A computer system security model verification method

The computer firewall extracts and analyzes attack data and combines the random forest algorithm to build a security verification model, which solves the problem that traditional methods cannot accurately verify the computer system security model, and achieves accurate restoration of attack scenarios and characteristics and improves system security.

CN118536124BActive Publication Date: 2024-11-26王延丽
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410882227.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-03
Publication Date
2024-11-26
Estimated Expiration
2044-07-03

AI Technical Summary

Technical Problem

Traditional computer system security model verification methods cannot accurately restore attack scenarios and potential attack characteristics, resulting in the inability to accurately verify the computer system.

Method used

Through the computer firewall, the attack interception data is extracted, the attack type is identified, the attack behavior pattern is analyzed, the attack behavior mapping and situational awareness are carried out, the intrusion detection rules are designed, the security verification model is built based on the random forest algorithm, and the system security risk detection and configuration optimization are carried out.

Benefits of technology

Accurate model verification of computer systems is realized, security and defense capabilities are improved, potential threats are identified and responded in a timely manner, system configuration is optimized, and attack risks are reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118536124B_ABST
    Figure CN118536124B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of security model verification, and in particular to a computer system security model verification method. The method comprises the following steps: extracting attack interception data from a computer system and identifying attack types to obtain attack type data; analyzing attack behavior patterns and performing attack behavior mapping processing on the attack type data to obtain attack type behavior mapping data; performing attack behavior situation awareness and system abnormality feedback analysis based on the attack type behavior mapping data to obtain system abnormality feedback data; designing intrusion detection rules based on the system abnormality feedback data to obtain intrusion detection rule data; constructing a computer system security verification model for the intrusion detection rule data, performing system security risk detection, and optimizing system configuration to obtain system configuration optimization data. The present invention makes the verification technology more accurate by optimizing the security model verification technology.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of security model verification technology, and in particular to a method for verifying computer system security models. Background Technology

[0002] Computer system security model verification is a crucial means of ensuring that computer systems can withstand various attacks and threats during design and operation. It encompasses knowledge of cryptography, access control, security protocols, vulnerability analysis, security policies, and security management. Cryptography provides technologies such as encryption algorithms and digital signatures to protect the confidentiality and integrity of data; access control technologies restrict user access to system resources; security protocols are communication rules built upon cryptography to ensure security during communication; vulnerability analysis is a key step in discovering potential vulnerabilities and weaknesses in the system for timely patching; and security policies and management are crucial for formulating and implementing system security policies, including risk assessment, security training, and security awareness. However, traditional computer system security model verification methods suffer from the inability to accurately recreate attack scenarios and potential attack characteristics, thus hindering accurate model verification of computer systems. Summary of the Invention

[0003] Therefore, it is necessary to provide a computer system security model verification method to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a computer system security model verification method is provided, the method comprising the following steps:

[0005] Step S1: Extract attack interception data from the computer system through the computer firewall to obtain firewall attack interception data; identify attack types based on the firewall attack interception data to obtain attack type data; analyze attack behavior patterns in the attack type data to obtain attack type behavior pattern data.

[0006] Step S2: Perform attack behavior mapping processing on the attack type behavior pattern data to obtain attack type behavior mapping data; perform attack behavior situation awareness based on the attack type behavior mapping data to obtain attack behavior situation awareness data; perform system anomaly feedback analysis on the computer system based on the attack type behavior mapping data and the attack behavior situation awareness data to obtain system anomaly feedback data.

[0007] Step S3: Design intrusion detection rules based on system anomaly feedback data to obtain intrusion detection rule data; construct a computer system security verification model based on the intrusion detection rule data using the random forest algorithm to obtain the system security verification model;

[0008] Step S4: Perform system security risk detection based on the system security verification model to obtain system security risk data; optimize system configuration based on system security risk data to obtain system configuration optimization data.

[0009] This invention utilizes computer firewalls to extract attack interception data from computer systems, enabling timely identification of malicious attacks and protecting system security. Attack type identification based on firewall attack interception data determines specific attack types, helping security teams better understand and respond to different attack types. Attack behavior pattern analysis of attack type data reveals attacker behavior patterns and methods, aiding in predicting future attacks and improving system security and defense capabilities. Attack behavior mapping of attack type behavior pattern data associates attack behaviors with specific systems or resources, facilitating a better understanding of attacker goals and intentions. Attack behavior situational awareness based on attack type behavior mapping data allows for real-time monitoring and analysis of attacker behavior dynamics, rapid identification and response to potential threats, and improved system security. Furthermore, system anomaly feedback analysis of the computer system based on attack type behavior mapping data and attack behavior situational awareness data can identify… This invention addresses anomalies in the system, helping to promptly identify and fix potential vulnerabilities or security risks, improving system stability and reliability. Intrusion detection rule design based on system anomaly feedback data can extract patterns and rules related to known intrusion behaviors, helping the system to promptly detect and intercept intrusion attempts, thus improving system security. Building a computer system security verification model based on the random forest algorithm for intrusion detection rule data can establish an effective model for judging and evaluating the system's security status and risk level. System security risk detection based on the system security verification model can assess existing security risks and potential threats, helping system administrators or security teams to take timely measures to address and prevent risks, improving system security. Analyzing and processing system security risk data can identify configuration vulnerabilities, weaknesses, or insecure settings in the system, and provide corresponding configuration optimization suggestions, helping system administrators improve system configuration and reduce the probability and impact of system attacks. Therefore, this invention provides a computer system security model verification method that solves the problem of traditional computer system security model verification methods failing to accurately recreate attack scenarios and potential attack characteristics, thus hindering accurate model verification of computer systems. This invention accurately recreates attack scenarios and potential attack characteristics, enabling accurate model verification of computer systems.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Extract attack interception data from the computer system through the computer firewall to obtain firewall attack interception data;

