An adaptive intrusion response game system and method for data services under intrusion attacks
Through the adaptive intrusion response game system, the use of intrusion detection, risk assessment and game modeling technologies to dynamically adjust defense strategies, identify and respond to abnormal data, solving the problem that traditional protection methods cannot effectively respond to intrusion attacks, and realizing the energy Internet system Real-time threat monitoring and response improves system security performance.
Patent Information
- Application Number
- CN202410303722.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-18
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-03-18
AI Technical Summary
Traditional data security protection methods cannot effectively avoid intrusion attacks and lack timely response capabilities, resulting in energy Internet systems being unable to provide data services normally when facing hacker attacks.
An adaptive intrusion-response game system was designed, including security threat perceptron, strategy game generator, data filtering identifier, data recovery and real-time updater. Through intrusion detection, risk assessment, game model generation and data visualization technologies, Dynamically adjust defense policies, identify and respond to abnormal data, and update attack information in real time.
Real-time threat monitoring and response to the energy Internet system is realized, which can minimize system losses, ensure the timeliness and integrity of data, provide the best defense strategy, and improve the security performance of the system.
Smart Images

Figure CN118264443B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of network security, and in particular relates to an adaptive intrusion response game system and method for data services under intrusion attacks. Background Art
[0002] Energy Internet is a new type of energy system that realizes the interconnection and optimal configuration of energy through digital upgrades, information transformation and intelligent management. In Energy Internet, Internet of Things technology and big data analysis can monitor and process various data in real time to improve the efficiency of energy production, transmission, storage and consumption. However, since the equipment and system are in a networked state, there is a risk of hacker intrusion and attack. In order to ensure the security and stability of the energy system, researchers proposed an adaptive intrusion response game technology, which automatically adjusts the defense strategy in the system according to different intrusion attack situations to deal with security threats.
[0003] Adaptive intrusion response game is a technology that uses game theory to dynamically adjust the security defense strategy in the system to deal with various security threats by establishing a game model between intruders and defenders. This technology divides the system into two opposing game participants, namely attackers and defenders. The attacker's goal is to break through the system's security defenses, obtain confidential information or disrupt the normal operation of the system, while the defender needs to improve the system's security performance to prevent the attacker's intrusion. Intrusion response can be divided into active response and passive response. Active response techniques include revoking TCP connections, disconnecting network connections, and blocking abnormal hosts internally, and passive response techniques include alerts and notifications and isolating untrusted connections.
[0004] The development of energy internet has made energy systems more intelligent and efficient, but it has also brought security risks. With the popularization of the Internet and the continuous development of technology, the means and methods of hacker attacks are also constantly changing and upgrading, which has brought great challenges to the security of energy internet. Attackers can attack energy systems through vulnerability scanning, password cracking, denial of service attacks, etc., causing serious security threats and losses.
[0005] Traditional data security protection methods are all aimed at preventing intrusion attacks, but these technical means are difficult to truly prevent the occurrence of intrusion attacks and provide timely responses after an intrusion occurs. Ultimately, they are unable to provide normal business system data services in the energy Internet.
[0006] In summary, traditional intrusion detection technology is a passive defense technology that cannot effectively monitor activities in the energy Internet and lacks active defense capabilities and adaptive response capabilities to intrusions.
[0007] Therefore, how to improve the security performance of the energy Internet and prevent various intrusion attacks is the technical problem that the present invention aims to solve. Summary of the invention
[0008] The object of the present invention is to provide an adaptive intrusion response game system and method for data services under intrusion attacks, so as to solve the problems raised in the above background technology.
[0009] The object of the present invention is achieved as follows: an adaptive intrusion response game system for data services under intrusion attacks, characterized in that: the system includes a security threat sensor, a strategy game generator, a data filtering identifier, and a data recovery and real-time updater;
[0010] The security threat sensor includes an intrusion detection system and a risk assessment module, and the strategy game generator includes an attack and defense strategy generation module and an attack and defense game competition module;
[0011] The data filtering identifier includes a data identification module and a data visualization module, and the data recovery and real-time updater includes a data recovery module and a real-time data update module.
