Bilinear attention mechanism IoT attack and defense method and system
Through the bilinear attention mechanism, the Internet of Things network traffic is analyzed and feature extraction is performed, and the network attack type judgment is determined by combining the Softmax function, which solves the problems of incomplete monitoring and insufficient offensive and defense models when facing network attacks, and improves the Internet of Things security and attack recognition efficiency.
Patent Information
- Application Number
- CN202410969977.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-07-19
AI Technical Summary
In the prior art, when facing cyber attacks, the real-time monitoring of the Internet of Things is not comprehensive and accurate enough, and it is impossible to effectively establish an offensive and defensive model for different types of cyber attacks, resulting in low security of the Internet of Things.
The bilinear attention mechanism is used to perform PCA analysis and feature extraction of network traffic logs, integrate key tuple information of the network layer and the transmission layer, distinguish network attack types through Softmax function, and build an IoT attack and defense model based on the attack type, select Optimal defense strategy.
It improves the recognition efficiency of network attacks and the security of IoT, and establishes a complete defense system that can quickly extract attack characteristics and select the optimal defense strategy.
Smart Images

Figure CN118523958B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of Internet of Things attack and defense technology, and in particular to an Internet of Things attack and defense method and system with a dual linear attention mechanism. Background Art
[0002] The bilinear attention mechanism is a new type of attention mechanism that combines the ideas of the bilinear model and the attention mechanism. It can effectively process multi-channel input data, extract key information, and improve the performance of the model. In the field of IoT attack and defense, the bilinear attention mechanism can be applied to data processing, feature extraction, attack detection, and other aspects, providing strong technical support for IoT security.
[0003] In the existing technology, when the Internet of Things faces network attacks, the real-time monitoring of network attacks is often not comprehensive and accurate enough. Furthermore, it is impossible to establish a good attack and defense model for different types of network attacks, resulting in low security of the Internet of Things.
[0004] Therefore, there is an urgent need for a bilinear attention mechanism IoT attack and defense method and system to solve the above problems. Summary of the invention
[0005] The purpose of the present invention is to provide a bilinear attention mechanism Internet of Things attack and defense method and system: to solve the technical problems that the existing solutions cannot perform real-time monitoring of network attacks, which is often not comprehensive and accurate, and cannot establish a good attack and defense model for different types of network attacks, resulting in low security of the Internet of Things.
[0006] The purpose of the present invention can be achieved through the following technical solutions:
[0007] On the one hand, the bilinear attention mechanism IoT attack and defense method includes:
[0008] Obtaining a network traffic log within a management period, performing PCA analysis on the network traffic log, and obtaining a first feature associated with a network attack, wherein the management period is a period during which the network is subjected to a network attack;
[0009] Acquire data packets within the management period, and acquire key tuple information of the network layer and the transport layer in the data packets according to the protocol types and fields of different data packets, wherein the key tuple information includes a source IP address, a destination IP address, a source port number, a destination port number, and a protocol number, and perform feature extraction on the key tuple information to obtain a second feature associated with the network attack;
[0010] The first feature and the second feature are fused into attack features through the bilinear attention mechanism, and then input into the fully connected layer after power normalization, and the network attack type is identified through the Softmax function;
[0011] Build an IoT attack and defense model based on the type of network attack and select the optimal defense strategy.
[0012] Further, performing PCA analysis on the network traffic log to obtain the first feature associated with the network attack includes the following steps:
[0013] (1) comparing the network traffic logs based on key performance indicators of the network traffic to determine abnormal network traffic information, wherein the key performance indicators of the network traffic include a packet size and a packet rate, determining whether the packet size exceeds a preset packet size threshold, if so, the packet is abnormal network traffic information, and determining whether the packet rate exceeds a preset packet rate threshold, if so, the packet is abnormal network traffic information;
[0014] (2) Based on the abnormal network traffic information, a matrix X is formed and the covariance matrix S of the samples of the matrix X is calculated;
[0015] (3) Calculate the eigenvectors of the covariance matrix S , ,…, and eigenvalues t=1,2,…,n;
[0016] (4) Project the data into the feature vector space, using the formula:
[0017] ;
[0018] Among them, the BV value is the value of the corresponding dimension in the original sample;
[0019] (5) Based on the maximum The corresponding feature vector is recorded as the first feature associated with the network attack.