[0012] Step S12: Analyze the attack targets of the firewall attack interception data to obtain the attack target data;

[0013] Step S13: Identify the attack type based on the firewall attack interception data and the attack target data to obtain attack type data;

[0014] Step S14: Perform attack behavior pattern analysis on the attack type data to obtain attack type behavior pattern data.

[0015] This invention's firewall can monitor and record network traffic. By extracting firewall attack interception data, it can detect potential malicious attacks, such as network intrusions, port scans, and denial-of-service attacks. The firewall attack interception data provides real-time attack information, enabling security teams to respond quickly and take appropriate countermeasures, such as blocking the attack source IP address and modifying access control rules, thereby reducing the harm caused by malicious attacks. The firewall attack interception data records detailed information on network activity and attack events. Auditing and analyzing this data helps security teams understand attack trends, detect system vulnerabilities, and improve security measures and strategies accordingly. Attack target analysis can determine the attacker's goals, attack methods, and attack techniques, helping to identify weaknesses and sensitive resources in the system. By analyzing attack target data, the threat level and potential losses of different attack behaviors can be assessed, providing guidance for subsequent security defenses and optimizations. Attack target analysis may reveal unknown vulnerabilities or security weaknesses in the system, allowing the security team to fix them in a timely manner. These vulnerabilities enhance system security. By analyzing and correlating firewall attack interception data and attack target data, different types of attacks can be identified, such as network scanning, malware propagation, and authentication attacks. This helps to accurately understand the attacker's strategies and methods. Based on attack type data, security teams can adjust and optimize defense strategies such as firewalls and intrusion detection systems, strengthening the interception and defense against known attack types, thereby improving system security and protection capabilities. Attack type behavior pattern data can reveal the behavioral patterns and habits of malicious attackers. By analyzing this data, possible future attack behaviors can be predicted, and corresponding preventive measures can be taken. Attack type behavior pattern data helps identify abnormal behaviors in the system, such as a large number of login failures and abnormal data transmission. It can promptly detect attack behaviors or potential threats and take appropriate response measures. By analyzing attack type behavior pattern data, existing security protection strategies and measures can be optimized, improving the system's ability to detect and block known and unknown attack behaviors.

[0016] Preferably, step S14 includes the following steps:

[0017] Step S141: Extract the interception logs from the firewall attack interception data to obtain the interception log data;

[0018] Step S142: Perform attack type log matching on the attack type data based on the interception log data to obtain attack type log matching data;

[0019] Step S143: Extract network traffic of attack type from attack type log matching data to obtain attack type network traffic data;

[0020] Step S144: Perform instantaneous geometric growth analysis on network traffic data based on attack type to obtain instantaneous geometric growth data of traffic;

[0021] Step S145: Identify potential growth patterns in the instantaneous geometric growth data of traffic to obtain potential growth pattern data of traffic.

[0022] Step S146: Analyze the attack behavior patterns of the attack type data based on the potential growth pattern data of traffic to obtain attack type behavior pattern data.

[0023] This invention extracts detailed information about attack behavior from intercept logs, including attack time, source IP address, and target, providing data support for subsequent analysis and processing. Interception log data can track the source and destination of attack activities, helping to determine attacker behavior patterns and strategies. The data records provided by the intercept logs help security teams audit security incidents and ensure compliance with relevant requirements. By matching and analyzing attack information in the intercept log data, attack behaviors can be associated with corresponding attack types, further refining and classifying attack data. Obtaining attack type log matching data improves the accuracy and reliability of attack type data, thus providing a more precise understanding of attack characteristics and patterns. By extracting network traffic from attack type log matching data, it can be associated with specific attack types, further analyzing and understanding attack behavior and characteristics. Attack type network traffic data provides an overall overview of attack activities, including the quantity, frequency, source, and target of attack traffic, providing a foundation for further analysis and modeling. The data on instantaneous geometric growth in traffic is used to identify abnormal network traffic, i.e., traffic growth that is faster than normal, thereby discovering potential malicious attacks or network anomalies. By analyzing the magnitude and trend of traffic growth, the impact of attack behavior on network performance and services can be assessed, and security teams can be helped to determine the threat level of malicious attacks. Data on potential traffic growth patterns is used to identify common traffic growth patterns in malicious attacks, such as the attack's initiation process, duration, and peak periods, which helps to build attack behavior models. By analyzing potential traffic growth patterns, security teams can predict the development trend of attack behavior and take corresponding security strategies and measures for response and defense. By analyzing the relationship between potential traffic growth pattern data and attack type data, we can gain a deeper understanding of the patterns and evolution of attack behavior, such as the attack's initiation stage, duration, and changing trends. Attack type behavior pattern data provides security teams with a basis for optimizing security strategies, enabling them to develop more effective security measures and defense strategies for specific attack types.

