Directional network attack identification method and device based on generative adversarial network

Through the targeted network attack identification method based on the generative adversarial network, the adversarial training model is used to identify and defend against network attacks in real time, and the problem of inability to efficiently identify and defend against network attacks in the existing technology is solved, and efficient security protection of power grid data is achieved.

CN119030731APending Publication Date: 2024-11-26ELECTRIC POWER SCI RES INST OF STATE GRID XINJIANG ELECTRIC POWER CO LTD +1
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410880758.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-02
Publication Date
2024-11-26

Smart Images

  • Figure CN119030731A_ABST
    Figure CN119030731A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of power grid data security, and discloses a generative adversarial network-based directional network attack identification method and device, and the method comprises the steps: obtaining historical intrusion data of a power grid and corresponding historical response data; constructing a confrontation training model; performing data enrichment on the historical intrusion data to obtain intrusion training data, and inputting the intrusion training data into an adversarial training model for adversarial training to obtain training response data; and capturing the generated intrusion data, importing the intrusion data into the adversarial training model, and carrying out network attack identification and defense through output response data. The system corresponds to the method. With the adoption of the method and the device, when the network intrusion behavior is generated, identification and defense of the intrusion data are carried out by calling the historical response data or the training response data, and the corresponding identification and defense strategy can be generated in real time through the adversarial training model based on the specific condition of the intrusion data, so that comprehensive and efficient network security protection is realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of power grid data security technology, and specifically to a method and device for identifying directed network attacks based on a generative adversarial network. Background Art

[0002] Power grid data is an important resource data, and its security is highly valued by all major power grid companies. In order to improve the security of power grid data, all companies will screen and verify power grid data in their daily enterprise construction. Existing network attack defense methods usually use honeypots or real-time defense methods. Specifically, honeypots lure network attack intrusions by arranging induction points, and obtain corresponding defense identification strategies based on the analysis of network attack methods. Real-time defense is to identify and defend network intrusions that appear at any time by arranging high-intensity firewalls. These two methods constitute typical network security methods. However, the use of honeypots to capture and analyze attacks and then defend against them is highly targeted and cannot ensure efficient network attack identification and defense at the first time of intrusion behavior. There are great network security risks. The use of firewalls for real-time defense requires powerful network security technology to arrange highly secure firewalls to achieve identification and defense of network attacks. Limited by the capabilities of network security technology, there is also a situation where network security protection cannot be carried out comprehensively and efficiently.

[0003] Therefore, there is an urgent need for a network security technology with high security and comprehensive protection. Summary of the invention

[0004] The purpose of this application is to provide a method and device for identifying targeted network attacks based on a generative adversarial network to solve the technical problems raised in the above background.

[0005] To achieve the above objectives, this application discloses the following technical solutions:

[0006] In a first aspect, the present application provides a method for identifying a targeted network attack based on a generative adversarial network, the method comprising the following steps:

[0007] S1: Obtain historical intrusion data and corresponding historical response data of the power grid;

[0008] S2: Building an adversarial training model based on the historical intrusion data and the corresponding historical response data;

[0009] S3: enriching the historical intrusion data, and inputting the intrusion training data obtained after the data enrichment into the adversarial training model for adversarial training to obtain a number of training response data;

[0010] S4: Capture the intrusion data generated in the network node. When the capture is successful, import the intrusion data into the adversarial training model, and output the corresponding response data through the adversarial training model to identify and defend against network attacks; the response data includes historical response data and training response data.

[0011] Preferably, the historical intrusion data includes attack frequency, behavior characteristics and intrusion objects; the historical response data includes response mode and response object.

[0012] Preferably, the S4 further comprises: performing data analysis on the captured intrusion data to obtain the attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data;

[0013] The network attack identification and defense are performed by outputting corresponding response data through the adversarial training model, specifically including:

[0014] The adversarial training model matches the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data with the historical intrusion data and intrusion training data, and uses the historical response data or training response data corresponding to the successfully matched historical intrusion data or intrusion training data as an identification and defense strategy to identify and defend the intrusion data.

