quality issues such as high latency and unstable communication. With the continuous development of the Internet, inter-network traffic redirection technology has played a crucial role in improving the quality of inter-network services and optimizing user experience. By redirecting user traffic from one network to another or to an acceleration platform, it can ensure service quality, reduce communication latency, and make communication more stable.

[0003] However, when processing the same type of business, existing technologies often only allow these businesses to be uploaded to the same acceleration platform for acceleration, without fully considering the load balancing issue of the acceleration platform.

[0004] In existing technologies, inter-network traffic splitting only uploads services of the same type to the same acceleration platform for acceleration according to a fixed splitting strategy, without considering the load balancing problem of the acceleration platform. This results in uneven load and low performance utilization of the acceleration platform, and the inability to fully schedule and utilize the resource performance of the acceleration platform. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and storage medium for targeted traffic splitting based on service characteristics, in order to solve the defects of uneven load and low performance utilization of acceleration platforms caused by fixed traffic splitting strategies in the prior art during inter-network traffic splitting.

[0006] This invention provides a business-feature-based targeted traffic splitting method, including:

[0007] Obtain information about acceleration services and acceleration platforms;

[0008] Based on the acceleration service information, obtain the corresponding attribute feature information;

[0009] Group the attribute feature information into groups, and add corresponding business feature tags to the attribute feature information in the same group;

[0010] The attribute feature information with the business feature tag and the acceleration platform information are input into a preset prediction model to obtain acceleration prediction information corresponding to the attribute feature information. The acceleration prediction information represents the power consumption and acceleration time of the business traffic corresponding to the acceleration business information on different acceleration platforms.

[0011] Based on the acceleration prediction information, a traffic splitting strategy is determined, which is determined based on minimizing power consumption and acceleration time.

[0012] Based on the traffic distribution strategy information, the traffic corresponding to the accelerated service information is allocated to the corresponding acceleration platform.

[0013] According to the business feature-based targeted traffic splitting method provided by the present invention, the step of determining the traffic splitting strategy information based on the acceleration prediction information includes:

[0014] Based on the acceleration prediction information, the power consumption value and corresponding acceleration time value for acceleration on each acceleration platform individually are determined; based on each of the power consumption values ​​and corresponding acceleration time values, the optimal acceleration scheme for combining the acceleration platforms is determined, and the traffic splitting strategy information is generated. The optimal acceleration scheme is determined by minimizing power consumption and acceleration time.

[0015] According to the business feature-based targeted traffic splitting method provided by the present invention, before obtaining the acceleration service information and acceleration platform information, the method further includes:

[0016] Obtain the business access requests of each user;

[0017] Based on the service access request, obtain the corresponding network status information;

[0018] Based on the network status information, determine whether accelerated processing is needed;

[0019] When accelerated processing is required, the business information corresponding to the business access request is marked as accelerated business information.

[0020] According to the service feature-based targeted traffic routing method provided by the present invention, after allocating the service traffic corresponding to the accelerated service information to the corresponding acceleration platform according to the traffic routing strategy information, the method further includes:

[0021] Based on the acceleration service information, obtain the corresponding service information;

[0022] Perform quality analysis on the aforementioned business service information and obtain the analysis results.

[0023] Based on the analysis results, adjust the traffic diversion strategy information;

[0024] The business service information includes business logs and network status, and the analysis results are obtained based on traffic geographic distribution, network quality, and application identification.

[0025] According to the business feature-based targeted traffic routing method provided by the present invention, the step of performing quality analysis on the business service information and obtaining analysis result information includes:

[0026] Based on the aforementioned service information, determine the intra-network latency and extra-network latency;

[0027] The sum of the intra-network latency and the extra-network latency is taken as the application latency;

[0028] Based on the application latency, determine whether there is a quality issue;

[0029] When a quality issue exists, the latency quality issue information is determined based on the intra-network latency and the extra-network latency.

[0030] Add the time delay quality defect information to the analysis result information.