[0024] Preferably, the analysis of instantaneous geometric growth in network traffic based on attack type data includes the following steps:

[0025] Time-series analysis is performed on network traffic data of attack types to obtain time-series data of network traffic of attack types.

[0026] The network traffic time series data of attack type is smoothed by performing time series smoothing on the network traffic time series data of attack type to obtain traffic time series smoothed data.

[0027] Based on the traffic time-series smoothing data, an instantaneous traffic growth Gaussian process analysis was performed on the network traffic data of attack types to obtain instantaneous traffic growth Gaussian process data;

[0028] Based on the Gaussian process data of instantaneous flow growth, the flow time-series smoothed data is processed by flow time-series segmentation to obtain flow time-series segmented data.

[0029] Based on the Gaussian process data of instantaneous flow growth, the geometric series growth of flow time-series segmented data is calculated to obtain segmented flow geometric growth data;

[0030] Based on the Gaussian process data of instantaneous flow rate growth and the segmented geometrical growth data of flow rate, an analysis of the instantaneous geometrical growth of flow rate was conducted to obtain the instantaneous geometrical growth data of flow rate.

[0031] This invention, through time-series analysis of attack-type network traffic data, can obtain the changing trends and periodicity of attack-type network traffic. This helps to better understand the dynamic behavior of traffic. By smoothing the time-series data of attack-type network traffic, noise and outliers can be removed, making the data more stable and reliable. By performing Gaussian process analysis on the smoothed time-series traffic data, the probability distribution information of traffic growth can be obtained. This helps to understand the growth of attack-type network traffic and identify abnormal growth events. Based on the Gaussian process data of instantaneous traffic growth, segmenting the smoothed time-series traffic data can more accurately describe the characteristics and trends of traffic changes. Based on the Gaussian process data of instantaneous traffic growth, geometric progression calculation of the growth of segmented time-series traffic can obtain the geometric growth rate of traffic in each time period. This helps to quantify the growth trend and speed of traffic. Based on the Gaussian process data of instantaneous traffic growth and segmented geometric growth data, instantaneous geometric multiple growth analysis of traffic can be performed. This can provide more comprehensive traffic growth indicators, helping analysts understand the growth magnitude and speed of attack-type network traffic.

[0032] Preferably, step S2 includes the following steps:

[0033] Step S21: Perform attack behavior mapping processing on the attack type behavior pattern data based on the firewall attack interception data to obtain attack type behavior mapping data;

[0034] Step S22: Perform attack behavior frequency statistics on the attack type behavior mapping data to obtain attack behavior frequency data;

[0035] Step S23: Analyze the attack heat map distribution of the computer system based on the attack type behavior mapping data and the attack behavior frequency data to obtain attack heat map data;

[0036] Step S24: Perform attack behavior situation awareness on the attack heat map data based on the attack type behavior mapping data to obtain attack behavior situation awareness data;

[0037] Step S25: Perform system anomaly feedback analysis on the computer system based on attack type behavior mapping data and attack behavior situational awareness data to obtain system anomaly feedback data.

[0038] This invention maps attack type behavior pattern data to specific behavioral patterns based on firewall attack interception data. This helps to more accurately understand the behavioral characteristics of different types of attacks, providing a useful data foundation for subsequent analysis. Frequency statistics of the attack type behavior mapping data can calculate the frequency of different attack behaviors. This helps to identify which attack behaviors are more common, allowing for more targeted responses and prevention. Based on the attack type behavior mapping data and attack behavior frequency data, attack heat map distribution analysis of computer systems can identify the areas or parts of the system most heavily attacked. This can be used to identify weak points and vulnerable areas in the system, facilitating focused protection and defense. Analyzing attack heat map data based on the attack type behavior mapping data enables attack behavior situational awareness. This means timely understanding of the current attack behavior situation and trends, identifying potential attack threats, and taking timely defensive and response measures. Based on the attack type behavior mapping data and attack behavior situational awareness data, system anomaly feedback analysis can be performed. This helps to identify abnormal behaviors and events in computer systems and provide corresponding feedback information. By analyzing abnormal feedback data, abnormal situations in the system can be quickly identified, providing a basis for further safe operations and interventions.

[0039] Preferably, step S24 includes the following steps:

[0040] Step S241: Capture intrusion attempts on the attack heat map data based on the attack type behavior mapping data to obtain attack type intrusion capture data;

[0041] Step S242: Perform an intrusion progressive strength assessment on the attack type intrusion capture data to obtain intrusion progressive strength data;

[0042] Step S243: Perform biased progressive concealment analysis on the progressive intensity data of intrusion to obtain biased progressive concealment data of intrusion.

[0043] Step S244: Perform intrusion intensity compensation on the intrusion progressive intensity data based on the intrusion bias progressive concealment data to obtain intrusion progressive intensity compensation data.

[0044] Step S245: Based on the intrusion progressive intensity compensation data and the intrusion bias progressive concealment data, perform attack behavior situational awareness on the attack heat map data to obtain attack behavior situational awareness data.