[0012] Preferably, the intrusion detection system is used to detect intrusion behaviors in the network, and the risk assessment module assesses the risks and threats existing in the network according to the output of the intrusion detection system, and provides a basis for subsequent decision-making;
[0013] The attack and defense strategy generation module performs quantitative analysis on the gains and losses of both parties;
[0014] The attack and defense game competition module obtains the optimal strategy for each time period to conduct competitive games, and after multiple iterations, reaches the optimal parameters for both the attack and defense sides;
[0015] After marking all data as read, the data identification module selects a core object as a seed, classifies the density reachable points of the object into one category, and forms a larger range of data object sets until all core objects have been traversed;
[0016] The data visualization module is used to display network traffic and attack source IP address information, allowing users to better identify and classify intrusion events. Users can better identify and classify intrusion events.
[0017] Preferably, the data recovery module is used to remove the identified abnormal data, and use a normal data set of different data types to take the arithmetic mean to replace the abnormal data and perform normal operations;
[0018] The real-time data updating module updates the real-time attack data, including the type, source, and target information of the attack.
[0019] An adaptive intrusion response game method for data services under intrusion attacks, characterized in that the method comprises the following steps:
[0020] Step S1: construct an intrusion detection system and input the collected data into the intrusion detection system to detect intrusion behaviors in the network;
[0021] Step S2: Perform risk assessment on the data input into the intrusion detection system to assess the risks and threats that exist;
[0022] Step S3: Establish an intrusion response game model to generate the optimal strategy according to different backgrounds and conditions;
[0023] Step S4: Screen the data of the optimal strategy, filter out abnormal data, take effective measures and visualize the data so that users can choose the optimal response strategy;
[0024] Step S5: Recover the damaged or deleted data, and continuously detect and update the current real-time data.
[0025] Preferably, in step S2, risk assessment is performed on the data input into the intrusion detection system, specifically:
[0026] Step S2-1: Establish a risk assessment model for potential threats, including the attack methods and techniques that malicious attackers may use, and assess the probability and impact of each threat;
[0027] Step S2-2: Determine the type of risk to be assessed, including asset assessment, vulnerability assessment, and risk assessment;
[0028] Step S2-3: Develop corresponding response strategies based on the risk assessment results;
[0029] For high-risk attacks, immediate measures need to be taken for defense and repair; for low-risk attacks, choose to monitor and observe and wait for further developments.
[0030] Preferably, in step S3, an intrusion response game model is established to generate an optimal strategy according to different backgrounds and conditions, specifically:
[0031] Step S3-1: When the intrusion detection system detects that the data is under attack, the intrusion response game model quantifies the gains and losses of both parties;
[0032] The intrusion response game model is established as T1, where the attacker is A and the defender is D. The attacker chooses the common method A1 or the new method A2, and the defender uses the abuse detection method D1 or the anomaly detection method D2. The strategies of the attacker and the defender are A i and D j , and their respective profit functions are defined as RA and R D , the interests of the attacker and defender are R ij (A) and R ij (D);
[0033] Step S3-2: Formulate a mixed strategy for the intrusion detection system and the user, solve the expected utility function of the payment matrix of both parties, obtain the Nash equilibrium point of the mixed strategy, and provide optimal strategy support.
