Cross-domain communication network traffic classification method and device, equipment and storage medium
The traffic classification model trained in deep feature is used to identify and classify different data flows in cross-domain communication networks, solving the problems of complexity and priority differences in data flow types in cross-domain communications, and achieving more efficient and stable network services.
Patent Information
- Application Number
- CN202411268739.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2024-12-06
AI Technical Summary
The transmission bandwidth and delay requirements for different data stream types in cross-domain communication networks are different, and the priority of data streams is also different in different scenarios, making it difficult for the existing technology to achieve high-reliability intelligent traffic classification, affecting the cross-domain communication network. stability.
Through the traffic classification model trained based on deep features, the traffic data to be analyzed is obtained and input into the pre-trained target model, and the depth features are extracted to identify complex patterns and dynamic behaviors, thereby improving the accuracy and stability of traffic classification.
Provide more reliable and efficient network services in complex cross-domain communication environments, improve the accuracy and robustness of traffic classification models, reduce computing resource consumption, and prevent overfitting.
Smart Images

Figure CN119089248A_ABST
Abstract
Description
Technical Field
[0001] Embodiments of the present invention relate to the technical field of cross-domain communication network traffic analysis, and in particular, to a cross-domain communication network traffic classification method, device, equipment and storage medium. Background Art
[0002] The technical difficulty of the cross-domain communication network intelligent traffic classification method lies mainly in the fact that cross-domain communication network traffic includes different data stream types, such as voice, text, traffic and video, and each type of data stream has different transmission bandwidth and latency requirements. In addition, depending on different cross-domain communication scenarios, each specific data stream may have different priorities. Therefore, achieving highly reliable cross-domain communication network intelligent traffic classification is crucial to building a stable cross-domain communication network.
[0003] Classification technology based on machine learning can classify business traffic by analyzing the statistical characteristics of business and applications. It can integrate the characteristics of various fields and then build a suitable feature data set for classification. Although machine learning technology has great potential, it is necessary to fully screen the data set and consider factors such as time dimension, computing power and business needs to ensure the accuracy and reliability of the classification model. Summary of the invention
[0004] The embodiments of the present invention provide a method, apparatus, device and storage medium for classifying cross-domain communication network traffic, which can improve the classification accuracy and stability of the model, thereby providing more reliable and efficient network services in a complex cross-domain communication environment.
[0005] In a first aspect, an embodiment of the present invention provides a method for classifying cross-domain communication network traffic, the method comprising:
[0006] Obtain the traffic data to be analyzed of the target communication network; input the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model obtained by deep feature training based on sample traffic data.
[0007] In a second aspect, an embodiment of the present invention provides a cross-domain communication network traffic classification device, the device comprising:
[0008] The to-be-analyzed traffic acquisition module is used to acquire the to-be-analyzed traffic data of the target communication network;
[0009] The traffic classification module is used to input the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model obtained by deep feature training based on sample traffic data.
[0010] In a third aspect, an embodiment of the present invention provides a computer device, the computer device comprising:
[0011] one or more processors;
[0012] A memory for storing one or more programs;
[0013] When the one or more programs are executed by the one or more processors, the one or more processors implement the cross-domain communication network traffic classification method described in any embodiment.
[0014] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the cross-domain communication network traffic classification method described in any embodiment.
[0015] The technical solution provided by the embodiment of the present invention obtains the traffic data to be analyzed of the target communication network; inputs the traffic data to be analyzed into a pre-trained target traffic classification model to obtain the target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model trained based on the deep features of the sample traffic data. The technical solution of the embodiment of the present invention can improve the classification accuracy and stability of the model by training the traffic classification model based on the deep features of the traffic data, thereby providing more reliable and efficient network services in complex cross-domain communication environments. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 It is a flow chart of a cross-domain communication network traffic classification method provided by an embodiment of the present invention;
[0017] Figure 2 It is a flow chart of another cross-domain communication network traffic classification method provided by an embodiment of the present invention;
[0018] Figure 3 It is a workflow diagram for classifying cross-domain communication network traffic provided by an embodiment of the present invention;
[0019] Figure 4 It is a structural schematic diagram of a cross-domain communication network traffic classification device provided by an embodiment of the present invention;
[0020] Figure 5 It is a structural schematic diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0021] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.