[0015] Preferably, when data matching is unsuccessful, the adversarial training model calculates the historical intrusion data with the highest similarity to the attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data and the historical response data or training response data corresponding to the intrusion training data as emergency identification and defense strategies for emergency adversarial identification and defense; and performs real-time adversarial training based on the intrusion data, and arranges the training response data generated by the adversarial training process based on the intrusion data at the corresponding network nodes in real time for identification and defense.

[0016] Preferably, the real-time adversarial training based on intrusion data specifically includes:

[0017] The intrusion data is anatomically analyzed to obtain its corresponding attack frequency characteristics, behavior characteristics and intrusion object characteristics, and the historical intrusion data and intrusion training data are matched using the attack frequency characteristics, behavior characteristics and intrusion object characteristics as matching methods respectively. The historical response data or training response data corresponding to all the matched historical intrusion data and intrusion training data are strategically combined, and the obtained combined identification defense strategy is used as a training response data generated by the adversarial training process based on the intrusion data.

[0018] Preferably, the data enrichment includes at least one of data splitting, data reorganization, and data derivation.

[0019] In a second aspect, the present application provides a directed network attack identification system based on a generative adversarial network, comprising:

[0020] A historical data acquisition module is configured to: acquire historical intrusion data of the power grid and corresponding historical response data;

[0021] A model building module is configured to: build an adversarial training model based on the historical intrusion data and the corresponding historical response data;

[0022] The adversarial training module is configured to: enrich the historical intrusion data, and input the intrusion training data obtained after the data enrichment into the adversarial training model for adversarial training, and obtain a number of training response data recognition;

[0023] The intrusion confrontation module is configured to: capture the intrusion data generated in the network node, and when the capture is successful, import the intrusion data into the confrontation training model, and output the corresponding response data through the confrontation training model to identify and defend against network attacks; the response data includes historical response data and training response data.

[0024] Preferably, the historical intrusion data includes attack frequency, behavior characteristics and intrusion objects; the historical response data includes response mode and response object;

[0025] The intrusion resistance module is further configured to: perform data analysis on the captured intrusion data to obtain the attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data;

[0026] The network attack identification and defense are performed by outputting corresponding response data through the adversarial training model, specifically including:

[0027] The adversarial training model matches the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data with the historical intrusion data and intrusion training data, and uses the historical response data or training response data corresponding to the successfully matched historical intrusion data or intrusion training data as an identification and defense strategy to identify and defend the intrusion data.

[0028] Preferably, when data matching is unsuccessful, the adversarial training model calculates the historical intrusion data with the highest similarity to the attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data and the historical response data or training response data corresponding to the intrusion training data as emergency identification and defense strategies for emergency adversarial identification and defense; and performs real-time adversarial training based on the intrusion data, and arranges the training response data generated by the adversarial training process based on the intrusion data at the corresponding network nodes in real time for identification and defense.

[0029] Preferably, the real-time adversarial training based on intrusion data specifically includes:

[0030] The intrusion data is anatomically analyzed to obtain its corresponding attack frequency characteristics, behavior characteristics and intrusion object characteristics, and the historical intrusion data and intrusion training data are matched using the attack frequency characteristics, behavior characteristics and intrusion object characteristics as matching methods respectively. The historical response data or training response data corresponding to all the matched historical intrusion data and intrusion training data are strategically combined, and the obtained combined identification defense strategy is used as a training response data generated by the adversarial training process based on the intrusion data.