[0034] Preferably, in step S3-2, a hybrid strategy of the intrusion detection system and the user is formulated, the expected utility function of the payment matrix of both parties is solved, the Nash equilibrium point of the hybrid strategy is obtained, and optimal strategy support is provided, specifically:
[0035] When the attacker's strategy A i and defender strategy D j When selected, the attack and defense game function is described as T1 = {(A, D), (F A ,F D ),(R A ,R D )}, where F A ={A1,A2},F D ={D1,D2},F A and F D They are respectively represented as the strategic space of the attacker and the defender;
[0036] The strategy matrix of the game model is M, which is expressed as follows:
[0037]
[0038] Step S3-3: Assume that the average detection accuracy of the anomaly detection method for new attacks is β, and the average detection accuracy of the misuse detection method for common attacks is α;
[0039] When M 11 =(A1,D1), the attacker chooses the common attack method, and the defender uses the misuse detection method with a detection accuracy of α:
[0040]
[0041] The benefit P gained by attacking cluster i i (t) is equal to the cost C i ′*Avg,C i ′*Avg represents the protection cluster C i The average value of ′, Q i (t) represents the attack cluster C at time t i (t) the consideration;
[0042] The defender's benefit is simplified to: R 11(D) = R(t) - C i (t)-(1-α)P i (t);
[0043] Assume that the defender uses the misuse detection method and anomaly detection method with probabilities p and 1-p, and the attacker uses the common attack method and the new attack method with probabilities q and 1-q, and obtain the value (p,q) when the attacker and the defender obtain the maximum benefit;
[0044] The binary utility matrix is used to obtain the total benefit function R of the defender and the attacker. D and R A :
[0045] R D =p(1-q)R 12 (D)+(1-p)qR 21 (D)+pqR 11 (D)+εR 22 (D)
[0046] R A =p(1-q)R 12 (A)+(1-p)qR 21 (A)+pqR 11 (A)+εR 22 (A)
[0047] Where ε is (1-p)(1-q);
[0048] The matrix Y and Z in R ij (D) and R ij (A) Bring in R D and R A :
[0049]
[0050] Find R A and R D The maximization strategy is:
[0051] Make R D Maximize R D p and R A The partial derivative of q is set equal to 0 to obtain:
[0052]
[0053] get:
[0054]
[0055] That is, strategy F A and F D for:
[0056]
[0057] When p = q, the maximum benefit for both the attacker and the defender is obtained; when α is a fixed value, the larger β is, The larger it is, the higher the accuracy of the anomaly detection method is; the attacker rationally reduces the probability of adopting new attack methods; the defender increases the probability of adopting misuse detection methods;
[0058] When β is a fixed value, the larger α is, The smaller it is, the higher the accuracy of the surface misuse detection method is, and the attacker will avoid using common attacks, as will the defender;
[0059] The attack and defense game competition module combines the optimal strategies of the attacker and defender with the ecological race competition model. The center of the game competition between the attacker and defender is the various resources in the energy Internet. It iterates and evolves according to the strategies under different conditions, and finally reaches the balance point between the two sides.
[0060] Quantifying the attack and defense sides respectively, we get:
[0061]
[0062] Among them A n and D n Represent the number of attackers and defenders respectively; γ1 and γ2 represent the recovery rate of attackers and defenders respectively; Q1 and Q2 represent the maximum capacity of the two sides; the strategic strength of both sides is F Ai and F Dj .
[0063] Preferably, the binary utility matrix is:
[0064]
[0065] Among them, the matrices Y and Z represent the utility matrices of the attacker and defender respectively, and the rows and columns represent the strategies of the attacker and defender.
[0066] Preferably, the abnormal data in step S4 is data that is not classified into a category, that is, the abnormal data. The data recovery module removes the identified abnormal data and replaces the abnormal data with the arithmetic mean of a normal data set of different data types to perform normal operations.
[0067] Compared with the prior art, the present invention has the following improvements and advantages:
[0068] 1. Through the intrusion detection system, detection and analysis are first carried out, and then risk assessment is carried out. Response strategies are formulated for different risk assessment results. By establishing a game model, a hybrid strategy is generated, and after optimizing the parameter values, abnormal data is classified, and the abnormal data is restored or deleted. Finally, the data is monitored and updated in real time to ensure the timeliness and accuracy of the data, and to ensure the integrity and availability of the system data.
[0069] 2. The method of the present invention adopts an adaptive approach to select the optimal response strategy, and adaptively responds to network attacks on data services, which can minimize system losses; a real-time data updater module is used to monitor and update these data in real time to ensure the timeliness and accuracy of the data for subsequent security analysis and intrusion response. BRIEF DESCRIPTION OF THE DRAWINGS
[0070] Figure 1 It is the overall framework diagram of the system of the present invention.
[0071] Figure 2 Flow chart of the method of the present invention.
[0072] Figure 3 FIG. 1 is a schematic diagram of the transmission line of the present invention in the IEEE 14-bus system.
[0073] Figure 4 Schematic diagram of the voltage values of the remaining nodes using the optimal strategy in the first time period.
[0074] Figure 5 Schematic diagram of the voltage values of the remaining nodes using the optimal strategy in the second time period.
[0075] Figure 6 It is a schematic diagram of the final voltage value when the method of the present invention is not adopted. DETAILED DESCRIPTION
[0076] The present invention is further summarized below with reference to the accompanying drawings.