[0031] According to the business feature-based targeted traffic splitting method provided by the present invention, the step of grouping the attribute feature information and adding corresponding business feature tags to the attribute feature information in the same group includes:

[0032] Based on a preset number of clusters, a corresponding number of the attribute feature information is randomly selected as the first centroid;

[0033] The attribute feature information is used to calculate the centroid distance to each of the first centroids;

[0034] Based on the centroid distance, the attribute feature information is grouped into the same set with the nearest first centroid, forming a cluster set;

[0035] Based on each of the cluster sets, obtain the second centroid and calculate the change between the second centroid and the corresponding first centroid.

[0036] When the change is greater than or equal to a preset threshold, the second centroid is used as the new first centroid, and the second centroid is reacquired.

[0037] When the change is less than the preset threshold, a business feature label is generated based on the second centroid, and the business feature label is added to each attribute feature information in the cluster set corresponding to the second centroid.

[0038] This invention also provides a service-characteristic-based targeted traffic splitting device, comprising:

[0039] The acquisition module is used to acquire acceleration service information and acceleration platform information;

[0040] The processing module is used to obtain corresponding attribute feature information based on the acceleration service information, group the attribute feature information into groups, and add corresponding service feature tags to the attribute feature information in the same group.

[0041] The analysis module is used to input the attribute feature information with the business feature tag and the acceleration platform information into a preset prediction model to obtain acceleration prediction information corresponding to the attribute feature information.

[0042] The traffic splitting module is used to determine traffic splitting strategy information based on the acceleration prediction information, and to allocate the service traffic corresponding to the acceleration service information to the corresponding acceleration platform according to the traffic splitting strategy information.

[0043] The acquisition module, the processing module, the analysis module, and the traffic splitting module cooperate to execute the above-described targeted traffic splitting method based on business characteristics.

[0044] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the service feature-based targeted traffic routing method as described above.

[0045] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the business feature-based targeted traffic routing method as described above.

[0046] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the business feature-based targeted traffic diversion method as described above.

[0047] The service-feature-based targeted traffic splitting method, apparatus, device, and storage medium provided by this invention have at least the following beneficial effects: By acquiring acceleration service information and acceleration platform information, it is possible to identify the services that users need to accelerate and the various acceleration platforms that can provide acceleration services. Based on the acceleration service information, attribute feature information corresponding to each service is extracted to determine the acceleration requirements. The attribute feature information is grouped to group similar services together and add service feature tags for easy differentiation and processing later. A prediction model processes the attribute feature information with service feature tags and the acceleration platform information to predict the power consumption and acceleration time of each service on different acceleration platforms, generating acceleration prediction information. Then, based on the acceleration prediction information and the principle of minimizing power consumption and acceleration time, a traffic splitting strategy is determined to allocate the service traffic corresponding to the acceleration service information to the corresponding acceleration platform according to the traffic splitting strategy. By grouping different business types and adding business feature tags, it is easier to perform traffic splitting more accurately. At the same time, by predicting the power consumption and acceleration time of each business on different acceleration platforms, traffic splitting strategy information is determined based on the principle of minimizing power consumption and acceleration time. After the business traffic is split and accelerated according to the traffic splitting strategy information, the overall load of power consumption and acceleration time of each acceleration platform is more balanced, which is conducive to providing stable acceleration services and making full use of the resource performance of the acceleration platform. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0049] Figure 1 This is one of the flowcharts of the business feature-based targeted traffic diversion method provided by the present invention.

[0050] Figure 2 This is the second flowchart of the business feature-based targeted traffic diversion method provided by the present invention.

[0051] Figure 3 This is the third flowchart of the business feature-based targeted traffic diversion method provided by the present invention.

[0052] Figure 4 This is the fourth flowchart of the business feature-based targeted traffic diversion method provided by the present invention.

[0053] Figure 5 This is the fifth flowchart of the business feature-based targeted traffic diversion method provided by the present invention.

[0054] Figure 6 This is the sixth flowchart of the business feature-based targeted traffic diversion method provided by the present invention.

[0055] Figure 7 This is a schematic diagram of the business feature-based targeted traffic diversion system provided by the present invention.