[0031] Beneficial effects: The method and device for identifying targeted network attacks based on a generative adversarial network of the present application are based on the construction of an adversarial training model to enrich the reserve of response data, wherein the response data is composed of historical response data and training response data. When a network intrusion occurs, the intrusion data is identified and defended by retrieving historical response data or training response data, and based on the specific circumstances of the intrusion data, the adversarial training model can be used to generate corresponding identification and defense strategies in real time, thereby achieving comprehensive and efficient network security protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0033] Figure 1 A flowchart of a method for identifying targeted network attacks based on a generative adversarial network provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the present application are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0035] In this article, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprising..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0036] In a first aspect, this embodiment discloses a method such as Figure 1 The method for identifying directed network attacks based on a generative adversarial network shown in the figure comprises the following steps:

[0037] S1: Obtain historical intrusion data and corresponding historical response data of the power grid;

[0038] S2: Building an adversarial training model based on the historical intrusion data and the corresponding historical response data;

[0039] S3: enriching the historical intrusion data, and inputting the intrusion training data obtained after the data enrichment into the adversarial training model for adversarial training to obtain a number of training response data;

[0040] S4: Capture the intrusion data generated in the network node. When the capture is successful, import the intrusion data into the adversarial training model, and output the corresponding response data through the adversarial training model to identify and defend against network attacks; the response data includes historical response data and training response data.

[0041] In this embodiment, the data enrichment includes at least one of data splitting, data reorganization, and data derivation. The historical intrusion data includes attack frequency, behavior characteristics, and intrusion objects; the historical response data includes response methods and response objects. In this regard, S4 also includes: performing data analysis on the captured intrusion data to obtain the attack frequency, behavior characteristics, and intrusion objects corresponding to the intrusion data;

[0042] The network attack identification and defense are performed by outputting corresponding response data through the adversarial training model, specifically including:

[0043] The adversarial training model matches the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data with the historical intrusion data and intrusion training data, and uses the historical response data or training response data corresponding to the successfully matched historical intrusion data or intrusion training data as an identification and defense strategy to identify and defend the intrusion data.

[0044] Furthermore, when data matching is unsuccessful, the adversarial training model calculates the historical intrusion data with the highest similarity to the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data and the historical response data or training response data corresponding to the intrusion training data as emergency identification and defense strategies for emergency adversarial identification and defense; and performs real-time adversarial training based on the intrusion data, and arranges the training response data generated by the adversarial training process based on the intrusion data at the corresponding network nodes in real time for identification and defense.

[0045] As a preferred implementation of this embodiment, the real-time adversarial training based on intrusion data specifically includes:

[0046] The intrusion data is anatomically analyzed to obtain its corresponding attack frequency characteristics, behavior characteristics and intrusion object characteristics, and the historical intrusion data and intrusion training data are matched using the attack frequency characteristics, behavior characteristics and intrusion object characteristics as matching methods respectively. The historical response data or training response data corresponding to all the matched historical intrusion data and intrusion training data are strategically combined, and the obtained combined identification defense strategy is used as a training response data generated by the adversarial training process based on the intrusion data.

[0047] Based on the above, the targeted network attack identification method based on generative adversarial networks of this embodiment is based on the construction of an adversarial training model to enrich the reserve of response data, wherein the response data is composed of historical response data and training response data. When a network intrusion occurs, the intrusion data is identified and defended by retrieving historical response data or training response data, and based on the specific circumstances of the intrusion data, the corresponding identification and defense strategy can be generated in real time through the adversarial training model, thereby achieving comprehensive and efficient network security protection.

[0048] In a second aspect, this embodiment discloses a directed network attack identification system based on a generative adversarial network, including:

[0049] A historical data acquisition module is configured to: acquire historical intrusion data of the power grid and corresponding historical response data;

[0050] A model building module is configured to: build an adversarial training model based on the historical intrusion data and the corresponding historical response data;

[0051] The adversarial training module is configured to: enrich the historical intrusion data, and input the intrusion training data obtained after the data enrichment into the adversarial training model for adversarial training, and obtain a number of training response data recognition;

[0052] The intrusion confrontation module is configured to: capture the intrusion data generated in the network node, and when the capture is successful, import the intrusion data into the confrontation training model, and output the corresponding response data through the confrontation training model to identify and defend against network attacks; the response data includes historical response data and training response data.