[0022] Figure 1 It is a flow chart of a cross-domain communication network traffic classification method provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the scenario of classifying the traffic type of cross-domain communication network traffic data. The method can be executed by a cross-domain communication network traffic classification device, which can be implemented by software and / or hardware.
[0023] like Figure 1 As shown, the cross-domain communication network traffic classification method includes the following steps:
[0024] S110: Obtain traffic data to be analyzed of the target communication network.
[0025] The target communication network may be a cross-domain communication network that needs to classify network traffic. The traffic data to be analyzed may be traffic data corresponding to the target communication network that needs to be distinguished by traffic type. Exemplarily, the cross-domain communication traffic between the target communication network and other cross-domain communication networks, and the cross-domain communication traffic in the target communication network may be used as the traffic data to be analyzed.
[0026] S120: Input the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed.
[0027] Among them, the target traffic classification model can be a model that is finally obtained to classify the cross-domain communication network traffic. Among them, the target traffic classification model can be trained based on the deep features of the sample traffic data. The deep features can be features that represent hidden patterns and relationships in the data. In this method, the deep features characterize the heterogeneous information in the cross-domain environment and the relationship between different traffic types. By extracting this feature, the model can identify complex patterns in the traffic data, capture the dynamic behavior and distribution characteristics of the cross-domain communication network traffic data, and convert the original traffic data into a feature type that is easier for the deep learning model to recognize, thereby improving the accuracy and robustness of the traffic classification model.
[0028] Further, the target traffic classification result may be a traffic classification result of the target communication network. Exemplarily, the target traffic classification result includes types such as original data type, traffic alarm type, result information type, control instruction type, and status information type. Specifically, the traffic data to be analyzed may be input into a pre-trained target traffic classification model to obtain the target traffic classification result of the traffic data to be analyzed.
[0029] The technical solution provided by the embodiment of the present invention obtains the traffic data to be analyzed of the target communication network; inputs the traffic data to be analyzed into a pre-trained target traffic classification model to obtain the target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model trained based on the deep features of the sample traffic data. The technical solution of the embodiment of the present invention can improve the classification accuracy and stability of the model by training the traffic classification model based on the deep features of the traffic data, thereby providing more reliable and efficient network services in a complex cross-domain communication environment.
[0030] Figure 2 It is another flow chart of a cross-domain communication network traffic classification method provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the scenario of classifying the traffic type of cross-domain communication network traffic data. Based on the above embodiment, this embodiment further illustrates how to obtain a target traffic classification model based on deep feature training of sample traffic data. The device can be implemented by software and / or hardware and integrated in a computer device with application development function.
[0031] like Figure 2 As shown, the cross-domain communication network traffic classification method includes the following steps:
[0032] S210: Input the sample traffic data into a preliminary traffic classification model to obtain a sample traffic classification result.
[0033] Among them, the preliminary traffic classification model can be an original cross-domain communication network traffic classification model that has not been pre-trained. Furthermore, the sample traffic data can be preset traffic data about the target communication network. The sample traffic data can contain corresponding preset type labels for evaluating the classification results of the initial traffic classification model and adjusting the model according to the evaluation results. The sample traffic classification result can be the classification result of the traffic type of the sample traffic data by the preliminary traffic classification model. The preset preset type label can be a traffic type label that is pre-marked for the sample traffic data. Among them, the sample traffic data includes a preset type label. The preset type label includes: at least one of the original data type, traffic alarm type, result information type, control instruction type and status information type.
[0034] Optionally, the preliminary traffic classification model includes a feature extraction subnetwork and a feature analysis subnetwork. The sample traffic data is input into the preliminary traffic classification model to obtain a sample traffic classification result, including: inputting the sample traffic data into the feature extraction subnetwork to obtain deep feature data of the sample traffic data; inputting the deep feature data into the feature analysis subnetwork to obtain a sample traffic classification result.