[0077] like Figure 1 As shown, an adaptive intrusion response game system for data services under intrusion attacks, the system includes a security threat sensor, a strategy game generator, a data filtering identifier, and a data recovery and real-time updater;
[0078] Security threat sensors include intrusion detection systems and risk assessment modules. Intrusion detection systems detect intrusion behaviors in the network, usually including intrusion detection technologies based on rules, statistics, machine learning or deep learning;
[0079] The risk assessment module evaluates the possible risks and threats in the network based on the output of the intrusion detection system and provides a basis for subsequent decision-making.
[0080] The strategy game generator includes an attack and defense strategy generation module and an attack and defense game competition module. When the data detection system detects an attack, the attack and defense strategy generation module conducts a quantitative analysis of the gains and losses of both parties; the attack and defense game competition module conducts a competitive game after obtaining the optimal strategy for each time period, and after multiple iterations, it reaches the optimal parameters for both the attack and defense parties; the strategy game generator defines variables based on the impact of IDS response and user intrusion.
[0081] The data filter identifier includes a data identification module and a data visualization module. After marking all data as read, the data identification module selects a core object as a seed, classifies the density reachable points of the object into one category, and forms a larger range of data object sets until all core objects have been traversed; data that is not classified into one category is abnormal data;
[0082] The data visualization module is used to display network traffic and attack source IP address information, allowing users to better identify and classify intrusion events. In terms of response, the data visualization module can display the effects, time and other information of different response strategies, allowing users to choose the optimal response strategy. The data visualization module can improve the efficiency and accuracy of data analysis and decision-making, helping users better protect system security.
[0083] The data recovery and real-time updater includes a data recovery module and a real-time data update module. The data recovery module is used to remove the identified abnormal data and replace the abnormal data with the arithmetic mean of a normal data set of different data types to perform normal operations; the data encryptor and decryptor can be configured in the data recovery module to protect the confidentiality and integrity of sensitive data;
[0084] The real-time data update module collects and updates real-time attack data, including the type, source, target, etc. of the attack. The real-time data update module can help the system detect intrusions in a timely manner and update system data and rules in a timely manner to ensure the security and stability of the system.
[0085] like Figure 2 As shown, an adaptive intrusion response game method for data services under intrusion attacks includes the following steps:
[0086] Step S1: construct an intrusion detection system and input the collected data into the intrusion detection system to detect intrusion behaviors in the network;
[0087] Intrusion Detection System (IDS) is an important part of the adaptive intrusion response game method. IDS is adaptive, that is, it can dynamically adjust and optimize according to the current attack situation and system status to improve detection accuracy and reduce false alarm rate; intrusion detection system usually includes the following parts:
[0088] Data collection: responsible for collecting traffic, system logs and other data, and pre-processing the data;
[0089] Anomaly detection: Based on machine learning, statistical analysis and other technologies, the collected data is analyzed to determine whether there is any abnormal behavior.
[0090] Rule detection: Match and detect network traffic, system logs and other data based on predefined rule sets.
[0091] Alarm generation: When abnormal behavior is detected or a rule is matched, a corresponding alarm is generated and relevant personnel or other security systems are notified for further response and processing.
[0092] Log records: record system operation status, abnormal events, alarm information, etc. for subsequent analysis and auditing.
[0093] Step S2: Perform risk assessment on the data input into the intrusion detection system to assess the risks and threats that exist;
[0094] Step S2-1: Model potential threats, including the attack methods and techniques that malicious attackers may use, and assess the probability and impact of each threat;
[0095] Model potential threats, specifically:
[0096] Input various custom attack methods in advance, classify them according to their threat levels, and define the threat level for each;
[0097] Table 1 shows the attack threat classification level
[0098]
[0099]
[0100] As shown in Table 1, six attack methods of different intensities are classified. The smaller the value of the level, the greater the intensity of the attack method and the greater the threat to the system. Therefore, for attack methods with lower intensity, such as social engineering attacks, the system can be monitored and observed first; in the face of attack scenarios with higher intensity, corresponding defense strategies need to be adopted according to different attack intensities, and the best strategies for each are generated according to the strategy game generator.