[0045] This invention analyzes and captures attack heatmap data based on attack type behavior mapping data, enabling the detection and identification of intrusion attempts within a system. This facilitates the timely capture and recording of detailed information on malicious intrusion behavior, providing valuable data support for subsequent analysis and countermeasures. Analyzing and evaluating the captured intrusion data by attack type allows for the determination of the progressive intensity of the intrusion behavior. This helps assess the threat and severity of the intrusion, providing security teams with quantitative reference indicators to better understand the severity of the intrusion and implement appropriate defense and response measures. Biased progressive concealment analysis of intrusion intensity data identifies whether there is a trend towards progressive concealment in intrusion behavior. This helps uncover potential covert methods and progressive attack strategies employed by attackers, leading to better identification and response to these potential threats. Compensation adjustments are made to the intrusion progressive intensity data based on the biased progressive concealment characteristics, providing more accurate intrusion intensity information. By adjusting compensation data, the threat level of intrusion behavior can be better reflected, providing more accurate data support for subsequent defense and response decisions. Analyzing attack heatmap data using intrusion asymptotic intensity compensation data and intrusion bias asymptotic concealment data enables attack behavior situational awareness. This means that it is possible to understand the development trend and changes of attack behavior in real time, identify high-risk areas and attack trends in the system, and help to take appropriate security measures and defense strategies in a timely manner to ensure system security.

[0046] Preferably, biased progressive concealment analysis of intrusion progressive intensity data includes the following steps:

[0047] Multi-level neural network learning is performed on the intrusion progressive intensity data based on the feedforward neural network to obtain intrusion progressive neural network data.

[0048] Intrusion bias data is obtained by capturing intrusion bias intrusion intensity data based on intrusion progressive neural network data.

[0049] Intrusion logic strategy deduction is performed on intrusion bias data and intrusion progressive intensity data to obtain intrusion logic deduction strategy data;

[0050] Intrusion logic deduction strategy data is used to simulate intrusion combination methods to obtain intrusion method combination simulation data.

[0051] Based on the intrusion logic deduction strategy data, intrusion method combination simulation data, and intrusion bias data, biased progressive concealment analysis is performed to obtain intrusion biased progressive concealment data.

[0052] This invention is based on multi-level neural network learning using feedforward neural networks: by using feedforward neural networks, multi-level neural network learning can be performed on intrusion asymptotic intensity data. This means that the system automatically extracts and learns relevant features and patterns from the input data (intrusion asymptotic intensity) through the training and learning process of the neural network, thereby obtaining intrusion asymptotic neural network data. This pre-trained intrusion asymptotic neural network data is then applied to the intrusion asymptotic intensity data. Through the predictive and classification capabilities of the neural network model, the intrusion bias of the intrusion asymptotic intensity data is captured and judged. This step helps determine the bias of intrusion behavior, i.e., attack trends or attacker tendencies. Based on the intrusion bias data, further analysis and deduction are performed on the intrusion asymptotic intensity data. This includes inferring the logical strategy behind the intrusion behavior, the methods, means, and strategies that the attacker may employ, as well as possible attack paths and evolution processes. In this step, intrusion logic deduction strategy data is used to simulate intrusion combination methods. This means simulating and predicting the likelihood of intrusions exhibiting a bias towards gradual, covert behavior to understand the different combinations of methods and techniques attackers might employ. In the final step, a biased gradual covert analysis is conducted, comprehensively considering intrusion logic deduction strategy data, intrusion method combination simulation data, and intrusion bias data. This analysis aims to identify and understand the gradual covert strategies and potential threats that attackers might employ, providing a more in-depth security risk assessment and effective defense measures.

[0053] Preferably, step S25 includes the following steps:

[0054] Step S251: Simulate abnormal behavior state of the computer system based on attack type behavior mapping data and attack behavior situation awareness data to obtain abnormal behavior state simulation data; wherein the abnormal behavior state simulation data includes CPU abnormal behavior state data and memory abnormal behavior state data.

[0055] Step S252: Analyze the abnormal clock frequency fluctuations in the CPU abnormal behavior status data to obtain abnormal clock frequency fluctuation data;

[0056] Step S253: Perform read rate loss simulation calculation on the memory abnormal behavior state data to obtain memory read rate loss data;

[0057] Step S254: Based on the abnormal fluctuation data of clock frequency, evaluate the memory access failure probability of memory read rate loss data to obtain memory access failure probability data;

[0058] Step S255: Perform system anomaly feedback analysis on the computer system based on clock frequency abnormal fluctuation data, memory read rate loss data, and memory access failure probability data to obtain system anomaly feedback data.

[0059] This invention simulates abnormal behavior states of a computer system to obtain abnormal behavior state simulation data. This data includes CPU and memory abnormal behavior state data, which helps simulate and understand the behavior patterns and states of a computer system under attack or abnormal conditions. It provides a detailed description and visualization of abnormal computer system behavior, aiding in further analysis and identification of potential threats. It helps detect and identify abnormal fluctuations in CPU clock frequency, potentially indicating an attack or malfunction, and provides a quantitative assessment of CPU abnormal behavior, supporting subsequent analysis and response. It helps assess the extent of memory read rate loss, potentially indicating a potential attack or malfunction, and provides a quantitative assessment of memory abnormal behavior, supporting subsequent analysis and response. It helps assess the potential risks of memory access failures and provides a quantitative assessment of related probabilities, understanding the likelihood of memory access failures. This helps optimize system configuration and enhance system stability. It provides a comprehensive analysis and assessment of abnormal computer system conditions, helping to alert system administrators or security teams to potential attacks or malfunctions in the computer system and enabling timely and appropriate response measures.