[0035] Among them, the deep feature data can be data about the deep features of the sample traffic data. Among them, the deep features can be features that represent hidden patterns and relationships in the data. In this method, the deep feature characterizes the heterogeneous information in the cross-domain environment and the relationship between different traffic types. By extracting this feature, the model can identify complex patterns in the traffic data, capture the dynamic behavior and distribution characteristics of the traffic data of the cross-domain communication network, and convert the original traffic data into a feature type that is easier to identify by the deep learning model, thereby improving the accuracy and robustness of the traffic classification model. Furthermore, the deep feature data can be input into the feature analysis subnetwork so that the feature analysis subnetwork performs data analysis on the deep feature data to obtain the sample traffic classification result.
[0036] Optionally, the preliminary traffic classification model also includes a feature preprocessing unit, which, before inputting the deep feature data into the feature analysis subnetwork, also includes: inputting the deep feature data into the feature preprocessing unit so that the feature preprocessing unit determines the influence parameters and accuracy parameters corresponding to the deep feature data, and adjusts the analysis weights corresponding to the deep feature data based on the influence parameters and accuracy parameters.
[0037] Among them, the influence parameter indicates the degree of contribution of the deep feature data to the classification performance of the model. The larger the influence parameter of the deep feature, the higher the contribution of the deep feature to the classification performance. Furthermore, if the influence parameter of the deep feature is low, the analysis weight corresponding to the deep feature can be reduced accordingly. Exemplarily, if the influence parameter of a certain depth feature is less than the preset depth parameter threshold, and the amount of calculation corresponding to the depth feature is large (for example, specific message lengths and traffic delays require complex calculations), the analysis weight corresponding to the depth feature can be reduced to 0, that is, the depth feature is excluded. By excluding non-essential depth features, features that contribute more to classification, such as source IP and destination IP, are retained, which enables the model to effectively classify traffic while reducing computing resource consumption.
[0038] Among them, the analysis weight can be a parameter indicating the importance of the analysis of the deep feature. That is, if the analysis weight of the deep feature is high, the deep feature can be analyzed in detail, and the feature analysis result has a greater impact on the final output result; if the analysis weight of the deep feature is low, the deep feature can be briefly analyzed, and the feature analysis result has a smaller impact on the final output result. Furthermore, the accuracy parameter indicates the degree of influence of the deep feature data on the classification accuracy of the model. The larger the accuracy parameter of the deep feature, the higher the degree of influence of the deep feature on the classification accuracy of the model. Specifically, if the accuracy parameter of the deep feature is small, the analysis weight corresponding to the deep feature can be reduced accordingly; if the accuracy parameter is large, the analysis weight corresponding to the deep feature can be increased accordingly. This improvement not only enhances the classification accuracy of the model on complex data sets, but also improves the robustness of the model, effectively preventing the occurrence of overfitting.
[0039] Optionally, before training the preliminary traffic classification model based on the sample traffic data, the sample traffic data can also be divided based on the traffic type to obtain multiple sample traffic clusters; for each sample traffic cluster, the sample distribution density corresponding to the sample traffic cluster is determined, and when the sample distribution density is less than a reference density threshold, the sample traffic cluster is sampled and processed, and the obtained sampling data is added to the sample traffic cluster.
[0040] Among them, the sample traffic cluster can be a collection of sample traffic data of the same type. Specifically, based on the preset type label corresponding to the sample traffic data, the data traffic of the same type can be divided into the same cluster, thereby obtaining a sample traffic cluster. Furthermore, the sample distribution density can be a parameter indicating the density of sample data in the sample traffic cluster. When the sample distribution density is less than the reference density threshold, it means that there are fewer traffic data in the sample traffic cluster, so the sample traffic cluster can be sampled to generate new sample data, and the obtained sample data can be added to the sample traffic cluster. Through this process, the model can better learn the characteristics of minority samples during training, thereby improving its ability to recognize minority traffic.
[0041] S220. Determine a classification loss function according to the sample traffic classification result and a preset type label of the sample traffic data.
[0042] The classification loss function may be a function representing the error degree of the preliminary traffic classification model in traffic classification. The error degree of the preliminary traffic classification model in traffic classification may be known through the classification loss function. Specifically, the sample traffic classification result is compared with the preset type label of the sample traffic data, and then the classification loss function is determined according to the comparison result.