[0101] Step S2-2: Determine the type of risk to be assessed, including asset assessment, vulnerability assessment, and risk assessment;
[0102] Threat assessment: Evaluate the attack methods in the attack scenario to provide a basis for formulating corresponding response strategies.
[0103] Vulnerability assessment: Evaluate possible vulnerabilities and weaknesses in data services.
[0104] Risk Assessment: Combine threat and asset assessment with vulnerability assessment to assess and quantify the risk of the entire system. Through the assessment of risks, priority response measures need to be taken.
[0105] Step S2-3: Develop corresponding response strategies based on the risk assessment results;
[0106] Adaptive response strategy: According to the risk assessment results, formulate corresponding response strategies. For high-risk attacks, immediate measures need to be taken to defend and repair. For low-risk attacks, you can choose to monitor and observe and wait for further developments.
[0107] Step S3: Establish an intrusion response game model to generate the optimal strategy according to different backgrounds and conditions;
[0108] Step S3-1: When the intrusion detection system detects that the data is under attack, the intrusion response game model quantifies the gains and losses of both parties;
[0109] The intrusion response game model is established as T1, where the attacker is A and the defender is D. The attacker chooses the common method A1 or the new method A2, and the defender uses the abuse detection method D1 or the anomaly detection method D2. The strategies of the attacker and the defender are A i and D j , and their respective profit functions are defined as R A and R D , the interests of the attacker and defender are R ij (A) and R ij (D);
[0110] Step S3-2: Formulate a hybrid strategy for the intrusion detection system and the user, solve the expected utility function of the payment matrix of both parties, obtain the Nash equilibrium point of the hybrid strategy, and provide optimal strategy support;
[0111] When the attacker's strategy A i and defender strategy D j When selected, the attack and defense game function is described as T1 = {(A, D), (F A ,F D ),(R A,R D )}, where F A ={A1,A2},F D ={D1,D2},F A and F D They are respectively represented as the strategic space of the attacker and the defender;
[0112] The strategy matrix of the game model is M, which is expressed as follows:
[0113]
[0114] Step S3-3: Assume that the average detection accuracy of the anomaly detection method for new attacks is β, and the average detection accuracy of the misuse detection method for common attacks is α;
[0115] When M 11 =(A1,D1), the attacker selects the attack method and the defender uses the misuse detection method with a detection accuracy of α:
[0116]
[0117] The benefit P gained by attacking cluster i i (t) is equal to the cost C′ i *Avg, C′ i *Avg represents the protected cluster C′ i The average value, Q i (t) represents the attack cluster C at time t i (t) the consideration;
[0118] Matrices Y and Z represent the utility matrices of the attacker and defender, respectively, where the rows and columns represent the strategies of the attacker and defender. The binary utility matrix is expressed as follows:
[0119] The binary utility matrix is:
[0120]
[0121] Among them, the matrices Y and Z represent the utility matrices of the attacker and defender respectively, and the rows and columns represent the strategies of the attacker and defender.
[0122] The defender's benefit is simplified to: R 11 (D) = R(t) - C i (t)-(1-α)P i (t);
[0123] Assume that the defender uses the misuse detection method and anomaly detection method with probabilities p and 1-p, and the attacker uses the common attack method and the new attack method with probabilities q and 1-q, and obtain the value (p,q) when the attacker and the defender obtain the maximum benefit;
[0124] The binary utility matrix is used to obtain the total benefit function R of the defender and the attacker. D and R A :
[0125] R D =p(1-q)R 12 (D)+(1-p)qR 21 (D)+pqR 11 (D)+εR 22 (D)
[0126] R A =p(1-q)R 12 (A)+(1-p)qR 21 (A)+pqR 11 (A)+εR 22 (A)
[0127] Where ε is (1-p)(1-q);
[0128] The matrix Y and Z in R ij (D) and R ij (A) Bring in R D and R A :
[0129]
[0130] Find R A and R D The maximization strategy is:
[0131] Make R D Maximize R D p and R A The partial derivative of q is set equal to 0 to obtain:
[0132]
[0133] get:
[0134]
[0135] That is, strategy F A and F D for:
[0136]
[0137] When p = q, the maximum benefit for both the attacker and the defender is obtained; when α is a fixed value, the larger β is, The larger it is, the higher the accuracy of the anomaly detection method is; the attacker rationally reduces the probability of adopting new attack methods; the defender increases the probability of adopting misuse detection methods;
[0138] When β is a fixed value, the larger α is, The smaller it is, the higher the accuracy of the surface misuse detection method is, and the attacker will avoid using common attacks, as will the defender.
