A method, device, equipment and medium for detecting encrypted HTTPS traffic DDoS attacks based on deep learning

By collecting data and extracting features from encrypted HTTPS traffic using deep learning methods based on XDP and STANet, this approach addresses the privacy risks and high computational resource consumption issues inherent in existing DDoS attack detection technologies, achieving efficient and accurate attack detection.

CN118646596BActive Publication Date: 2024-11-08CHINA ACADEMY OF INFORMATION & COMM
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411096155.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-08
Estimated Expiration
2044-08-12

AI Technical Summary

Technical Problem

Existing technologies for detecting DDoS attacks under encrypted HTTPS traffic suffer from problems such as privacy risks during the decryption process, high computational resource consumption, limitations of detection algorithms, and low detection rates.

Method used

Traffic data is collected using XDP technology. The encrypted traffic data is preprocessed and features are extracted using the STANet deep learning model. The output score vector is generated using grayscale image conversion and attention mechanism, and detection is performed in conjunction with alarm notifications.

Benefits of technology

It achieves accurate identification of DDoS attacks on encrypted HTTPS traffic, improves the accuracy and targeting of detection, reduces computing resource consumption, and avoids privacy data leakage.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118646596B_ABST
    Figure CN118646596B_ABST
Patent Text Reader

Abstract

The disclosed embodiment discloses a method and apparatus, device and medium for detecting DDoS attacks on encrypted HTTPS traffic based on deep learning, wherein the method comprises: first, based on the XDP technology, traffic data is collected from network devices to generate data to be processed. Afterwards, based on the attack characteristics of the encrypted DDoS, the generated data to be processed is converted and extracted to generate encrypted traffic data. Then, based on the STANet deep learning model, the encrypted traffic data is processed to generate an output score vector, and the probability of the traffic data category is calculated based on the output score vector to generate a traffic detection report. Finally, based on the generated traffic detection report, a preset measure is responded to the alarm notification to complete the detection of the encrypted HTTPS traffic DDoS attack. The disclosed embodiment realizes the accurate identification of DDoS attacks and improves the accuracy of detection through the above-mentioned detection method. In application, compared with the existing detection method, it is more targeted and has a higher detection rate.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to encrypted traffic DDoS attack detection technology, and in particular to an encrypted HTTPS traffic DDoS attack detection method and apparatus, device and medium based on deep learning. Background Art

[0002] DDoS attack, or Distributed Denial of Service attack, is a means of network attack. Its purpose is to overload the target server or network resources through a large amount of network traffic or requests, so that normal users cannot access or use these resources. According to statistics, the proportion of web traffic encrypted with HTTPS has exceeded 90% worldwide, and the global network is moving towards a fully encrypted era. In the scenario of encrypted traffic, how to accurately detect DDoS attacks has posed new challenges to traditional network security detection and defense methods. It is also a difficult problem in the current industry.

[0003] The existing detection methods for encrypted traffic DDoS attacks mainly include: detection methods after decrypting the traffic, detection methods that rely on specific encryption protocols, and detection methods based on ensemble learning.

[0004] Among them: the detection method that performs traffic decryption involves sensitive information such as user privacy data and system private keys because it needs to decrypt HTTPS traffic, and the decryption and re-encryption process consumes more computing resources. The detection method that relies on a specific encryption protocol can only detect specific encryption protocols, and the detection algorithm has great limitations. The detection method based on ensemble learning does not make targeted detection of the characteristics of encrypted traffic DDoS attacks, resulting in a low detection rate.

[0005] Therefore, one or more methods are needed to solve the above problems.

[0006] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0007] The purpose of the present disclosure is to provide a method and apparatus, device and medium for detecting DDoS attacks on encrypted HTTPS traffic based on deep learning, thereby overcoming one or more problems caused by limitations and defects of related technologies at least to a certain extent.