[0043] S230. Adjust the preliminary traffic classification model according to the classification loss function to obtain a target traffic classification model.
[0044] The target traffic classification model may be a model finally obtained for classifying cross-domain communication network traffic. Specifically, the model parameters in the preliminary traffic classification model may be adjusted according to the classification loss function to obtain the target traffic classification model.
[0045] Optionally, a preliminary traffic classification model and sample traffic data may be obtained; the preliminary traffic classification model may be unsupervisedly trained based on the sample traffic data to obtain a target traffic classification model. For example, a clustering algorithm (such as K-means, DBSCAN) may be used to perform preliminary classification of cross-domain communication network traffic, and secondary classification may be performed in combination with supervised learning, thereby reducing reliance on labeled data.
[0046] S240: Obtain the traffic data to be analyzed of the target communication network.
[0047] The target communication network may be a cross-domain communication network that needs to classify network traffic. The traffic data to be analyzed may be traffic data corresponding to the target communication network that needs to be distinguished by traffic type. Exemplarily, the cross-domain communication traffic between the target communication network and other cross-domain communication networks, and the cross-domain communication traffic in the target communication network may be used as the traffic data to be analyzed.
[0048] S250: Input the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed.
[0049] The target traffic classification result may be a traffic classification result of the target communication network. Exemplarily, the target traffic classification result includes types such as original data type, traffic alarm type, result information type, control instruction type, and status information type. Specifically, the traffic data to be analyzed may be input into a pre-trained target traffic classification model to obtain the target traffic classification result of the traffic data to be analyzed.
[0050] Exemplarily, the traffic data to be analyzed can be input into the feature extraction subnetwork in the target traffic classification model to obtain the target deep feature data; the target deep feature data can be input into the feature analysis subnetwork in the target traffic classification model to obtain the target traffic classification result. The target deep feature data represents the data of the deep features corresponding to the traffic data to be analyzed.
[0051] For example, in order to better understand the technical solution provided by the present invention, a specific embodiment is introduced below: Figure 3 FIG. 1 is a flowchart of a cross-domain communication network traffic classification process provided by an embodiment of the present invention. Figure 3 As shown in Figure 1, the workflow for cross-domain communication network traffic classification includes two aspects: model training and model use. The specific model training includes the following steps:
[0052] 1) Preparation of cross-domain communication network traffic data and classification labels: First, traffic data sets from the cross-domain communication network are collected and divided into five categories, with each category of traffic data having a specific priority.
[0053] 2) Data is divided into training set and test set: The cross-domain communication network traffic data is divided into training set and test set, where 80% of the data is used as the training set and 20% of the data is used as the test set.
[0054] 3) Preprocessing of cross-domain communication network traffic data: Preprocessing of real-time cross-domain communication network traffic data, mainly including: extracting session data (extracting network session data from traffic data), traffic balancing (balancing traffic distribution and deleting redundant data), image generation (converting byte stream network traffic data into images that can be recognized by the model) and IDX conversion (converting to the IDX file format commonly used in machine learning).
[0055] 4) Feature extraction of cross-domain communication network traffic data (extracting deep feature representation of cross-domain communication network traffic data): In cross-domain communication network traffic classification, deep features usually refer to complex feature representations automatically extracted by deep learning models, which can capture hidden patterns and relationships in the data. Specifically, deep features in this method characterize the heterogeneous information in the cross-domain environment and the relationship between different traffic types. This feature is automatically extracted in multiple hidden layers through a deep neural network, gradually forming a high-level abstract representation. By extracting this feature, the model can identify complex patterns in traffic data, capture the dynamic behavior and distribution characteristics of cross-domain communication network traffic data, and convert the original traffic data into feature types that are easier for deep learning models to recognize, thereby improving the accuracy and robustness of the traffic classification model.
[0056] 5) Input cross-domain communication network traffic data and related features: Input the preprocessed data and related features processed by feature engineering into the cross-domain communication network traffic classification model.
[0057] 6) Build a cross-domain communication network traffic classification model based on improved XGBoost.
[0058] 7) Use the training set to train the constructed classification model: Use the training set to train the cross-domain communication network traffic classification model, so that the model gradually adjusts parameters according to the feature data and the corresponding label data to improve the data fitting ability.