[0139] The attack and defense game competition module combines the optimal strategies of the attacker and defender with the ecological race competition model. The center of the game competition between the attacker and defender is the various resources in the energy Internet. Iterate and evolve according to the strategies under different conditions, and finally reach the balance point between the two sides:
[0140] Quantify the attack and defense sides respectively, and get the formula and formula A and D represent the number of attackers and defenders, respectively; α and β represent the recovery rates of attackers and defenders, respectively; X and Y represent the maximum capacity of the two sides; the strategic strength of both sides is F Ai and F Dj ;
[0141] An attack and defense game model is established, and a strategy game generator is obtained by combining the attack and defense strategy generation module.
[0142] The strategy game generator takes the intrusion detection model and response strategy as participants in the game, establishes an adaptive intrusion response game model, and then helps game theory researchers generate different types of hybrid strategies to study issues such as the equilibrium point of the game and strategy optimization.
[0143] Step S4: Screen the data of the optimal strategy, filter out abnormal data, take effective measures and visualize the data so that users can choose the optimal response strategy;
[0144] Step S5: Recover the damaged or deleted data, and continuously detect and update the current real-time data.
[0145] The main workflow of an adaptive intrusion response game method for data services under intrusion attacks is:
[0146] Step 1: First, data collection is performed, which is responsible for collecting data such as traffic and system logs, and preprocessing the data.
[0147] Step 2: Perform anomaly detection, analyze the collected data based on machine learning, statistical analysis and other technologies, and determine whether there is any abnormal behavior.
[0148] Step 3: Generate an alarm. When abnormal behavior is detected or a rule is matched, a corresponding alarm is generated.
[0149] Step 4: Record system operation status, abnormal events, alarm information, etc.
[0150] Step 5: Input attack methods in different scenarios into the risk assessment module in advance, customize the attack methods and establish a threat level index.
[0151] Step 6: Evaluate the alarm behavior detected in step 3 to obtain the corresponding threat level index.
[0152] Step 7: Combine threat and asset assessment with vulnerability assessment to assess and quantify the risk of the entire system. Through the assessment of risk, priority response measures need to be taken.
[0153] Step 8: Based on the risk assessment results, for low-risk attacks, you can choose to monitor and observe and wait for further developments; for other risk attacks, enter the strategy game generator.
[0154] Step 9: Adaptive intrusion response game modeling for data services under intrusion attacks, generating the optimal strategy F for both the attacker and the defender according to different backgrounds and conditions Ai and F Dj
[0155] Step 10: Combine the optimal strategies of the attacking and defending parties in step 9 with the racial competition model to establish an attack-defense game competition model. After iterating the optimal strategies at different times, the optimal balance point between the two parties is reached to achieve the minimum loss of the system.
[0156] Step 10: Enter the target identification module. After all data are marked as read, the target identification module selects a core object as a seed and classifies the density reachable points of the object into one category. A larger range of data object sets is formed until all core objects have been traversed. Those data that are not classified into one category are abnormal data. If all data have been read, proceed to step 11, otherwise return to step 10 to continue.
[0157] Step 11: The data enters the data visualization module, which can display the effects, time and other information of different response strategies. It can improve the efficiency and accuracy of data analysis and decision-making.
[0158] Step 12: Enter the data recovery module and restore the damaged or deleted data as much as possible through backup data or other means to ensure the integrity and availability of system data. Go to step 7.
[0159] Step 13: Finally, the real-time data updater module monitors and updates the data in real time to ensure the timeliness and accuracy of the data for subsequent security analysis and intrusion response.