[0008] According to one aspect of the present disclosure, a method for detecting DDoS attacks on encrypted HTTPS traffic based on deep learning is provided, comprising:

[0009] Based on XDP technology, traffic data is collected from network devices to generate data to be processed;

[0010] Based on the attack characteristics of encrypted DDoS, the encrypted traffic data is generated by converting and extracting the data to be processed;

[0011] Based on the STANet deep learning model, the encrypted traffic data is processed to generate an output score vector, and based on the output score vector, a traffic detection report is generated by calculating the probability of traffic data categories;

[0012] Based on the traffic detection report, the detection of encrypted HTTPS traffic DDoS attacks is completed by responding to the alarm notification with preset measures.

[0013] In an exemplary embodiment of the present disclosure, collecting traffic data from network devices includes:

[0014] Based on the splitting mirror technology, a bypass network is generated by deploying network bypass in the network device driver;

[0015] Based on the XDP technology, the flow data is collected in the bypass network to generate collected data;

[0016] Based on the eBPF technology, the collected data is optimized and defined to generate the data to be processed.

[0017] In an exemplary embodiment of the present disclosure, the data to be processed is converted and extracted, including:

[0018] Based on the HTTPS protocol information, generating related data by first cleaning the data to be processed;

[0019] Based on the protocol layering principle, layered data is generated by performing a second cleaning on the associated data;

[0020] Based on the abnormal situation correction principle, the layered data is cleaned for a third time to generate cleaned data;

[0021] Based on the principle of protocol standardization, the cleaned data is converted into a preset unified format to generate conversion data;

[0022] Based on the preset characteristic parameters, the encrypted traffic data x is generated by extracting the converted data, where x=[x1, x2, ..., xN], where N is the number of the preset characteristic parameters.

[0023] In an exemplary embodiment of the present disclosure, the preset characteristic parameters include flow ID, source IP, source port, destination IP, destination port, protocol, TLS version, encryption suite, TLS extension, elliptic curve parameters, elliptic curve point format, JA3 fingerprint, flow duration, timestamp, total number of request packets, total number of reply packets, total number of request bytes, total number of reply bytes, maximum request packet length, minimum request packet length, average request message length, request message length variance, maximum reply packet length, minimum reply packet length, average reply message length, reply message length variance, traffic pps, traffic bps, data packet interval time, data packet interval time standard deviation, maximum data packet interval time, minimum data packet interval time, number of PSH flag packets, number of URG flag packets, number of SYN flag packets, number of ACK flag packets, number of RST flag packets, number of FIN flag packets, uplink and downlink traffic ratio, shortest flow active time, longest flow active time, average flow active time, standard deviation of flow active time, shortest flow idle time, longest flow idle time, average flow idle time, standard deviation of flow idle time, and data tag.

[0024] In an exemplary embodiment of the present disclosure, the encrypted traffic data is processed to generate an output score vector, including:

[0025] Based on grayscale image conversion, a grayscale image is generated by converting the grayscale value of the encrypted traffic data;

[0026] Based on a convolutional neural network model, a grayscale image feature is generated by extracting features from the grayscale image;

[0027] Based on the spatial attention mechanism, the grayscale image feature is flattened into a two-dimensional matrix, and the position weight is calculated through the self-attention mechanism to generate a weight vector, and the weight vector is multiplied by the two-dimensional matrix to generate a spatial image feature;

[0028] Based on the time attention mechanism, after the time frame of the grayscale image feature is represented as a frame vector, the time weight is calculated through the self-attention mechanism to generate a weight frame, and the weight frame is multiplied by the frame vector to generate a time frame feature;

[0029] By performing multi-scale feature fusion on the spatial graph features and the time frame features, an output score vector z is generated, z=[ , ],in For normal traffic category, Indicates the attack traffic category.

[0030] In an exemplary embodiment of the present disclosure, based on the output score vector, a flow detection report is generated by calculating the flow data category probability, including:

[0031] Based on the output score vector, by the formula Calculate the probability of the normal traffic category to generate a normal probability ;

[0032] Based on the output score vector, by the formula Calculate the probability of the attack traffic category to generate the attack probability ;

[0033] in Output score for normal traffic, Output score for attack traffic;

[0034] If the attack probability Greater than normal probability , then the traffic data is judged to be attack traffic, and an attack traffic detection report is generated;

[0035] If the attack probability No greater than normal probability , then the traffic data is judged to be normal traffic and a normal traffic detection report is generated.