[0060] Preferably, the memory access failure probability assessment based on memory read rate loss data includes the following:

[0061] The abnormal frequency fluctuation data is decomposed into abnormal frequency fluctuation trends to obtain abnormal frequency fluctuation trend decomposition data.

[0062] Based on the abnormal frequency fluctuation trend decomposition data, memory read rate loss state space modeling is performed on the memory read rate loss data to obtain the read rate loss state space model.

[0063] Based on the read rate loss state space model, nonlinear relationships are captured in the memory read rate loss data to obtain nonlinear relationship data of read loss.

[0064] Based on the nonlinear relationship data of the loss, nonparametric statistical probability calculation is performed on the memory read rate loss data to obtain the read rate nonparametric probability data;

[0065] The memory access failure probability is evaluated based on the nonlinear relationship data of read loss and the nonparametric probability data of read rate, and the memory access failure probability data is obtained.

[0066] This invention first decomposes the abnormal frequency fluctuation data into anomaly frequency fluctuation trends. This step aims to detect and extract abnormal frequency fluctuation trends for subsequent analysis. Using the decomposed data, state-space modeling is performed on the memory read rate loss data. This process maps the memory read rate loss data into a state-space model for further analysis and processing. Based on the read rate loss state-space model, the nonlinear relationships of the memory read rate loss data are captured. The purpose of this step is to represent the nonlinear relationships as a data model to better understand and analyze the characteristics of memory read rate loss. Based on the nonlinear relationship data of read loss, nonparametric statistical methods are used to calculate the probability of memory read rate loss. This process can estimate the probability distribution of memory read rate loss based on the actual data distribution to further assess the likelihood of memory access failures. Combining the nonlinear relationship data of read loss and the nonparametric probability data of read rate, a memory access failure probability assessment is performed. This step calculates the probability of memory access failures based on the nonlinear relationships and probability data, used to determine the likelihood and risk of memory access failures.

[0067] Preferably, step S3 includes the following steps:

[0068] Step S31: Design intrusion detection rules based on attack behavior situational awareness data and system anomaly feedback data to obtain intrusion detection rule data;

[0069] Step S32: Encode the intrusion detection rule data to obtain intrusion detection rule encoded data;

[0070] Step S33: Based on the random forest algorithm, construct a computer system security verification model using intrusion detection rule encoding data, attack behavior situational awareness data, and system anomaly feedback data to obtain the system security verification model.

[0071] This invention combines intrusion detection rule encoding data, attack behavior situational awareness data, and system anomaly feedback data to obtain comprehensive and multi-dimensional security verification data. Such integrated data analysis can improve system security and identify and predict potential threats and vulnerabilities. The random forest algorithm, a machine learning algorithm, performs classification, regression, and anomaly detection by ensembling multiple decision trees. Building a computer system security verification model based on the random forest algorithm leverages its highly scalable, efficient, and robust characteristics to provide accurate prediction and evaluation results. By constructing this system security verification model, a comprehensive security assessment and verification of the system can be performed. This model can identify potential security vulnerabilities and threats and provide corresponding defense suggestions and measures. This enhances system security and effectively responds to various attacks and intrusions.

[0072] The beneficial effects of this invention are as follows: By extracting attack interception data from a computer system through a computer firewall, malicious attacks can be identified in a timely manner, thereby protecting the security of the computer system. Attack type identification based on firewall attack interception data can determine the specific attack type, helping the security team better understand and respond to different types of attacks. Attack behavior pattern analysis of attack type data can reveal the attacker's behavioral patterns and methods, helping to predict future possible attacks and improve system security and defense capabilities. Attack behavior mapping processing of attack type behavior pattern data can associate attack behavior with specific systems or resources, helping to better understand the attacker's goals and intentions. Attack behavior situational awareness based on attack type behavior mapping data can monitor and analyze the attacker's behavioral dynamics in real time, quickly identify and respond to potential threats, and improve system security. System anomaly feedback analysis of the computer system can be performed based on attack type behavior mapping data and attack behavior situational awareness data. This invention can identify anomalies in the system, helping to promptly identify and fix potential vulnerabilities or security risks, improving system stability and reliability. Designing intrusion detection rules based on system anomaly feedback data can extract patterns and rules related to known intrusion behaviors, helping the system to promptly detect and intercept intrusion attempts, improving system security. Building a computer system security verification model based on the random forest algorithm for intrusion detection rule data can establish an effective model for judging and evaluating the system's security status and risk level. System security risk detection based on the system security verification model can assess existing security risks and potential threats, helping system administrators or security teams to take timely measures to address and prevent risks, improving system security. Analyzing and processing system security risk data can identify configuration vulnerabilities, weaknesses, or insecure settings in the system, and provide corresponding configuration optimization suggestions, helping system administrators improve system configuration and reduce the probability and impact of system attacks. Therefore, this invention provides a computer system security model verification method that solves the problem of traditional computer system security model verification methods failing to accurately recreate attack scenarios and potential attack characteristics, thus hindering accurate model verification of computer systems. It accurately recreates attack scenarios and potential attack characteristics, enabling accurate model verification of computer systems. Attached Figure Description

[0073] Figure 1 This is a flowchart illustrating the steps of a computer system security model verification method.