[0059] 8) Use the test set to test the trained classification model: Use the test set to test and evaluate the trained cross-domain communication network traffic classification model, obtain the classification results predicted by the model, and calculate indicators such as the model classification accuracy to evaluate the performance of the model.
[0060] 9) Model iterative tuning: Optimize the model according to the test evaluation results, and improve the performance of the cross-domain communication network traffic classification model through update iteration and model tuning until the termination conditions are met, such as reaching the specified performance indicators or the model convergence performance no longer improves.
[0061] 10) Save and export training network parameters: On the premise that the model iteration tuning termination conditions are met, save the model and export the training network parameters of the cross-domain communication network traffic classification model.
[0062] 11) Complete the cross-domain communication network traffic classification model training.
[0063] Furthermore, the model usage includes the following steps:
[0064] 1) Deploy and load the trained network parameters: After the model training is completed, import the trained cross-domain communication network traffic model, deploy and load its network parameters.
[0065] 2) Construct a cross-domain communication network traffic interface: periodically collect cross-domain communication network information according to the cross-domain communication network node controller, and construct an API interface to process the input of cross-domain communication network traffic.
[0066] 3) Preprocessing of real-time cross-domain communication network traffic data: Preprocessing of real-time cross-domain communication network traffic data mainly includes: extracting session data, traffic balancing, image generation and IDX conversion.
[0067] 4) Feature extraction of real-time cross-domain communication network traffic data: Extract deep feature representation of real-time cross-domain communication network traffic for subsequent model inference and classification.
[0068] 5) Input cross-domain communication network traffic data and related characteristics.
[0069] 6) Classification using the trained intelligent classification model: Using the trained cross-domain communication network traffic classification model, different types of cross-domain communication network traffic data are intelligently classified in real time according to the bandwidth and latency requirements of the cross-domain communication network.
[0070] 7) Generate cross-domain communication network traffic classification results: For each cross-domain communication network traffic data sample, the intelligent classification model generates a predicted category or category probability result.
[0071] 8) Output and application of cross-domain communication network traffic classification results: Output the cross-domain communication network traffic classification results and make a visual display, and use them as direct or indirect information input for subsequent intelligent decision-making.
[0072] 9) End the cross-domain communication network traffic classification model reasoning.
[0073] The technical solution provided by the embodiment of the present invention obtains a sample traffic classification result by inputting sample traffic data into a preliminary traffic classification model; determines a classification loss function based on the sample traffic classification result and a preset type label of the sample traffic data; adjusts the preliminary traffic classification model based on the classification loss function to obtain a target traffic classification model; obtains traffic data to be analyzed of a target communication network; inputs the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed. The technical solution of the embodiment of the present invention can improve the classification accuracy and stability of the model by training the traffic classification model based on the deep features of the traffic data, thereby providing more reliable and efficient network services in complex cross-domain communication environments.
[0074] Figure 4 It is a structural diagram of a cross-domain communication network traffic classification device provided by an embodiment of the present invention. The embodiment of the present invention can be applied to the scenario of classifying the traffic type of cross-domain communication network traffic data. The device can be implemented by software and / or hardware and integrated into a computer device with application development function.
[0075] like Figure 4 As shown, the cross-domain communication network traffic classification device includes: a traffic acquisition module 310 for analysis and a traffic classification module 320.
[0076] The traffic acquisition module 310 for analysis is used to obtain the traffic data for analysis of the target communication network; the traffic classification module 320 is used to input the traffic data for analysis into a pre-trained target traffic classification model to obtain the target traffic classification result of the traffic data for analysis; wherein the target traffic classification model is a model obtained by deep feature training based on sample traffic data.
[0077] The technical solution provided by the embodiment of the present invention obtains the traffic data to be analyzed of the target communication network; inputs the traffic data to be analyzed into a pre-trained target traffic classification model to obtain the target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model trained based on the deep features of the sample traffic data. The technical solution of the embodiment of the present invention can improve the classification accuracy and stability of the model by training the traffic classification model based on the deep features of the traffic data, thereby providing more reliable and efficient network services in complex cross-domain communication environments.