[0036] In an exemplary embodiment of the present disclosure, the detection of encrypted HTTPS traffic DDoS attacks is completed by responding to the alarm notification with preset measures, including:

[0037] Based on the normal flow detection report, the detection system continues to operate according to the preset operation logic;

[0038] Based on the attack traffic detection report, the detection system establishes alarm information and generates an alarm notification;

[0039] Based on the alarm notification, the automatic response system completes the detection of the encrypted HTTPS traffic DDoS attack by responding to the traffic data with preset measures.

[0040] In one aspect of the present disclosure, a deep learning-based encrypted HTTPS traffic DDoS attack detection device is provided, comprising:

[0041] The data collection module is used to collect traffic data from network devices and generate data to be processed;

[0042] A data preprocessing module, used to generate encrypted traffic data by converting and extracting the data to be processed;

[0043] A STANet deep learning detection algorithm module, used to generate an output score vector by processing the encrypted traffic data, and generate a traffic detection report by calculating the probability of traffic data categories based on the output score vector;

[0044] The alarm management module is used to complete the detection of encrypted HTTPS traffic DDoS attacks by responding to alarm notifications with preset measures.

[0045] In one aspect of the present disclosure, there is provided an electronic device, comprising:

[0046] Processor; and

[0047] A memory having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by the processor, implement the method according to any one of the above items.

[0048] In one aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method according to any one of the above items is implemented.

[0049] Based on the embodiment of the present disclosure, first, based on the XDP technology, traffic data is collected from network devices to generate data to be processed. Afterwards, based on the attack characteristics of encrypted DDoS, the generated data to be processed is converted and extracted to generate encrypted traffic data. Then, based on the STANet deep learning model, the encrypted traffic data is processed to generate an output score vector, and the probability of the traffic data category is calculated based on the output score vector to generate a traffic detection report. Finally, based on the generated traffic detection report, the alarm notification is responded to with preset measures to complete the detection of encrypted HTTPS traffic DDoS attacks. The embodiment of the present disclosure achieves accurate identification of DDoS attacks and improves the accuracy of detection through the above-mentioned detection method. In application, compared with existing detection methods, it is more targeted and has a higher detection rate.

[0050] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure.

[0051] The technical solution of the present disclosure is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings, which constitute a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0053] The present disclosure may be more clearly understood from the following detailed description with reference to the accompanying drawings, in which:

[0054] Figure 1 This is a flowchart of a method for detecting DDoS attacks on encrypted HTTPS traffic based on deep learning according to an embodiment of the disclosed method;

[0055] Figure 2 A STANet model detection logic flow chart of a method for detecting encrypted HTTPS traffic DDoS attacks based on deep learning according to an embodiment of the disclosed method;

[0056] Figure 3 This is a structural block diagram of a deep learning-based encrypted HTTPS traffic DDoS attack detection device according to an embodiment of the disclosed method;

[0057] Figure 4 A block diagram of an electronic device according to an embodiment of the method disclosed herein. DETAILED DESCRIPTION

[0058] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be comprehensive and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted.

[0059] In addition, the described features, structures or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced without one or more of the specific details, or other methods, components, materials, devices, steps, etc. may be adopted. In other cases, known structures, methods, devices, implementations, materials or operations are not shown or described in detail to avoid blurring the various aspects of the present disclosure.

[0060] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or these functional entities or parts of functional entities may be implemented in one or more software hardened modules, or these functional entities may be implemented in different networks and / or processor devices and / or microcontroller devices.

[0061] In the embodiments of the present disclosure, a method for detecting DDoS attacks on encrypted HTTPS traffic based on deep learning is first provided; Figure 1 As shown in , the encrypted HTTPS traffic DDoS attack detection method based on deep learning may include the following steps:

[0062] Step S110, based on the XDP technology, traffic data is collected from network devices to generate data to be processed.

[0063] Step S120, based on the attack characteristics of encrypted DDoS, the encrypted traffic data is generated by converting and extracting the data to be processed.

[0064] Step S130, based on the STANet deep learning model, by processing the encrypted traffic data, generating an output score vector, and based on the output score vector, by calculating the traffic data category probability, generating a traffic detection report.