[0020] Furthermore, obtaining key tuple information of the network layer and the transport layer in the data packet according to the protocol type and field of different data packets includes:
[0021] Link layer protocol analysis, determining the message type based on the protocol number field in the data packet header;
[0022] Network layer protocol analysis to determine whether the message is an IP message. If not, the corresponding network layer message is further analyzed. If it is an IP message, the message is judged whether IP fragments appear according to its fragmentation mark, and the next step is to decide whether to perform fragment reassembly or obtain specific information of the IP packet header.
[0023] Transport layer protocol analysis to obtain the port number or ACK, SYN, FIN, and RST identification information during TCP connection establishment;
[0024] Application layer protocol analysis to determine whether it is an HTTP message. If so, obtain the URL, cookie information, and GET / POST / PUT request method in the payload of the HTTP message and parse the payload.
[0025] Furthermore, feature extraction is performed on the key tuple information to obtain the second feature associated with the network attack, including:
[0026] Count the number of statistical packets, average packet length, total data volume of the flow and duration of the flow based on key tuple information;
[0027] The importance of each feature is evaluated by the random forest algorithm for the number of data packets, average data packet length, total data volume of the flow, and duration of the flow. Features are extracted in order according to their importance, and the obtained features are recorded as the second features associated with network attacks.
[0028] Furthermore, the first feature and the second feature are fused into attack features through the bilinear attention mechanism, and then input into the fully connected layer after power normalization, and the network attack type is identified through the Softmax function, including:
[0029] Step 1: The first feature , the second feature The fusion is performed according to the following formula:
[0030] ;
[0031] in, Indicates the number of training sample packets, , Represents the full connection parameters, , , are the dimensions of the first and second characteristics, respectively;
[0032] Step 2: Use the attention mechanism for pooling operation, and the attention size is set to 16;
[0033] Step 3: The attention mechanism is:
[0034] ;
[0035] ;
[0036] ;
[0037] in, , for The weight matrix of represents the bias, express Cyber attack risk score, Expressed as The weight of the attack feature, is the attack feature after fusion;
[0038] Step 4: Use the Softmax function to identify the type of network attack: Convert it into a probability distribution so that the sum of the probabilities of all categories is 1, and determine the type of network attack by probability.
[0039] Furthermore, the IoT attack and defense model is constructed based on the network attack type, and the optimal defense strategy is selected, including:
[0040] Establish Five tuples, and define:
[0041] Gather for the players, For the defenders, for the attacker;
[0042] The defender type refers to the real type of the attack target corresponding to the network attack type. Indicates real service, Represents a honeypot;
[0043] is the action space of the defender and attacker, where The defender's action space, Represents the transmission of real service signals, Indicates the transmission of honeypot signal; is the attacker’s action space, Indicates that the attacker chooses to attack, Indicates that the attacker chooses to give up the attack; the defender provides four services: 1) When the defender type is When the defender maintains its own service to send a signal , providing deterrence services to send signals ; 2) When the defender type is When the defender provides a decoy service to send a signal , maintain the honeypot service to send signals ;
[0044] is the attacker’s game belief set, represents the attacker's prior belief about the defender type, satisfying , represents the attacker's posterior belief about the defender type, It is calculated by the attacker using Bayes’ rule based on the received signal and prior beliefs;
[0045] is the set of profit functions, represents the defender's payoff function, represents the attacker’s payoff function;
[0046] Substitute the network attack type into , based on the defender benefit function, determine the candidate defense strategy corresponding to the maximum defender benefit function value, wherein the candidate defense strategy corresponding to the maximum benefit function value is used as the defense strategy, and the candidate defense strategy is a preset defense strategy.
[0047] On the one hand, the bilinear attention mechanism IoT attack and defense system includes a feature extraction module, a fusion discrimination module, and an attack and defense module;
[0048] The feature extraction module is used to obtain the network traffic log within the management period, perform PCA analysis on the network traffic log, and obtain the first feature associated with the network attack, wherein the management period is the period when the network is subjected to the network attack; obtain the data packet within the management period, obtain the key tuple information of the network layer and the transport layer in the data packet according to the protocol type and field of different data packets, wherein the key tuple information includes the source IP address, the destination IP address, the source port number, the destination port number and the protocol number, and perform feature extraction on the key tuple information to obtain the second feature associated with the network attack;
[0049] The fusion discrimination module is used to fuse the first feature and the second feature into attack features through the bilinear attention mechanism, and then input them into the fully connected layer after power normalization, and discriminate the network attack type through the Softmax function;
[0050] The attack and defense module is used to build an IoT attack and defense model based on the type of network attack and select the optimal defense strategy.