[0078] In an optional embodiment, the traffic classification module 320 is specifically used to: input the traffic data to be analyzed into the feature extraction subnetwork in the target traffic classification model to obtain target deep feature data; input the target deep feature data into the feature analysis subnetwork in the target traffic classification model to obtain the target traffic classification result.
[0079] In an optional embodiment, the cross-domain communication network traffic classification device also includes: a first model training module, used to: obtain a preliminary traffic classification model and sample traffic data; perform unsupervised training on the preliminary traffic classification model based on the sample traffic data to obtain the target traffic classification model.
[0080] In an optional embodiment, the cross-domain communication network traffic classification device also includes: a second model training module, used to: input sample traffic data into a preliminary traffic classification model to obtain a sample traffic classification result; wherein the sample traffic data includes a preset type label; determine a classification loss function based on the sample traffic classification result and the preset type label of the sample traffic data; adjust the preliminary traffic classification model according to the classification loss function to obtain the target traffic classification model.
[0081] In an optional embodiment, the preliminary traffic classification model includes a feature extraction subnetwork and a feature analysis subnetwork, and the second model training module includes: a sample traffic classification unit, used to: input the sample traffic data into the feature extraction subnetwork to obtain deep feature data of the sample traffic data; input the deep feature data into the feature analysis subnetwork to obtain the sample traffic classification result.
[0082] In an optional embodiment, the sample traffic classification unit includes: a feature processing subunit, used to: input the depth feature data into the feature preprocessing unit before inputting the depth feature data into the feature analysis subnetwork, so that the feature preprocessing unit determines the influence parameter and accuracy parameter corresponding to the depth feature data, and adjusts the analysis weight corresponding to the depth feature data based on the influence parameter and accuracy parameter; wherein the influence parameter represents the contribution degree of the depth feature data to the model classification performance; and the accuracy parameter represents the influence degree of the depth feature data on the model classification accuracy.
[0083] In an optional embodiment, the second model training module also includes: a sample preprocessing unit, which is used to: divide the sample flow data based on the flow type to obtain multiple sample flow clusters; for each sample flow cluster, determine the sample distribution density corresponding to the sample flow cluster, and when the sample distribution density is less than a reference density threshold, perform sampling processing on the sample flow cluster, and add the obtained sampling data to the sample flow cluster.
[0084] The cross-domain communication network traffic classification device provided in the embodiment of the present invention can execute the cross-domain communication network traffic classification method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0085] Figure 5 A schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Figure 5 A block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 5 The computer device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention. The computer device 12 can be any terminal device with computing capability and can be configured in a cross-domain communication network traffic classification device.
[0086] like Figure 5 As shown, the computer device 12 is in the form of a general-purpose computing device. The components of the computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that connects various system components (including the system memory 28 and the processing unit 16).
[0087] The bus 18 may be one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. For example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.
[0088] The computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 12, including volatile and non-volatile media, removable and non-removable media.
[0089] The system memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache 32. The computer device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 5 not shown, usually called a "hard drive"). Although Figure 5 Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The system memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.
[0090] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. Program modules 42 generally perform the functions and / or methods of the embodiments described herein.
[0091] The computer device 12 may also communicate across domains with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), one or more devices that enable a user to interact with the computer device 12, and / or any device that enables the computer device 12 to communicate across domains with one or more other computing devices (e.g., network cards, modems, etc.). Such cross-domain communication may be performed via an input / output (I / O) interface 22. Furthermore, the computer device 12 may also communicate across domains with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 20. Figure 5 As shown, the network adapter 20 communicates with other modules of the computer device 12 across domains via the bus 18. It should be understood that although Figure 5 Not shown, other hardware and / or software modules may be used in conjunction with computer device 12, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0092] The processing unit 16 executes various functional applications and data processing by running the program stored in the system memory 28, for example, implementing the cross-domain communication network traffic classification method provided in the embodiment of the present invention, the method comprising:
[0093] Obtain the traffic data to be analyzed of the target communication network; input the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model obtained by deep feature training based on sample traffic data.