[0065] Step S140, based on the traffic detection report, the detection of the encrypted HTTPS traffic DDoS attack is completed by responding to the alarm notification with preset measures.

[0066] Below, as Figure 2 As shown, a method for detecting DDoS attacks on encrypted HTTPS traffic based on deep learning in an embodiment of the present disclosure will be further described.

[0067] In step S110, traffic data may be collected from network devices based on the XDP technology to generate data to be processed.

[0068] In some optional embodiments of this example: First, based on the splitting and mirroring technology, the entire detection system is deployed in the network device driver in a serial or bypass manner, and when using network bypass, network traffic is received by splitting and mirroring.

[0069] Afterwards, based on the XDP technology, the flow data is collected in the bypass network to generate collection data. The XDP technology can process the flow data immediately when it reaches the network card driver layer, which is more efficient than the performance of the traditional kernel network stack processing. At the same time, it can support most high-speed network cards and has a wider range of applications. And because XDP is part of the Linux kernel, it also has a stable API and good compatibility.

[0070] Finally, combined with eBPF technology, the collection content and optimization process of the collected data can be defined by writing eBPF programs (for example, in terms of collection content, the collected original message collection protocol data can be defined to include: HTTP (encrypted part), TLS, TCP, IP, Ethernet, etc.), and finally a set of preliminarily collected and processed data to be processed is generated. The written eBPF program can make the reception and processing of message data have good adaptability and flexibility, and can also ensure re-mounting and updating at any time.

[0071] In step S120, based on the attack characteristics of encrypted DDoS, encrypted traffic data is generated by converting and extracting the data to be processed.

[0072] In some optional embodiments of this example: since the collected traffic is all processes of the target server, it is necessary to perform data cleaning on the collected data to be processed.

[0073] The first step is to obtain HTTPS protocol-related traffic based on HTTPS protocol information cleaning, including TCP connection establishment, TLS handshake phase, HTTPS application data interaction, TCP link disconnection and other traffic, while removing irrelevant protocol traffic such as ARP protocol traffic, UDP protocol traffic, etc., and finally obtain related data.

[0074] In the second step, the filtered associated data is cleaned again according to the protocol layering principle, and the data irrelevant to the model training is deleted or anonymized to obtain layered data.

[0075] The third step is to delete the abnormal data and correct the missing data based on the abnormal situation correction principle to obtain the cleaned data after complete cleaning.

[0076] Afterwards, based on the protocol standardization principle, the cleaned data is converted into a unified and standardized format to generate converted data. For example, the TCP protocol ID is set to 6 after standardization, and the time format can be uniformly set to YYYY-MM-DD hh:mm:ss format.

[0077] Finally, according to the characteristics of the encrypted DDoS attack, based on the designed and used characteristic parameters, the characteristic vector is extracted from the collected conversion data message field to generate encrypted traffic data x, x=[x1,x2,...,xN], where N is the number of the preset characteristic parameters.

[0078] In some optional embodiments of this example, the preset characteristic parameters include but are not limited to flow ID, source IP, source port, destination IP, destination port, protocol, TLS version, encryption suite, TLS extension, elliptic curve parameters, elliptic curve point format, JA3 fingerprint, flow duration, timestamp, total number of request packets, total number of reply packets, total number of request bytes, total number of reply bytes, maximum request packet length, minimum request packet length, average request message length, request message length variance, maximum reply packet length, minimum reply packet length, average reply message length, reply message length variance, traffic pps, traffic bps, data packet interval time, data packet interval time standard deviation, maximum data packet interval time, minimum data packet interval time, number of PSH flag packets, number of URG flag packets, number of SYN flag packets, number of ACK flag packets, number of RST flag packets, number of FIN flag packets, uplink and downlink traffic ratio, shortest flow active time, longest flow active time, average flow active time, standard deviation of flow active time, shortest flow idle time, longest flow idle time, average flow idle time, standard deviation of flow idle time, data tag, etc.

[0079] In step S130, based on the STANet deep learning model, the encrypted traffic data is processed to generate an output score vector, and based on the output score vector, the traffic data category probability is calculated to generate a traffic detection report.