[0051] Compared with the existing solutions, the present invention achieves the following beneficial effects:
[0052] The present invention obtains the network traffic log within the management period, performs PCA analysis on the network traffic log, and obtains the first feature associated with the network attack; obtains the data packet within the management period, obtains the key tuple information of the network layer and the transport layer in the data packet according to the protocol type and field of different data packets, and extracts the key tuple information to obtain the second feature associated with the network attack; the first feature and the second feature are fused into the attack feature through the bilinear attention mechanism, and then input into the fully connected layer after power normalization, and the network attack type is distinguished through the Softmax function. The present invention can quickly extract the attack features of the network attack, and distinguish the network attack type through the Softmax function, thereby improving the recognition efficiency of the network attack.
[0053] The present invention constructs an IoT attack and defense model based on the type of network attacks, selects the optimal defense strategy, establishes a complete defense system, and improves the security of the IoT. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] 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 will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.
[0055] Figure 1 It is a workflow diagram of a bilinear attention mechanism Internet of Things attack and defense method according to an embodiment of the present invention.
[0056] Figure 2 It is a system block diagram of a dual-linear attention mechanism IoT attack and defense method according to an embodiment of the present invention; DETAILED DESCRIPTION
[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0058] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.
[0059] This embodiment provides a bilinear attention mechanism IoT attack and defense method. Figure 1 This is a workflow diagram of a bilinear attention mechanism IoT attack and defense method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:
[0060] Step S101: obtaining a network traffic log within a management period, performing PCA analysis on the network traffic log, and obtaining a first feature associated with a network attack, wherein the management period is a period during which the network is subjected to a network attack;
[0061] Step S102: obtaining data packets within the management period, obtaining key tuple information of the network layer and the transport layer in the data packets according to the protocol types and fields of different data packets, wherein the key tuple information includes the source IP address, the destination IP address, the source port number, the destination port number and the protocol number, and performing feature extraction on the key tuple information to obtain a second feature associated with the network attack;
[0062] Step S103: The first feature and the second feature are fused into attack features through a bilinear attention mechanism, and then input into a fully connected layer after power normalization, and the network attack type is identified through a Softmax function;
[0063] Step S104: construct an IoT attack and defense model based on the network attack type and select the optimal defense strategy.
[0064] In summary, the present invention obtains the network traffic log within the management period, performs PCA analysis on the network traffic log, and obtains the first feature associated with the network attack; obtains the data packet within the management period, obtains the key tuple information of the network layer and the transport layer in the data packet according to the protocol type and field of different data packets, and extracts the key tuple information to obtain the second feature associated with the network attack; the first feature and the second feature are fused into the attack feature through the bilinear attention mechanism, and then input into the fully connected layer after power normalization, and the network attack type is distinguished by the Softmax function. The present invention can quickly extract the attack features of the network attack, and distinguish the network attack type by the Softmax function, thereby improving the recognition efficiency of the network attack. The IoT attack and defense model is constructed based on the network attack type, the optimal defense strategy is selected, a complete defense system is established, and the security of the IoT is improved.
[0065] In some embodiments, performing PCA analysis on the network traffic log to obtain a first feature associated with the network attack includes the following steps:
[0066] (1) comparing the network traffic logs based on key performance indicators of the network traffic to determine abnormal network traffic information, wherein the key performance indicators of the network traffic include a packet size and a packet rate, determining whether the packet size exceeds a preset packet size threshold, if so, the packet is abnormal network traffic information, and determining whether the packet rate exceeds a preset packet rate threshold, if so, the packet is abnormal network traffic information;
[0067] (2) Based on the abnormal network traffic information, a matrix X is formed and the covariance matrix S of the samples of the matrix X is calculated;
[0068] (3) Calculate the eigenvectors of the covariance matrix S , ,…, and eigenvalues t=1,2,…,n;
[0069] (4) Project the data into the feature vector space, using the formula:
[0070] ;
[0071] Among them, the BV value is the value of the corresponding dimension in the original sample;
[0072] (5) Based on the maximum The corresponding feature vector is recorded as the first feature associated with the network attack.
[0073] In some embodiments, obtaining key tuple information of the network layer and the transport layer in the data packet according to the protocol type and field of different data packets includes:
[0074] Link layer protocol analysis, determining the message type based on the protocol number field in the data packet header;
[0075] Network layer protocol analysis: determine whether the message is an IP message. If not, analyze the corresponding network layer message again. If it is an IP message, determine whether the message has IP fragments based on its fragmentation identifier, that is, decide whether to perform fragment reassembly or obtain specific information of the IP packet header in the next step;
[0076] Transport layer protocol analysis to obtain the port number or ACK, SYN, FIN, and RST identification information during TCP connection establishment;
[0077] Application layer protocol analysis to determine whether it is an HTTP message. If so, obtain the URL, cookie information, and GET / POST / PUT request method in the payload of the HTTP message and parse the payload.