[0056] Figure 8 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0058] The following is combined Figures 1-6 The present invention describes a business-feature-based targeted traffic splitting method, comprising:

[0059] S100: Obtain information on acceleration services and acceleration platform;

[0060] S200: Obtain the corresponding attribute feature information based on the acceleration service information;

[0061] S300: Group the attribute feature information into groups and add corresponding business feature tags to the attribute feature information in the same group; S400: Input the attribute feature information with the business feature tags and the acceleration platform information into a preset prediction model to obtain acceleration prediction information corresponding to the attribute feature information. The acceleration prediction information represents the power consumption and acceleration time of the service traffic corresponding to the acceleration service information on different acceleration platforms.

[0062] S500: Determine the shunting strategy information based on the acceleration prediction information, wherein the shunting strategy information is determined based on minimizing power consumption and acceleration time;

[0063] S600: Based on the traffic distribution strategy information, allocate the service traffic corresponding to the accelerated service information to the corresponding acceleration platform.

[0064] By acquiring acceleration service information and acceleration platform information, we can identify the services that users need to accelerate and the various acceleration platforms that can provide acceleration services. Based on the acceleration service information, we extract the corresponding attribute feature information for each service to determine the acceleration requirements. We group these attribute feature information to group similar services together and add service feature tags for easy differentiation and processing later. A predictive model processes the attribute feature information with service feature tags and the acceleration platform information to predict the power consumption and acceleration time of each service on different acceleration platforms, generating acceleration prediction information. Then, based on the acceleration prediction information and the principle of minimizing power consumption and acceleration time, we determine the traffic distribution strategy information to allocate the service traffic corresponding to the acceleration service information to the corresponding acceleration platforms according to the traffic distribution strategy information.

[0065] By grouping different business types and adding business feature tags, it is easier to perform traffic splitting more accurately. At the same time, by predicting the power consumption and acceleration time of each business on different acceleration platforms, traffic splitting strategy information is determined based on the principle of minimizing power consumption and acceleration time. After the business traffic is split and accelerated according to the traffic splitting strategy information, the overall load of power consumption and acceleration time of each acceleration platform is more balanced, which is conducive to providing stable acceleration services and making full use of the resource performance of the acceleration platform.

[0066] The services users perform can be categorized into different service types, each with its own attribute characteristics. Examples of some service types and their attribute characteristics are shown in the table below.

[0067]

[0068] Therefore, by determining the attribute characteristics of services that need to be accelerated based on the acceleration service information, it is easier to perform more accurate traffic distribution processing in the future.

[0069] Accelerated service information may include IP address, port information, DPI call detail records, packet characteristics, and service session information to determine the user's service type and corresponding attribute characteristics. Accelerated platform information may include the acceleration platform's identification information, acceleration parameter information, and acceleration level information.

[0070] Due to the diverse types of services on the network, processing each individual service directly would be extremely complex and time-consuming. By adding service feature tags, services with similar characteristics can be grouped together. These tags provide a clear basis for subsequent processing, thus simplifying the process. Furthermore, adding service feature tags enables explicit data classification, which also simplifies the subsequent processing by predictive models.

[0071] When users conduct inter-network services, the resulting service traffic is transmitted between different networks. During the transmission of service traffic, since there are multiple inter-network exits between the networks, different acceleration platforms can be used to access different inter-network exits. Different acceleration platforms have different acceleration settings and biases. The inter-network exit with the most suitable latency, stability and bandwidth is selected to accelerate the service and ensure the quality of service for inter-network services.

[0072] refer to Figure 2 In some embodiments of the business feature-based targeted traffic diversion method of the present invention, S500 includes: S510: determining the power consumption value and the corresponding acceleration time value for individual acceleration on each acceleration platform based on the acceleration prediction information.

[0073] S520: Based on each of the power consumption values ​​and the corresponding acceleration time values, determine the optimal acceleration scheme for combining the acceleration platforms for acceleration, and generate the traffic splitting strategy information. The optimal acceleration scheme is determined by minimizing power consumption and acceleration time.

[0074] By using predictive models to obtain the power consumption and acceleration time values ​​for each acceleration platform, and then through quantitative calculations to combine the power consumption and acceleration time of each platform, the optimal acceleration scheme with the minimum power consumption and acceleration time is determined. Based on this, and in addition to flexibly allocating different service types to various acceleration platforms, further acceleration methods can be combined across different platforms. This means that the same service is not restricted to a single acceleration platform during its duration but can switch between different platforms for acceleration, which helps to discover the optimal combined acceleration scheme, further improve the utilization efficiency of the acceleration platforms, and make the load on the acceleration platforms more balanced.