[0080] In some optional embodiments of this example, based on grayscale image conversion, the encrypted traffic data is first converted into a two-dimensional pixel matrix, and the converted pixel values ​​are filled into the pixel matrix of the grayscale image to obtain a grayscale image of the encrypted traffic data. In a specific embodiment, after filling, a 64x64 grayscale image is selected as a case for further deep learning model detection.

[0081] In this embodiment, the adopted STANet model is composed of a two-dimensional convolutional neural network model, and an attention layer based on time and space attention is added to the convolutional neural network model.

[0082] After the grayscale image passes through the convolutional layer of the convolutional neural network model (the convolution kernel is 5*5 and the step size is 1), the ReLU activation function is used to introduce nonlinearity. The spatial dimension characteristics of the grayscale image are reduced through the maximum pooling layer (the pooling window size is 2*2), reducing the number of parameters and the amount of calculation. Finally, through the fully connected layer, the temporal and spatial features are abstracted into high-dimensional features (i.e., grayscale image features) and output to the attention layer.

[0083] In the attention layer, the grayscale image features of each frame are flattened into a two-dimensional matrix (shape , where H and W represent the height and width of the grayscale image features, and C represents the number of channels), the self-attention mechanism is applied to calculate the spatial weight of each position, and the calculated weight vector is multiplied with the two-dimensional matrix to obtain the weighted spatial graph feature.

[0084] At the same time, in the attention layer, each frame is represented as a frame vector, and the self-attention mechanism is applied to calculate the time weight of each frame vector. The calculated weight frame is multiplied by the frame vector to obtain the weighted time frame feature. The spatial graph features and time frame features processed by the spatial and temporal attention mechanisms are subjected to multi-scale feature fusion to finally obtain the output score vector z, z=[ , ],in For normal traffic category, Indicates the attack traffic category.

[0085] In an exemplary embodiment of the present disclosure, based on the calculated output score vector, the formula Calculate the probability of normal traffic category and get the normal probability , through the formula Calculate the probability of DDoS attack traffic categories to get the attack probability (in Output score for normal traffic, Output score for attack traffic).

[0086] After that, the traffic category is judged. If the attack probability Greater than normal probability , then the traffic data is judged to be attack traffic, and an attack traffic detection report is generated. No greater than normal probability , then the traffic data is judged to be normal traffic and a normal traffic detection report is generated.

[0087] In step S140, based on the traffic detection report, the detection of the encrypted HTTPS traffic DDoS attack is completed by responding to the alarm notification with preset measures.

[0088] In some optional embodiments of this example, if the detection result is a normal traffic detection report, the detection system continues to run the program according to the preset normal operation logic. If the detection result is an attack traffic detection report, the detection system determines that there is a DDoS attack in the encrypted traffic and generates an alarm notification including the attack start time, the attacked IP address, the attack type, the attack level, the attack source IP and other alarm information.

[0089] The automatic response system is activated while notifying the administrator via email, SMS, system pop-up windows, etc., and automatically executes a series of preset response measures, such as traffic cleaning for encrypted DDoS attacks and blocking the IP address of the attack source.

[0090] In this way, by collecting, parsing, and preprocessing the data, extracting data features, converting images, and using the STANet model for detection, it is ultimately possible to accurately identify, detect, and issue alarms for encrypted DDoS attacks.

[0091] It should be noted that, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0092] In addition, in this exemplary embodiment, a deep learning-based encrypted HTTPS traffic DDoS attack detection device is also provided. Figure 3 As shown, the encrypted HTTPS traffic DDoS attack detection device 300 based on deep learning may include: a data acquisition module 310, a data preprocessing module 320, a STANet deep learning detection algorithm module 330 and an alarm management module 340. Among them:

[0093] The data collection module 310 is used to collect traffic data from network devices and generate data to be processed;

[0094] The data preprocessing module 320 is used to generate encrypted traffic data by converting and extracting the data to be processed;

[0095] STANet deep learning detection algorithm module 330, used to generate an output score vector by processing the encrypted traffic data, and generate a traffic detection report by calculating the traffic data category probability based on the output score vector;

[0096] The alarm management module 340 is used to complete the detection of encrypted HTTPS traffic DDoS attacks by responding to alarm notifications with preset measures.