[0078] In some embodiments, the first feature and the second feature are fused into an attack feature through a bilinear attention mechanism, and then input into a fully connected layer after power normalization, and the network attack type is discriminated through a Softmax function, including:
[0079] Step 1: The first feature , the second feature The fusion is performed according to the following formula:
[0080] ;
[0081] in, Indicates the number of training sample packets, , Represents the full connection parameters, , , are the dimensions of the first and second characteristics, respectively;
[0082] Step 2: Use the attention mechanism for pooling operation, and the attention size is set to 16;
[0083] Step 3: The attention mechanism is:
[0084] ;
[0085] ;
[0086] ;
[0087] in, , for The weight matrix of represents the bias, express Cyber attack risk score, Expressed as The weight of the attack feature, is the attack feature after fusion;
[0088] Step 4: Use the Softmax function to identify the type of network attack: Convert it into a probability distribution so that the sum of the probabilities of all categories is 1, and determine the type of network attack by probability.
[0089] In some embodiments, building an IoT attack and defense model based on the network attack type and selecting the optimal defense strategy include:
[0090] Establish Five tuples, and define:
[0091] Gather for the players, For the defenders, for the attacker;
[0092] The defender type refers to the real type of the attack target corresponding to the network attack type. Indicates real service, Represents a honeypot;
[0093] is the action space of the defender and attacker, where The defender's action space, Represents the transmission of real service signals, Indicates the transmission of honeypot signal; is the attacker’s action space, Indicates that the attacker chooses to attack, Indicates that the attacker chooses to give up the attack; the defender provides four services: 1) When the defender type is When the defender maintains its own service to send a signal , providing deterrence services to send signals ; 2) When the defender type is When the defender provides a decoy service to send a signal , maintain the honeypot service to send signals ;
[0094] is the attacker’s game belief set, represents the attacker's prior belief about the defender type, satisfying , represents the attacker's posterior belief about the defender type, It is calculated by the attacker using Bayes’ rule based on the received signal and prior beliefs;
[0095] is the set of profit functions, represents the defender's payoff function, represents the attacker’s payoff function;
[0096] Substitute the network attack type into , based on the defender benefit function, determine the candidate defense strategy corresponding to the maximum defender benefit function value, wherein the candidate defense strategy corresponding to the maximum benefit function value is used as the defense strategy, and the candidate defense strategy is a preset defense strategy.
[0097] This embodiment also provides a bilinear attention mechanism IoT attack and defense system. Figure 2 is a system block diagram of a dual linear attention mechanism IoT attack and defense method according to an embodiment of the present invention. Figure 2 As shown, the system includes a feature extraction module, a fusion discrimination module and an attack and defense module;
[0098] The feature extraction module is used to obtain the network traffic log within the management period, perform PCA analysis on the network traffic log, and obtain the first feature associated with the network attack, wherein the management period is the period when the network is subjected to the network attack; obtain the data packet within the management period, obtain the key tuple information of the network layer and the transport layer in the data packet according to the protocol type and field of different data packets, wherein the key tuple information includes the source IP address, the destination IP address, the source port number, the destination port number and the protocol number, and perform feature extraction on the key tuple information to obtain the second feature associated with the network attack;
[0099] The fusion discrimination module is used to fuse the first feature and the second feature into attack features through the bilinear attention mechanism, and then input them into the fully connected layer after power normalization, and discriminate the network attack type through the Softmax function;
[0100] The attack and defense module is used to build an IoT attack and defense model based on the type of network attack and select the optimal defense strategy.
[0101] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.