[0075] In some embodiments of the present invention, the score corresponding to each combined acceleration scheme can be calculated by setting power consumption weights and acceleration time weights, thereby determining the optimal acceleration scheme. For example, if the power consumption of attribute feature information labeled X is 4W and the acceleration time is 3S on acceleration platform A, and the power consumption is 3W and the acceleration time is 4S on acceleration platform B, after calculation based on the weights, the combined acceleration scheme with the highest score is: accelerate for 2S on acceleration platform A and accelerate for the remaining time on acceleration platform B, achieving the effect of minimizing both power consumption and acceleration time. This serves as the traffic distribution strategy for attribute feature information labeled X. In this way, different acceleration platforms can be flexibly combined, traffic distribution strategies can be formulated, and the resources of acceleration platforms can be fully utilized, which is beneficial to a more balanced load across each acceleration platform.

[0076] It is understood that the optimal acceleration scheme can combine different acceleration platforms for acceleration. In some embodiments of the present invention, after quantitative calculation, it may be that the power consumption and acceleration time are minimized when a single acceleration platform is used for acceleration. In this case, the optimal acceleration scheme is to continuously accelerate on the same acceleration platform.

[0077] refer to Figure 3 In some embodiments of the business feature-based targeted traffic diversion method of the present invention, before S100, the method further includes:

[0078] S010: Obtain the business access requests of each user;

[0079] S020: Obtain the corresponding network status information based on the service access request;

[0080] S030: Based on the network status information, determine whether accelerated processing is required;

[0081] S040: When accelerated processing is required, the business information corresponding to the business access request is marked as accelerated business information.

[0082] Since not every inter-network service requires acceleration, if it can operate within its appropriate attribute range without acceleration, then acceleration is unnecessary. Therefore, based on the user's service access request, the corresponding network status information is obtained to determine whether the user's service requires acceleration. If acceleration is required, the corresponding service information is marked as accelerated service information, and subsequent acceleration and traffic offloading will be performed to ensure the quality of service. If acceleration is not required, normal routing will be used, without passing through the acceleration platform.

[0083] Network status information can include latency, packet loss, jitter, retransmission rate, etc. It can be determined whether acceleration processing is needed by comparing thresholds. For example, if any one of the latency, packet loss, jitter, or retransmission rate exceeds a preset threshold, it is determined that acceleration processing is needed. Conversely, if it does not exceed the preset threshold or is within a suitable range, it is determined that acceleration processing is not needed.

[0084] refer to Figure 4 In some embodiments of the business feature-based targeted traffic routing method of the present invention, after S600, the method further includes:

[0085] S700: Obtain the corresponding service information based on the acceleration service information;

[0086] S800: Perform quality analysis on the service information and obtain the analysis results;

[0087] S900: Adjust the traffic diversion strategy information based on the analysis results;

[0088] The business service information includes business logs and network status, and the analysis results are obtained based on traffic geographic distribution, network quality, and application identification.

[0089] After traffic is routed according to the traffic splitting strategy, the service status of the services is tracked to obtain service information. Quality analysis is then performed on the service logs and network status within this information to determine the service quality during service execution. Based on the analysis results, the traffic splitting strategy is adjusted to optimize the acceleration service quality and ensure greater stability. Therefore, after initial traffic splitting based on acceleration predictions from a predictive model, adaptively adjusting the traffic splitting strategy based on actual acceleration performance helps to stabilize acceleration and guarantee service quality.

[0090] Business service information can include accelerated business session information, such as source and destination IP addresses, latency, jitter, packet loss, and retransmission rate, as well as business logs generated during the process. Quality analysis and optimization are performed from three dimensions: geographical distribution of business traffic, content service quality, and application identification and analysis.