[0074] Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S2.

[0075] Figure 3for Figure 2 A detailed flowchart illustrating the implementation steps of step S24.

[0076] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0077] The technical method of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention.

[0078] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0079] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0080] To achieve the above objectives, please refer to Figures 1 to 3 A method for verifying a computer system security model, the method comprising the following steps:

[0081] Step S1: Extract attack interception data from the computer system through the computer firewall to obtain firewall attack interception data; identify attack types based on the firewall attack interception data to obtain attack type data; analyze attack behavior patterns in the attack type data to obtain attack type behavior pattern data.

[0082] Step S2: Perform attack behavior mapping processing on the attack type behavior pattern data to obtain attack type behavior mapping data; perform attack behavior situation awareness based on the attack type behavior mapping data to obtain attack behavior situation awareness data; perform system anomaly feedback analysis on the computer system based on the attack type behavior mapping data and the attack behavior situation awareness data to obtain system anomaly feedback data.

[0083] Step S3: Design intrusion detection rules based on system anomaly feedback data to obtain intrusion detection rule data; construct a computer system security verification model based on the intrusion detection rule data using the random forest algorithm to obtain the system security verification model;

[0084] Step S4: Perform system security risk detection based on the system security verification model to obtain system security risk data; optimize system configuration based on system security risk data to obtain system configuration optimization data.

[0085] In this embodiment of the invention, reference Figure 1 The above is a flowchart illustrating the steps of a computer system security model verification method according to the present invention. In this example, the computer system security model verification method includes the following steps:

[0086] Step S1: Extract attack interception data from the computer system through the computer firewall to obtain firewall attack interception data; identify attack types based on the firewall attack interception data to obtain attack type data; analyze attack behavior patterns in the attack type data to obtain attack type behavior pattern data.

[0087] In this embodiment of the invention, a computer firewall is set up and the attack interception function is enabled. The firewall logs are monitored, the firewall attack interception data is extracted, the firewall attack interception data is analyzed and processed, the attack type is identified, and the result is recorded as attack type data. The attack type data is further analyzed, and the attack type behavior pattern data is obtained by pattern recognition and analysis of the attack behavior.

[0088] Step S2: Perform attack behavior mapping processing on the attack type behavior pattern data to obtain attack type behavior mapping data; perform attack behavior situation awareness based on the attack type behavior mapping data to obtain attack behavior situation awareness data; perform system anomaly feedback analysis on the computer system based on the attack type behavior mapping data and the attack behavior situation awareness data to obtain system anomaly feedback data.

[0089] In this embodiment of the invention, attack type behavior pattern data is processed to perform attack behavior mapping. Based on the attack type behavior mapping data, attack behavior situational awareness is conducted on the computer system. This includes real-time monitoring, analysis, and recording of known attack behaviors and abnormal behaviors. The attack behavior situational awareness data is combined with the attack type behavior mapping data to perform system anomaly feedback analysis on the computer system. This includes the identification of abnormal behaviors, cause analysis, and potential impact assessment. The system anomaly feedback data is analyzed to obtain system anomaly feedback results, which are used to further improve the system's security and defense capabilities.

[0090] Step S3: Design intrusion detection rules based on system anomaly feedback data to obtain intrusion detection rule data; construct a computer system security verification model based on the intrusion detection rule data using the random forest algorithm to obtain the system security verification model;

[0091] In this embodiment of the invention, intrusion detection rules are designed based on system anomaly feedback data. This includes formulating corresponding rules to detect potential intrusion behaviors based on known attack and anomaly patterns, such as specific network traffic patterns and abnormal file access. Intrusion detection rule data is generated based on the designed rules. These rules describe various attack and anomaly patterns. A computer system security verification model is constructed using the intrusion detection rule data based on the random forest algorithm. In this step, the random forest algorithm uses the intrusion detection rule data to construct a classifier model composed of multiple decision trees. This model is used to determine whether a new sample belongs to an attack or an anomaly, thus completing the construction of the system security verification model. This model can be used to classify new system behaviors and determine whether a security threat exists.

[0092] Step S4: Perform system security risk detection based on the system security verification model to obtain system security risk data; optimize system configuration based on system security risk data to obtain system configuration optimization data.

[0093] In this embodiment of the invention, system security risk detection is performed based on a system security verification model. The system behavior to be detected is input into the system security verification model, which classifies and assesses the risk of the behavior to determine whether it involves potential security risks. Based on the results of the system security risk detection, system security risk data is generated. This data describes detailed information about the security risks and potential threats present in the system. The system security risk data is analyzed and evaluated, and system configuration optimization is performed. Depending on the degree and scope of the security risks, corresponding adjustment and optimization measures are taken, including changing system configuration, updating security patches, and enhancing access control, to complete the system configuration optimization and obtain system configuration optimization data. This data records the configuration optimization operations performed and their results to improve the system's security and defense capabilities.