[0053] The historical intrusion data includes attack frequency, behavior characteristics and intrusion objects; the historical response data includes response methods and response objects;

[0054] The intrusion resistance module is further configured to: perform data analysis on the captured intrusion data to obtain the attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data;

[0055] The network attack identification and defense are performed by outputting corresponding response data through the adversarial training model, specifically including:

[0056] The adversarial training model matches the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data with the historical intrusion data and intrusion training data, and uses the historical response data or training response data corresponding to the successfully matched historical intrusion data or intrusion training data as an identification and defense strategy to identify and defend the intrusion data.

[0057] Furthermore, when data matching is unsuccessful, the adversarial training model calculates the historical intrusion data with the highest similarity to the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data and the historical response data or training response data corresponding to the intrusion training data as emergency identification and defense strategies for emergency adversarial identification and defense; and performs real-time adversarial training based on the intrusion data, and arranges the training response data generated by the adversarial training process based on the intrusion data at the corresponding network nodes in real time for identification and defense.

[0058] It is feasible that the real-time adversarial training based on the intrusion data specifically includes:

[0059] The intrusion data is anatomically analyzed to obtain its corresponding attack frequency characteristics, behavior characteristics and intrusion object characteristics, and the historical intrusion data and intrusion training data are matched using the attack frequency characteristics, behavior characteristics and intrusion object characteristics as matching methods respectively. The historical response data or training response data corresponding to all the matched historical intrusion data and intrusion training data are strategically combined, and the obtained combined identification defense strategy is used as a training response data generated by the adversarial training process based on the intrusion data.

[0060] It should be noted that the directed network attack identification system based on generative adversarial networks of this embodiment corresponds to the aforementioned directed network attack identification method based on generative adversarial networks. Therefore, the technical content and corresponding technical effects that are not disclosed in the directed network attack identification system based on generative adversarial networks can be referred to the description in the aforementioned directed network attack identification method based on generative adversarial networks, and this text will not elaborate on them here.

[0061] In the embodiments provided in the present application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: application specific integrated circuit (ASIC), digital signal processor (DSP), digital signal processing device (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), processor, controller, microcontroller, microprocessor, other electronic units designed to implement the functions described herein or their combination. For software implementation, part or all of the flow of the embodiment can be completed by instructing the relevant hardware through a computer program. When implemented, the above program can be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein the communication medium includes any medium that is convenient for transmitting a computer program from one place to another. The storage medium can be any available medium that a computer can access. The computer-readable storage medium can include but is not limited to RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer.

[0062] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for identifying targeted network attacks based on generative adversarial networks, characterized in that: The method comprises the following steps: S1: Obtain historical intrusion data and corresponding historical response data of the power grid; S2: Building an adversarial training model based on the historical intrusion data and the corresponding historical response data; S3: enriching the historical intrusion data, and inputting the intrusion training data obtained after the data enrichment into the adversarial training model for adversarial training to obtain a number of training response data; S4: Capture the intrusion data generated in the network node. When the capture is successful, import the intrusion data into the adversarial training model, and output the corresponding response data through the adversarial training model to identify and defend against network attacks; the response data includes historical response data and training response data.

2. The method for identifying targeted network attacks based on a generative adversarial network according to claim 1, characterized in that: The historical intrusion data includes attack frequency, behavior characteristics and intrusion objects; the historical response data includes response methods and response objects.

3. The method for identifying targeted network attacks based on generative adversarial networks according to claim 2 is characterized in that: The S4 also includes: performing data analysis on the captured intrusion data to obtain the attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data; The network attack identification and defense are performed by outputting corresponding response data through the adversarial training model, specifically including: The adversarial training model matches the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data with the historical intrusion data and intrusion training data, and uses the historical response data or training response data corresponding to the successfully matched historical intrusion data or intrusion training data as an identification and defense strategy to identify and defend the intrusion data.