[0097] The encrypted HTTPS traffic DDoS attack detection device based on deep learning in the embodiment of the present disclosure corresponds to the embodiment of the encrypted HTTPS traffic DDoS attack detection method based on deep learning in the present disclosure, and the relevant contents can be referenced to each other, which will not be repeated here. The beneficial technical effects corresponding to the encrypted HTTPS traffic DDoS attack detection device based on deep learning in the embodiment of the present disclosure can be referred to the corresponding beneficial technical effects of the corresponding exemplary method part above, which will not be repeated here.

[0098] It should be noted that although several modules or units of the encrypted HTTPS traffic DDoS attack detection device 300 based on deep learning are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be concretized in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0099] Below, reference Figure 4 The electronic device according to the embodiment of the present disclosure is described. The electronic device may be any one or both of the first device and the second device, or a stand-alone device independent of them, and the stand-alone device may communicate with the first device and the second device to receive the collected input signals from them.

[0100] Figure 4 A block diagram of an electronic device according to an embodiment of the present disclosure is illustrated.

[0101] like Figure 4 As shown, the electronic device includes one or more processors and memory.

[0102] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0103] The memory may store one or more computer program products, and the memory may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program products may be stored on the computer-readable storage medium, and the processor may run the computer program product to implement the various embodiments of the present disclosure described above and / or other desired functions.

[0104] In one example, the electronic device may further include: an input device and an output device, and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0105] In addition, the input device may also include, for example, a keyboard, a mouse, and the like.

[0106] The output device can output various information to the outside, including determined distance information, direction information, etc. The output device can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0107] Of course, to simplify, Figure 4 Only some of the components related to the present disclosure in the electronic device are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device may further include any other appropriate components.

[0108] In addition to the above methods and devices, an embodiment of the present disclosure may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present disclosure described in the above part of this specification.

[0109] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the disclosed embodiments, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0110] In addition, an embodiment of the present disclosure may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enables the processor to execute the steps of the method according to various embodiments of the present disclosure described in the above part of this specification.

[0111] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable 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.

[0112] The basic principles of the present disclosure are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, and are not limitations. The above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

[0113] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0114] The block diagrams of the devices, apparatuses, equipment, and systems involved in this disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including," "comprising," "having," and the like are open words, referring to "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or," and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0115] The method and apparatus of the present disclosure may be implemented in many ways. For example, the method and apparatus of the present disclosure may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present disclosure are not limited to the order specifically described above, unless otherwise specifically stated. In addition, in some embodiments, the present disclosure may also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present disclosure. Therefore, the present disclosure also covers a recording medium storing a program for executing the method according to the present disclosure.

[0116] It should also be noted that in the apparatus, device and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0117] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0118] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for detecting DDoS attacks on encrypted HTTPS traffic based on deep learning, characterized in that: include: Based on XDP technology, traffic data is collected from network devices to generate data to be processed; Based on the attack characteristics of encrypted DDoS, the encrypted traffic data is generated by converting and extracting the data to be processed; Based on the STANet deep learning model, the encrypted traffic data is processed to generate an output score vector, and based on the output score vector, a traffic detection report is generated by calculating the probability of traffic data categories; Based on the traffic detection report, the detection of the encrypted HTTPS traffic DDoS attack is completed by responding to the alarm notification with preset measures; By converting and extracting the data to be processed, including: Based on the HTTPS protocol information, generating related data by first cleaning the data to be processed; Based on the protocol layering principle, layered data is generated by performing a second cleaning on the associated data; Based on the abnormal situation correction principle, the layered data is cleaned for a third time to generate cleaned data; Based on the principle of protocol standardization, the cleaned data is converted into a preset unified format to generate conversion data; Based on the preset characteristic parameters, the encrypted traffic data x is generated by extracting the converted data, where x=[x1, x2, ..., xN], where N is the number of the preset characteristic parameters.