[0094] This invention utilizes computer firewalls to extract attack interception data from computer systems, enabling timely identification of malicious attacks and protecting system security. Attack type identification based on firewall attack interception data determines specific attack types, helping security teams better understand and respond to different attack types. Attack behavior pattern analysis of attack type data reveals attacker behavior patterns and methods, aiding in predicting future attacks and improving system security and defense capabilities. Attack behavior mapping of attack type behavior pattern data associates attack behaviors with specific systems or resources, facilitating a better understanding of attacker goals and intentions. Attack behavior situational awareness based on attack type behavior mapping data allows for real-time monitoring and analysis of attacker behavior dynamics, rapid identification and response to potential threats, and improved system security. Furthermore, system anomaly feedback analysis of the computer system based on attack type behavior mapping data and attack behavior situational awareness data can identify… This invention addresses anomalies in the system, helping to promptly identify and fix potential vulnerabilities or security risks, improving system stability and reliability. Intrusion detection rule design based on system anomaly feedback data can extract patterns and rules related to known intrusion behaviors, helping the system to promptly detect and intercept intrusion attempts, thus improving system security. Building a computer system security verification model based on the random forest algorithm for intrusion detection rule data can establish an effective model for judging and evaluating the system's security status and risk level. System security risk detection based on the system security verification model can assess existing security risks and potential threats, helping system administrators or security teams to take timely measures to address and prevent risks, improving system security. Analyzing and processing system security risk data can identify configuration vulnerabilities, weaknesses, or insecure settings in the system, and provide corresponding configuration optimization suggestions, helping system administrators improve system configuration and reduce the probability and impact of system attacks. Therefore, this invention provides a computer system security model verification method that solves the problem of traditional computer system security model verification methods failing to accurately recreate attack scenarios and potential attack characteristics, thus hindering accurate model verification of computer systems. This invention accurately recreates attack scenarios and potential attack characteristics, enabling accurate model verification of computer systems.

[0095] Preferably, step S1 includes the following steps:

[0096] Step S11: Extract attack interception data from the computer system through the computer firewall to obtain firewall attack interception data;

[0097] Step S12: Analyze the attack targets of the firewall attack interception data to obtain the attack target data;

[0098] Step S13: Identify the attack type based on the firewall attack interception data and the attack target data to obtain attack type data;

[0099] Step S14: Perform attack behavior pattern analysis on the attack type data to obtain attack type behavior pattern data.

[0100] In this embodiment of the invention, appropriate firewall rules are set and firewall functions are enabled to monitor and intercept network traffic of the computer system. Firewall logging is configured to ensure that attack interception data is recorded and saved. Firewall logs are monitored, and log entries containing attack information are extracted to form firewall attack interception data. This data may include key information such as the attack source IP address, the attack target IP address, the attack protocol, and the attack type. The attack target IP address or domain name in the firewall attack interception data is analyzed. Based on the characteristics of the attack target, such as related business systems, important resources, or sensitive data, the attack target is classified and summarized to obtain attack target data. Key information, such as attack protocols and attack behavior characteristics, is extracted from the firewall attack interception data. Based on a known attack type library or attack behavior pattern, data mining or machine learning methods are used to classify and identify attacks to obtain attack type data. The collected attack type data is analyzed to extract attack behavior patterns and rules. Statistical analysis, data mining, or machine learning techniques are used to perform pattern recognition and analysis on the attack type data to obtain attack type behavior pattern data. This behavior pattern data can help understand the attacker's strategies, attack methods, and potential threats, thus providing a foundation for subsequent intrusion detection rule design and system security risk detection.

[0101] This invention's firewall can monitor and record network traffic. By extracting firewall attack interception data, it can detect potential malicious attacks, such as network intrusions, port scans, and denial-of-service attacks. The firewall attack interception data provides real-time attack information, enabling security teams to respond quickly and take appropriate countermeasures, such as blocking the attack source IP address and modifying access control rules, thereby reducing the harm caused by malicious attacks. The firewall attack interception data records detailed information on network activity and attack events. Auditing and analyzing this data helps security teams understand attack trends, detect system vulnerabilities, and improve security measures and strategies accordingly. Attack target analysis can determine the attacker's goals, attack methods, and attack techniques, helping to identify weaknesses and sensitive resources in the system. By analyzing attack target data, the threat level and potential losses of different attack behaviors can be assessed, providing guidance for subsequent security defenses and optimizations. Attack target analysis may reveal unknown vulnerabilities or security weaknesses in the system, allowing the security team to fix them in a timely manner. These vulnerabilities enhance system security. By analyzing and correlating firewall attack interception data and attack target data, different types of attacks can be identified, such as network scanning, malware propagation, and authentication attacks. This helps to accurately understand the attacker's strategies and methods. Based on attack type data, security teams can adjust and optimize defense strategies such as firewalls and intrusion detection systems, strengthening the interception and defense against known attack types, thereby improving system security and protection capabilities. Attack type behavior pattern data can reveal the behavioral patterns and habits of malicious attackers. By analyzing this data, possible future attack behaviors can be predicted, and corresponding preventive measures can be taken. Attack type behavior pattern data helps identify abnormal behaviors in the system, such as a large number of login failures and abnormal data transmission. It can promptly detect attack behaviors or potential threats and take appropriate response measures. By analyzing attack type behavior pattern data, existing security protection strategies and measures can be optimized, improving the system's ability to detect and block known and unknown attack behaviors.