4. The method for identifying targeted network attacks based on a generative adversarial network according to claim 3 is characterized in that: When the data matching is unsuccessful, the adversarial training model calculates the historical intrusion data with the highest similarity in attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data and the historical response data or training response data corresponding to the intrusion training data as an emergency identification and defense strategy for emergency adversarial identification and defense based on the result of the data matching; Real-time adversarial training is performed based on the intrusion data, and the training response data generated by the adversarial training process based on the intrusion data is arranged at the corresponding network nodes in real time for identification and defense.

5. The method for identifying targeted network attacks based on generative adversarial networks according to claim 4 is characterized in that: The real-time adversarial training based on intrusion data specifically includes: The intrusion data is anatomically analyzed to obtain its corresponding attack frequency characteristics, behavior characteristics and intrusion object characteristics, and the historical intrusion data and intrusion training data are matched using the attack frequency characteristics, behavior characteristics and intrusion object characteristics as matching methods respectively. The historical response data or training response data corresponding to all the matched historical intrusion data and intrusion training data are strategically combined, and the obtained combined identification defense strategy is used as a training response data generated by the adversarial training process based on the intrusion data.

6. The method for identifying targeted network attacks based on a generative adversarial network according to claim 1, characterized in that: The data enrichment includes at least one of data splitting, data reorganization, and data derivation.

7. A directed network attack identification system based on generative adversarial networks, characterized in that: include: A historical data acquisition module is configured to: acquire historical intrusion data of the power grid and corresponding historical response data; A model building module is configured to: build an adversarial training model based on the historical intrusion data and the corresponding historical response data; The adversarial training module is configured to: enrich the historical intrusion data, and input the intrusion training data obtained after the data enrichment into the adversarial training model for adversarial training, and obtain a number of training response data recognition; The intrusion confrontation module is configured to: capture the intrusion data generated in the network node, and when the capture is successful, import the intrusion data into the confrontation training model, and output the corresponding response data through the confrontation training model to identify and defend against network attacks; the response data includes historical response data and training response data.

8. The directed network attack identification system based on generative adversarial network according to claim 7 is characterized in that: The historical intrusion data includes attack frequency, behavior characteristics and intrusion objects; the historical response data includes response methods and response objects; The intrusion resistance module is further configured to: perform data analysis on the captured intrusion data to obtain the attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data; The network attack identification and defense are performed by outputting corresponding response data through the adversarial training model, specifically including: The adversarial training model matches the attack frequency, behavior characteristics and intrusion objects corresponding to the intrusion data with the historical intrusion data and intrusion training data, and uses the historical response data or training response data corresponding to the successfully matched historical intrusion data or intrusion training data as an identification and defense strategy to identify and defend the intrusion data.

9. The directed network attack identification system based on generative adversarial network according to claim 8 is characterized in that: When the data matching is unsuccessful, the adversarial training model calculates the historical intrusion data with the highest similarity in attack frequency, behavior characteristics and intrusion object corresponding to the intrusion data and the historical response data or training response data corresponding to the intrusion training data as an emergency identification and defense strategy for emergency adversarial identification and defense based on the result of the data matching; Real-time adversarial training is performed based on the intrusion data, and the training response data generated by the adversarial training process based on the intrusion data is arranged at the corresponding network nodes in real time for identification and defense.

10. The method for identifying targeted network attacks based on generative adversarial networks according to claim 9, characterized in that: The real-time adversarial training based on intrusion data specifically includes: The intrusion data is anatomically analyzed to obtain its corresponding attack frequency characteristics, behavior characteristics and intrusion object characteristics, and the historical intrusion data and intrusion training data are matched using the attack frequency characteristics, behavior characteristics and intrusion object characteristics as matching methods respectively. The historical response data or training response data corresponding to all the matched historical intrusion data and intrusion training data are strategically combined, and the obtained combined identification defense strategy is used as a training response data generated by the adversarial training process based on the intrusion data.

Citation Information

Cited By

  • Pluggable multi-mode large model safety protection method

    CN121211443A

  • Generative adversarial network-based nuclear energy data secrecy security protection method and system

    CN121217415A