[0094] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the cross-domain communication network traffic classification method provided by any embodiment of the present invention is implemented, including:
[0095] Obtain the traffic data to be analyzed of the target communication network; input the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model obtained by deep feature training based on sample traffic data.
[0096] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.
[0097] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.
[0098] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
[0099] Computer program code for performing the operations of the present invention may be written in one or more programming languages or combinations thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0100] It should be understood by those skilled in the art that the modules or steps of the present invention described above can be implemented by a general-purpose computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, optionally, they can be implemented by a program code executable by a computer device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.
[0101] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims
1. A cross-domain communication network traffic classification method, It is characterized in that include: Obtaining traffic data to be analyzed from the target communication network; Inputting the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed; The target traffic classification model is a model obtained by deep feature training based on sample traffic data.
2. The method according to claim 1, It is characterized in that The training process of the target traffic classification model includes: Obtain preliminary traffic classification model and sample traffic data; The preliminary traffic classification model is unsupervisedly trained based on the sample traffic data to obtain the target traffic classification model.
3. The method according to claim 1, It is characterized in that The training process of the target traffic classification model also includes: Inputting the sample traffic data into a preliminary traffic classification model to obtain a sample traffic classification result; wherein the sample traffic data includes a preset type label; Determine a classification loss function according to the sample flow classification result and a preset type label of the sample flow data; The preliminary traffic classification model is adjusted according to the classification loss function to obtain the target traffic classification model.
4. The method according to claim 3, It is characterized in that The preliminary traffic classification model includes a feature extraction subnetwork and a feature analysis subnetwork. The sample traffic data is input into the preliminary traffic classification model to obtain the sample traffic classification result, including: Inputting the sample flow data into the feature extraction subnetwork to obtain deep feature data of the sample flow data; The deep feature data is input into the feature analysis subnetwork to obtain the sample traffic classification result.
5. The method according to claim 4, It is characterized in that The preliminary traffic classification model further includes a feature preprocessing unit, which, before inputting the deep feature data into the feature analysis subnetwork, further includes: Inputting the depth feature data into the feature preprocessing unit so that the feature preprocessing unit determines an influence parameter and an accuracy parameter corresponding to the depth feature data, and adjusts an analysis weight corresponding to the depth feature data based on the influence parameter and the accuracy parameter; Among them, the influence parameter represents the contribution degree of the deep feature data to the model classification performance; the accuracy parameter represents the influence degree of the deep feature data on the model classification accuracy.
6. The method according to claim 3, It is characterized in that Before training the preliminary traffic classification model based on the sample traffic data, the method further includes: Dividing the sample traffic data based on traffic type to obtain multiple sample traffic clusters; For each sample flow cluster, determine the sample distribution density corresponding to the sample flow cluster. When the sample distribution density is less than a reference density threshold, perform sampling processing on the sample flow cluster and add the obtained sampling data to the sample flow cluster.
7. The method according to claim 1, It is characterized in that The step of inputting the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed includes: Inputting the traffic data to be analyzed into the feature extraction subnetwork in the target traffic classification model to obtain target deep feature data; The target deep feature data is input into the feature analysis subnetwork in the target traffic classification model to obtain the target traffic classification result.
8. A cross-domain communication network traffic classification device, It is characterized in that The device comprises: The to-be-analyzed traffic acquisition module is used to acquire the to-be-analyzed traffic data of the target communication network; The traffic classification module is used to input the traffic data to be analyzed into a pre-trained target traffic classification model to obtain a target traffic classification result of the traffic data to be analyzed; wherein the target traffic classification model is a model obtained by deep feature training based on sample traffic data.
9. A computer device, It is characterized in that The computer device comprises: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the cross-domain communication network traffic classification method as described in any one of claims 1-7.
10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the cross-domain communication network traffic classification method as described in any one of claims 1-7 is implemented.
Citation Information
Patent Citations
Network traffic classification method and device, electronic equipment and storage medium
CN112054967A
Data generation method and system for unbalanced voice data set
CN112861928A
Network traffic identification method and device, equipment and storage medium
CN113762377A
Network traffic classification method and device, storage medium and electronic equipment
CN114745335A
Decision tree-based traditional Chinese medicine constitution identification method, system and equipment and storage medium
CN118568604A