[0091] Traditional network quality monitoring mostly focuses on network layer analysis. However, users' perception of network service quality has evolved from network metrics such as bandwidth to the service experience of a specific business. In short, users have shifted their focus from network indicators like bandwidth and download speed to the perceived quality of service within applications, such as video playback smoothness and webpage loading latency. This invention provides service quality analysis at the application layer, enabling comprehensive monitoring and analysis of service information such as intranet latency, internet latency, and server response latency to ensure the quality of service experienced by users at the application layer.

[0092] refer to Figure 5 In some embodiments of the service-feature-based targeted traffic diversion method of the present invention, S800 includes: S810: determining the intra-network latency and the extra-network latency based on the service information;

[0093] S820: The sum of the intra-network latency and the extra-network latency is used as the application latency;

[0094] S830: Determine whether there is a quality issue based on the application latency;

[0095] S840: When there is a quality issue, determine the latency quality issue information based on the intra-network latency and the extra-network latency; S850: Add the latency quality issue information to the analysis result information.

[0096] During quality analysis, based on both intra-network and extra-network latency, when quality issues such as stuttering or jitter occur, it's determined whether the source of the quality issue is within or outside the network. This latency quality issue information is then added to the analysis results. This allows for subsequent adjustments to traffic routing strategies based on the quality issue information, bypassing the issue to improve service quality. For example, if the quality issue is outside the network, the acceleration platform can be modified to change the inter-network exit point to bypass it. In this way, by determining whether quality issues occur based on intra-network and extra-network latency, and when such issues do occur, latency quality issue information is added to the analysis results.

[0097] It is understandable that the aforementioned latency and quality issues are only part of the quality analysis process, which may also include log analysis, comprehensive data analysis, and other processes. Log analysis analyzes whether the user's DNS configuration is incorrect based on user IP information and the type of application accessed, and analyzes whether the scheduling strategy is unreasonable based on the ISP affiliation and geographical location of the application's destination address. Comprehensive data analysis combines network quality and log data to analyze the causes of problems such as excessive latency within the network.

[0098] In some embodiments of the present invention, the method further includes: generating display information based on the analysis results, wherein the display information is used to display the service quality of the service.

[0099] By displaying information, various data obtained from quality analysis can be presented, facilitating subsequent operational and maintenance analyses. The displayed information can include end-to-end quality analysis, application type analysis, and more.

[0100] refer to Figure 6 In some embodiments of the business feature-based targeted traffic diversion method of the present invention, S300 includes: S310: randomly selecting a corresponding number of attribute feature information as the first centroid according to a preset number of clusters;

[0101] S320: The attribute feature information is used to calculate the centroid distance to each of the first centroids;

[0102] S330: Based on the centroid distance, the attribute feature information and the nearest first centroid are grouped into the same set to form a cluster set;

[0103] S340: Based on each of the cluster sets, obtain the second centroid and calculate the change between the second centroid and the corresponding first centroid;

[0104] S350: When the change is greater than or equal to a preset threshold, the second centroid is used as the new first centroid, and the second centroid is reacquired.

[0105] S360: When the change is less than the preset threshold, a business feature label is generated based on the second centroid, and the business feature label is added to each attribute feature information in the cluster set corresponding to the second centroid.

[0106] Based on the attribute features of each acceleration service, a clustering algorithm is used to group these features. The centroids of the cluster sets are then calculated as service feature labels. In this way, clustering is performed based on similarity according to the attribute features of each acceleration service, eliminating the need for manual pre-setting of service grouping information and enabling flexible, automatic grouping and generation of corresponding service feature labels.

[0107] The clustering algorithm can specifically be an implementation of an algorithm with clustering function, such as the K-Means algorithm.

[0108] The preset prediction model is obtained through pre-training. In some embodiments of the present invention, the prediction model can be obtained through the following steps:

[0109] Obtain historical inter-network service information to extract attribute characteristic information.

[0110] The K-Means clustering algorithm is used to group services with the same attribute features together, and labels are added to each group of services with the same attribute features to form sample data.

[0111] The sample data is divided into a test set and a training set according to a ratio (e.g., a 7:3 ratio). The training set data is input into the neural network for training. When the difference between the loss function data and the label data is within a preset threshold range, the neural network converges, and the initial neural network model is obtained.