[0102] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0103] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0104] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only for some logical functions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0105] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0106] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. The bilinear attention mechanism IoT attack and defense method is characterized by: Methods include: Obtaining a network traffic log within a management period, performing PCA analysis on the network traffic log, and obtaining a first feature associated with a network attack, wherein the management period is a period during which the network is subjected to a network attack; The PCA analysis of the network traffic log to obtain the first feature associated with the network attack includes the following steps: (1) comparing the network traffic logs based on key performance indicators of the network traffic to determine abnormal network traffic information, wherein the key performance indicators of the network traffic include a packet size and a packet rate, determining whether the packet size exceeds a preset packet size threshold, if so, the packet is abnormal network traffic information, and determining whether the packet rate exceeds a preset packet rate threshold, if so, the packet is abnormal network traffic information; (2) Based on the abnormal network traffic information, a matrix X is formed and the covariance matrix S of the samples of the matrix X is calculated; (3) Calculate the eigenvectors of the covariance matrix S , ,…, and eigenvalues t=1,2,…,n; (4) Project the data into the feature vector space, using the formula: ; Among them, the BV value is the value of the corresponding dimension in the original sample; (5) Based on the maximum The corresponding feature vector is recorded as the first feature associated with the network attack; Acquire data packets within the management period, and acquire key tuple information of the network layer and the transport layer in the data packets according to the protocol types and fields of different data packets, wherein the key tuple information includes a source IP address, a destination IP address, a source port number, a destination port number, and a protocol number, and perform feature extraction on the key tuple information to obtain a second feature associated with the network attack; Among them, feature extraction is performed on the key tuple information to obtain the second feature associated with the network attack, including: Count the number of statistical packets, average packet length, total data volume of the flow and duration of the flow based on key tuple information; The importance of each feature is evaluated by the random forest algorithm for the number of packets, average packet length, total data volume of the flow, and duration of the flow. The features are extracted in order according to the importance, and the obtained features are recorded as the second features associated with the network attack. The first and second features are fused into attack features through the bilinear attention mechanism, and then input into the fully connected layer after power normalization, and the network attack type is identified through the Softmax function: Step 1: The first feature , the second feature The fusion is performed according to the following formula: ; in, Indicates the number of training sample packets, , Represents the full connection parameters, , , are the dimensions of the first and second characteristics, respectively; Step 2: Use the attention mechanism for pooling operation, and the attention size is set to 16; Step 3: The attention mechanism is: ; ; ; in, , for The weight matrix of Indicates bias, express Cyber attack risk score, Expressed as The weight of the attack feature, is the attack feature after fusion; Step 4: Use the Softmax function to identify the type of network attack: Convert it into a probability distribution so that the sum of the probabilities of all categories is 1, and determine the type of network attack by probability; Build an IoT attack and defense model based on the type of network attack and select the optimal defense strategy.
2. The bilinear attention mechanism IoT attack and defense method according to claim 1 is characterized in that: The key tuple information of the network layer and transport layer in the data packet is obtained according to the protocol type and field of different data packets, including: Link layer protocol analysis, determining the message type based on the protocol number field in the data packet header; Network layer protocol analysis to determine whether the message is an IP message. If not, the corresponding network layer message is further analyzed. If it is an IP message, the message is judged whether IP fragments appear according to its fragmentation mark, and the next step is to perform fragment reassembly or obtain all the information of the IP packet header. Transport layer protocol analysis to obtain the port number or ACK, SYN, FIN, and RST identification information during TCP connection establishment; Application layer protocol analysis determines whether it is an HTTP message. If so, the URL, cookie information, and GET / POST / PUT request method are obtained from the HTTP message payload and the payload is analyzed.
3. The bilinear attention mechanism IoT attack and defense method according to claim 1 is characterized in that: Build an IoT attack and defense model based on the type of network attack and select the optimal defense strategy including: Establish Five tuples, and define: Gather for the players, For the defenders, for the attacker; The defender type refers to the real type of the attack target corresponding to the network attack type. Indicates real service, Represents a honeypot; is the action space of the defender and attacker, where The defender's action space, Represents the transmission of real service signals, Indicates the transmission of honeypot signal; is the attacker’s action space, Indicates that the attacker chooses to attack, Indicates that the attacker chooses to give up the attack; the defender provides four services: 1) When the defender type is When the defender maintains its own service to send a signal , providing deterrence services to send signals ; 2) When the defender type is When the defender provides a decoy service to send a signal , maintain the honeypot service to send signals ; is the attacker’s game belief set, represents the attacker's prior belief about the defender type, satisfying , represents the attacker's posterior belief about the defender type, It is calculated by the attacker using Bayes’ rule based on the received signal and prior beliefs; is the set of profit functions, represents the defender's payoff function, represents the attacker’s payoff function; Substitute the network attack type into , based on the defender benefit function, determine the candidate defense strategy corresponding to the maximum defender benefit function value, wherein the candidate defense strategy corresponding to the maximum benefit function value is used as the defense strategy, and the candidate defense strategy is a preset defense strategy.