[0160] In order to verify the effect of the method of the present invention, a simulation experiment is carried out on the IEEE 14-bus system; Figure 3 As shown, Figure 31-14 are all transmission lines, the capacity of transmission line 1 is 160MW, and the capacity of transmission lines 2-14 is 60MW; it is assumed that each busbar has a measurement unit and each branch has a measurement unit;
[0161] Table 2 shows the parameters of the IEEE14 node example;
[0162]
[0163] The experiment takes the node voltage of the IEEE 14-bus system as the center of contention between the attacker and the defender, and the final voltage obtained by the proposed game method is the reference object;
[0164] Table 3 Final voltage using the optimal strategy
[0165] node Initial voltage (V) Final voltage(V) 1 1.06 0.742 2 1.045 0.730 3 1.01 0.702 4 1.018 0.708 5 1.02 0.710 6 1.07 0.749 7 1.062 0.743 8 1.09 0.765 9 1.056 0.739 10 1.051 0.735 11 1.057 0.739 12 1.055 0.738 13 1.05 0.734 14 1.036 0.723
[0166] Table 4 Final voltage without using the optimal strategy
[0167]
[0168]
[0169] Table 3 is divided into two stages of optimal strategy iteration. Table 4 does not use the optimal strategy, takes the time unit as 30, and then combines Figure 4 , Figure 5 , Figure 6 The final voltage display of the node is observed, which proves the effectiveness and feasibility of the method of the present invention.
[0170] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims
1. An adaptive intrusion response game method for data services under intrusion attacks, characterized by: The method comprises the following steps: Step S1: construct an intrusion detection system and input the collected data into the intrusion detection system to detect intrusion behaviors in the network; Step S2: Perform risk assessment on the data input into the intrusion detection system to assess the risks and threats that exist; Step S3: Establish an intrusion response game model to generate the optimal strategy according to different backgrounds and conditions; An intrusion response game model is established to generate the optimal strategy according to different backgrounds and conditions, specifically: Step S3-1: When the intrusion detection system detects that the data is under attack, the intrusion response game model quantifies the gains and losses of both parties; The intrusion response game model is established as T1, where the attacker is A and the defender is D. The attacker chooses the common method A1 or the new method A2, and the defender uses the abuse detection method D1 or the anomaly detection method D2. The strategies of the attacker and the defender are A i and D j , and their respective profit functions are defined as R A and R D , the interests of the attacker and defender are R ij (A) and R ij (D); Step S3-2: Formulate an intrusion detection system and user hybrid strategy to provide optimal strategy support; Provide optimal strategy support, specifically: When the attacker's strategy A i and the defensive strategy D j When selected, the attack and defense game function is described as T1 = {(A, D), (F A ,F D ),(R A ,R D )}, where F A ={A1,A2},F D ={D1,D2},F A and F D They represent the strategic space of the attacker and defender respectively; A represents the attacker and D represents the defender; The strategy matrix of the game model is M, which is expressed as follows: Assume that the average detection accuracy of the anomaly detection method for the new method A2 is β, and the average detection accuracy of the abuse detection method for the common method A1 is α; When M 11 =(A1,D1), the attacker chooses the common attack method, and the defender uses the abuse detection method, with an average detection accuracy of α: R 11 (A)=(1-α)(P i (t)-Q i (t))-αQ i (t)=(1-α)P i (t)-Q i (t) R 11 (D)=(1-α)(R(t)-C i (t)-C′ i *Avg)+α(R(t)-C i (t)); =R(t)-C i (t)-(1-α)C′ i *Avg. The benefit P gained by attacking cluster i i (t) is equal to the cost C i ′*Avg,C i ′*Avg represents the protection cluster C i The average value of ′, Q i (t) represents the attack cluster C at time t i (t) the consideration; The defender's benefit is simplified to: R 11 (D) = R(t) - C i (t)-(1-α)P i (t); Assume that the defender uses the abuse detection method and anomaly detection method with probabilities p and 1-p, and the attacker uses the common attack method and the new attack method with probabilities q and 1-q, and obtain the value (p,q) when the attacker and the defender obtain the maximum benefit; The binary utility matrix is used to obtain the total benefit function R of the defender and the attacker. D and R A : The binary utility matrix is: Among them, the matrices Y and Z represent the utility matrices of the attacker and defender respectively, and the rows and columns represent the strategies of the attacker and defender; Where ε is (1-p)(1-q); The matrix Y and the matrix Z in R ij (D) and R ij (A) Bring in R D and R A : Find R A and R D The maximization strategy is: Make R D Maximize R D p and R A The partial derivative of q is set equal to 0 to obtain: get: That is, strategy F A and F D for: When p = q, the maximum benefit for both the attacker and the defender is obtained; when α is a fixed value, the larger β is, The larger the value, the higher the accuracy of the anomaly detection method; the attacker is less likely to use new attack methods; the defender is more likely to use abuse detection methods; When β is a fixed value, the larger α is, The smaller it is, the higher the accuracy of the abuse detection method is, and the attacker will avoid using common attacks, and the same is true for the defender; The attack and defense game competition module combines the optimal strategies of the attacker and defender with the competition model of ecology. The center of the game competition between the attacker and defender is the resources in the energy Internet. Iterate and evolve according to different strategies to finally reach the balance point between the two sides. Quantifying the attack and defense sides respectively, we get: Among them A n and D n Represent the number of attackers and defenders respectively; γ1 and γ2 represent the recovery rate of attackers and defenders respectively; Q1 and Q2 represent the maximum capacity of the two sides; the strategic strength of both sides is F Ai and F Dj ; Step S4: Filter the data of the optimal strategy, filter out abnormal data, and visualize the data for users to select the optimal response strategy; Step S5: Recover the damaged or deleted data, and continuously detect and update the current real-time data.