[0112] The test set data is input into the initial neural network model. When the accuracy of the test set reaches a preset threshold, the prediction model is obtained.

[0113] The following describes the service feature-based targeted traffic splitting device provided by the present invention. The service feature-based targeted traffic splitting device described below and the service feature-based targeted traffic splitting method described above can be referred to and correspond to each other.

[0114] This invention also provides a service-characteristic-based targeted traffic splitting device, comprising:

[0115] The acquisition module is used to acquire acceleration service information and acceleration platform information;

[0116] The processing module is used to obtain corresponding attribute feature information based on the acceleration service information, group the attribute feature information into groups, and add corresponding service feature tags to the attribute feature information in the same group.

[0117] The analysis module is used to input the attribute feature information with the business feature tag and the acceleration platform information into a preset prediction model to obtain acceleration prediction information corresponding to the attribute feature information.

[0118] The traffic splitting module is used to determine traffic splitting strategy information based on the acceleration prediction information, and to allocate the service traffic corresponding to the acceleration service information to the corresponding acceleration platform according to the traffic splitting strategy information.

[0119] The acquisition module, the processing module, the analysis module, and the traffic splitting module cooperate to execute the above-described targeted traffic splitting method based on business characteristics.

[0120] The acquisition module obtains acceleration service information and acceleration platform information, enabling it to identify the services the user needs to accelerate and the various acceleration platforms that can provide acceleration services. The processing module extracts the attribute feature information corresponding to each service based on the acceleration service information to determine the acceleration requirements. The processing module groups the attribute feature information to group similar services together and adds service feature tags for easy differentiation and processing later. The analysis module processes the attribute feature information with service feature tags and acceleration platform information using a prediction model to predict the power consumption and acceleration time of each service on different acceleration platforms, generating acceleration prediction information. Finally, based on the acceleration prediction information and the principle of minimizing power consumption and acceleration time, the analysis module determines the traffic distribution strategy information to allocate the service traffic corresponding to the acceleration service information to the corresponding acceleration platforms according to the traffic distribution strategy information.

[0121] By grouping different business types and adding business feature tags, it is easier to perform traffic splitting more accurately. At the same time, by predicting the power consumption and acceleration time of each business on different acceleration platforms, traffic splitting strategy information is determined based on the principle of minimizing power consumption and acceleration time. After the business traffic is split and accelerated according to the traffic splitting strategy information, the overall load of power consumption and acceleration time of each acceleration platform is more balanced, which is conducive to providing stable acceleration services and making full use of the resource performance of the acceleration platform.

[0122] refer to Figure 7 The present invention also provides a service feature-based targeted traffic splitting system, including a traffic splitting gateway and an acceleration platform. The traffic splitting gateway is connected to the acceleration platform and is used to execute the above-described service feature-based targeted traffic splitting method.

[0123] The acceleration platform can include a static service optimization module, a dynamic service optimization module, a log analysis module, and a network quality analysis module. The static service optimization module includes content network modules such as caching and CDN, which localize static services such as mobile application updates, patch downloads, and P2P downloads in user inter-network traffic, reducing inter-network access latency and improving user experience. The dynamic service optimization module connects to high-quality inter-network exits, sets up firewall devices for NAT, and converts discrete user source IP addresses into a centralized address range, optimizing dynamic resources such as key websites and games on the inter-network. The log analysis module analyzes the composition and flow of inter-network traffic, providing reliable data for future resource introduction. The network quality analysis module analyzes the intra-network latency, inter-network latency, and application response latency of various inter-network application accesses, providing reliable data for optimizing application access quality.

[0124] Figure 8 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 8 As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840. The processor 810, communications interface 820, and memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions from the memory 830 to execute the aforementioned service-characteristic-based targeted traffic routing method.

[0125] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0126] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the business feature-based targeted traffic diversion method provided by the above methods.

[0127] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the business feature-based targeted traffic routing methods provided by the above methods.