2. The method according to claim 1, characterized in that By collecting traffic data from network devices, including: Based on the splitting mirror technology, a bypass network is generated by deploying network bypass in the network device driver; Based on the XDP technology, the flow data is collected in the bypass network to generate collected data; Based on the eBPF technology, the collected data is optimized and defined to generate the data to be processed.

3. The method according to claim 1, characterized in that The preset characteristic parameters include flow ID, source IP, source port, destination IP, destination port, protocol, TLS version, encryption suite, TLS extension, elliptic curve parameters, elliptic curve point format, JA3 fingerprint, flow duration, timestamp, total number of request packets, total number of reply packets, total number of request bytes, total number of reply bytes, maximum request packet length, minimum request packet length, average request message length, request message length variance, maximum reply packet length, minimum reply packet length, average reply message length, reply message length variance, traffic pps, traffic bps, data packet interval time, data packet interval time standard deviation, maximum data packet interval time, minimum data packet interval time, number of PSH flag packets, number of URG flag packets, number of SYN flag packets, number of ACK flag packets, number of RST flag packets, number of FIN flag packets, uplink and downlink traffic ratio, shortest flow active time, longest flow active time, average flow active time, standard deviation of flow active time, shortest flow idle time, longest flow idle time, average flow idle time, standard deviation of flow idle time, and data tag.

4. The method according to claim 1, characterized in that: By processing the encrypted traffic data, an output score vector is generated, including: Based on grayscale image conversion, a grayscale image is generated by converting the grayscale value of the encrypted traffic data; Based on a convolutional neural network model, a grayscale image feature is generated by extracting features from the grayscale image; Based on the spatial attention mechanism, the grayscale image feature is flattened into a two-dimensional matrix, and the position weight is calculated through the self-attention mechanism to generate a weight vector, and the weight vector is multiplied by the two-dimensional matrix to generate a spatial image feature; Based on the time attention mechanism, after the time frame of the grayscale image feature is represented as a frame vector, the time weight is calculated through the self-attention mechanism to generate a weight frame, and the weight frame is multiplied by the frame vector to generate a time frame feature; By performing multi-scale feature fusion on the spatial graph features and the time frame features, an output score vector z is generated, z=[ , ],in For normal traffic category, Indicates the attack traffic category.

5. The method according to claim 4, characterized in that Based on the output score vector, a traffic detection report is generated by calculating the probability of the traffic data category, including: Based on the output score vector, by the formula Calculate the probability of the normal traffic category to generate a normal probability ; Based on the output score vector, by the formula Calculate the probability of the attack traffic category to generate the attack probability ; in Output score for normal traffic, Output score for attack traffic; If the attack probability Greater than normal probability , then the traffic data is judged to be attack traffic, and an attack traffic detection report is generated; If the attack probability No greater than normal probability , then the traffic data is judged to be normal traffic and a normal traffic detection report is generated.

6. The method according to claim 5, characterized in that By responding to alarm notifications with preset measures, the detection of encrypted HTTPS traffic DDoS attacks is completed, including: Based on the normal flow detection report, the detection system continues to operate according to the preset operation logic; Based on the attack traffic detection report, the detection system establishes alarm information and generates an alarm notification; Based on the alarm notification, the automatic response system completes the detection of the encrypted HTTPS traffic DDoS attack by responding to the traffic data with preset measures.

7. A deep learning-based encrypted HTTPS traffic DDoS attack detection device, the device adopts the method described in any one of claims 1-6, characterized in that: The device comprises: The data collection module is used to collect traffic data from network devices and generate data to be processed; A data preprocessing module, used to generate encrypted traffic data by converting and extracting the data to be processed; A STANet deep learning detection algorithm module, used to generate an output score vector by processing the encrypted traffic data, and generate a traffic detection report by calculating the traffic data category probability based on the output score vector; The alarm management module is used to complete the detection of encrypted HTTPS traffic DDoS attacks by responding to alarm notifications with preset measures.

8. An electronic device, characterized in that: include: a memory for storing a computer program product; A processor is used to execute the computer program product stored in the memory, and when the computer program product is executed, it implements the method described in any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method described in any one of claims 1 to 6 is implemented.