2. The adaptive intrusion response game method for data services under intrusion attacks according to claim 1 is characterized by: In step S2, risk assessment is performed on the data input into the intrusion detection system, specifically: Step S2-1: Model potential threats, including the attack methods and techniques used by malicious attackers, and assess the probability and impact of each threat; Step S2-2: Determine the type of risk to be assessed, including asset assessment, vulnerability assessment, and risk assessment; Step S2-3: Develop corresponding response strategies based on the risk assessment results; For high-risk attacks, immediate measures need to be taken for defense and repair; for low-risk attacks, choose to monitor and observe and wait for further developments.
3. The adaptive intrusion response game method for data services under intrusion attacks according to claim 1 is characterized by: In step S4, the abnormal data is data that is not classified into a category, that is, the abnormal data. The data recovery module removes the identified abnormal data and replaces the abnormal data with the arithmetic mean of a normal data set of different data types to perform normal operations.
4. The adaptive intrusion response game method for data services under intrusion attacks according to any one of claims 1 to 3 generates an adaptive intrusion response game system for data services under intrusion attacks, characterized in that: The system includes a security threat sensor, a strategy game generator, a data filtering identifier, and a data recovery and real-time updater; The security threat sensor includes an intrusion detection system and a risk assessment module, and the strategy game generator includes an attack and defense strategy generation module and an attack and defense game competition module; The data filtering identifier includes a data identification module and a data visualization module, and the data recovery and real-time updater includes a data recovery module and a real-time data update module.
5. The adaptive intrusion response game system for data services under intrusion attacks according to claim 4, characterized in that: The intrusion detection system is used to detect intrusion behaviors in the network. The risk assessment module assesses the risks and threats existing in the network based on the output of the intrusion detection system and provides a basis for subsequent decision-making; The attack and defense strategy generation module performs quantitative analysis on the gains and losses of both parties; The attack and defense game competition module obtains the optimal strategy for each time period to conduct competitive games, and after multiple iterations, reaches the optimal parameters for both the attack and defense sides; After marking all data as read, the data identification module selects a core object as a seed, classifies the density reachable points of the object into one category, and forms a data object set until all core objects have been traversed; The data visualization module is used to display network traffic and attack source IP address information, allowing users to identify and classify intrusion events.
6. The adaptive intrusion response game system for data services under intrusion attacks according to claim 4, characterized in that: The data recovery module is used to remove the identified abnormal data and use the arithmetic mean of data sets of different data types to replace the abnormal data to perform normal operations; The real-time data updating module updates the real-time attack data, including the type, source, and target information of the attack.
Citation Information
Patent Citations
Data mining based invasion detection system with Oracle as core
CN105468995A
Network risk analysis and optimal active defense method based on multi-level game model
CN108683664A
Network cooperative attack modeling and hazard quantitative analysis method based on game theory
CN115296830A
Power distribution network defense strategy design method based on dynamic game under network attack
CN115348064A
Industrial information physical system defense strategy generation method based on Bayesian random game
CN117040809A