[0128] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0129] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0130] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0131] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0132] All actions involving the acquisition of signals, information, or data in this application are carried out in accordance with the relevant data protection laws and policies of the country where the application is located, and with the authorization of the owner of the relevant device.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for traffic feature-based steering and offloading, characterized in that, The method comprises the following steps: obtaining acceleration service information and acceleration platform information; obtaining corresponding attribute feature information according to the acceleration service information; grouping the attribute feature information, and adding corresponding service feature labels to the attribute feature information in the same group; inputting the attribute feature information with the service feature labels and the acceleration platform information into a preset prediction model to obtain acceleration prediction information corresponding to the attribute feature information, wherein the acceleration prediction information represents the power consumption and acceleration time of the service traffic corresponding to the acceleration service information on different acceleration platforms; determining a distribution strategy information based on the acceleration prediction information, wherein the distribution strategy information is determined based on the minimization of power consumption and acceleration time; allocating the service traffic corresponding to the acceleration service information to the corresponding acceleration platform according to the distribution strategy information.

2. The method for service feature-based traffic steering and routing according to claim 1, wherein, The method further comprises the following steps before the step of obtaining the acceleration service information and the acceleration platform information: obtaining service access requests of each user; 3. The method for service feature-based traffic steering and routing of claim 1, wherein, obtaining corresponding network state information according to the service access requests; determining whether acceleration processing is needed according to the network state information; when acceleration processing is needed, marking the service information corresponding to the service access requests as acceleration service information. The method further comprises the following steps after the step of allocating the service traffic corresponding to the acceleration service information to the corresponding acceleration platform according to the distribution strategy information: obtaining corresponding service service information according to the acceleration service information; 4. The method for traffic feature-based steering of claim 1, wherein, performing quality analysis on the service service information to obtain analysis result information; adjusting the distribution strategy information according to the analysis result information; wherein the service service information comprises service logs and network states, and the analysis result information is obtained based on the analysis of traffic geographical distribution, network quality, and application identification. The method further comprises the following steps of performing quality analysis on the service service information to obtain analysis result information: determining in-network latency and out-of-network latency according to the service service information; 5. The method for service feature-based traffic steering and routing of claim 4, wherein, taking the sum of the in-network latency and the out-of-network latency as application latency; determining whether there is a quality difference according to the application latency; when there is a quality difference, determining latency quality difference point information according to the in-network latency and the out-of-network latency; adding the latency quality difference point information to the analysis result information. The method further comprises the following steps of grouping the attribute feature information and adding corresponding service feature labels to the attribute feature information in the same group: randomly selecting a corresponding number of attribute feature information as first centroids according to a preset number of clusters; 6. The method for traffic feature-based steering of claim 1, wherein, calculating the centroid distance between the attribute feature information and each first centroid respectively; ​ ​ According to the centroid distance, the attribute feature information is divided into the same set with the closest first centroid, forming a clustering set; According to each of the clustering set, a second centroid is obtained, and a change between the second centroid and the corresponding first centroid is calculated; When the change is greater than or equal to a preset threshold, the second centroid is taken as a new first centroid, and a second centroid is re-obtained; When the change is less than the preset threshold, a business feature label is generated according to the second centroid, and the business feature label is added to each of the attribute feature information in the clustering set corresponding to the second centroid.

7. A device for directing traffic based on traffic characteristics, characterized in that Comprise: An acquisition module is configured to acquire acceleration service information and acceleration platform information; A processing module is configured to acquire corresponding attribute feature information according to the acceleration service information, group each of the attribute feature information, and add corresponding business feature labels to the attribute feature information in the same group; An analysis module is configured to input the attribute feature information with the business feature labels and the acceleration platform information into a preset prediction model, and acquire acceleration prediction information corresponding to the attribute feature information; A shunting module is configured to determine shunting strategy information according to the acceleration prediction information, and distribute service traffic corresponding to the acceleration service information to a corresponding acceleration platform according to the shunting strategy information; The acquisition module, the processing module, the analysis module, and the shunting module cooperate to perform the business feature-based directional shunting method of any one of claims 1 to 6.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the business feature-based directional shunting method of any one of claims 1 to 6. 9.A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the business feature-based directional shunting method of any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the business feature-based directional shunting method of any one of claims 1 to 6. The computer program is executed by the processor to implement the business feature-based directional shunting method of any one of claims 1 to 6.

Images