[0102] Preferably, step S14 includes the following steps:

[0103] Step S141: Extract the interception logs from the firewall attack interception data to obtain the interception log data;

[0104] Step S142: Perform attack type log matching on the attack type data based on the interception log data to obtain attack type log matching data;

[0105] Step S143: Extract network traffic of attack type from attack type log matching data to obtain attack type network traffic data;

[0106] Step S144: Perform instantaneous geometric growth analysis on network traffic data based on attack type to obtain instantaneous geometric growth data of traffic;

[0107] Step S145: Identify potential growth patterns in the instantaneous geometric growth data of traffic to obtain potential growth pattern data of traffic.

[0108] Step S146: Analyze the attack behavior patterns of the attack type data based on the potential growth pattern data of traffic to obtain attack type behavior pattern data.

[0109] In this embodiment of the invention, interception logs are extracted from firewall attack interception data. These logs are parsed according to log specifications and formats. The extracted interception logs contain key attack information, such as timestamps, source IP addresses, target IP addresses, and attack types. The interception log data is analyzed, and interception logs are matched with corresponding attack types according to a predefined attack type log database. The resulting attack type log matching data contains attack type information corresponding to each interception log. Based on the source and target IP addresses in the attack type log matching data, corresponding network traffic data is filtered out. The extracted attack type network traffic data includes characteristics of the relevant network traffic at the time of the attack, such as packet size... By analyzing network traffic data for different attack types, including size and protocol type, and calculating the changes in traffic over instantaneous time periods, the growth trend of network traffic during an attack can be obtained by calculating the instantaneous geometric growth of traffic. Analysis and pattern recognition of this instantaneous geometric growth data aims to identify potential traffic growth patterns. Through statistical analysis, machine learning, and other techniques, the patterns of traffic growth during an attack, such as linear or exponential growth, can be determined. Analyzing potential traffic growth pattern data and correlating it with attack type data helps identify the behavioral patterns of different attack types. By analyzing these behavioral patterns, a better understanding of attackers' habits and attack methods can be achieved, providing a reference for defense and response.

[0110] This invention extracts detailed information about attack behavior from intercept logs, including attack time, source IP address, and target, providing data support for subsequent analysis and processing. Interception log data can track the source and destination of attack activities, helping to determine attacker behavior patterns and strategies. The data records provided by the intercept logs help security teams audit security incidents and ensure compliance with relevant requirements. By matching and analyzing attack information in the intercept log data, attack behaviors can be associated with corresponding attack types, further refining and classifying attack data. Obtaining attack type log matching data improves the accuracy and reliability of attack type data, thus providing a more precise understanding of attack characteristics and patterns. By extracting network traffic from attack type log matching data, it can be associated with specific attack types, further analyzing and understanding attack behavior and characteristics. Attack type network traffic data provides an overall overview of attack activities, including the quantity, frequency, source, and target of attack traffic, providing a foundation for further analysis and modeling. The data on instantaneous geometric growth in traffic is used to identify abnormal network traffic, i.e., traffic growth that is faster than normal, thereby discovering potential malicious attacks or network anomalies. By analyzing the magnitude and trend of traffic growth, the impact of attack behavior on network performance and services can be assessed, and security teams can be helped to determine the threat level of malicious attacks. Data on potential traffic growth patterns is used to identify common traffic growth patterns in malicious attacks, such as the attack's initiation process, duration, and peak periods, which helps to build attack behavior models. By analyzing potential traffic growth patterns, security teams can predict the development trend of attack behavior and take corresponding security strategies and measures for response and defense. By analyzing the relationship between potential traffic growth pattern data and attack type data, we can gain a deeper understanding of the patterns and evolution of attack behavior, such as the attack's initiation stage, duration, and changing trends. Attack type behavior pattern data provides security teams with a basis for optimizing security strategies, enabling them to develop more effective security measures and defense strategies for specific attack types.

[0111] Preferably, the analysis of instantaneous geometric growth in network traffic based on attack type data includes the following steps:

[0112] Time-series analysis is performed on network traffic data of attack types to obtain time-series data of network traffic of attack types.

[0113] The network traffic time series data of attack type is smoothed by performing time series smoothing on the network traffic time series data of attack type to obtain traffic time series smoothed data.

[0114] Based on the traffic time-series smoothing data, an instantaneous traffic growth Gaussian process analysis was performed on the network traffic data of attack types to obtain instantaneous traffic growth Gaussian process data;

[0115] Based on the Gaussian process data of instantaneous flow growth, the flow time-series smoothed data is processed by flow time-series segmentation to obtain flow time-series segmented data.

[0116] Based on the Gaussian process data of instantaneous flow growth, the geometric series growth of flow time-series segmented data is calculated to obtain segmented flow geometric growth data;

[0117] Based on the Gaussian process data of instantaneous flow rate growth and the segmented geometrical growth data of flow rate, an analysis of the instantaneous geometrical growth of flow rate was conducted to obtain the instantaneous geometrical growth data of flow rate.

[0118] In this embodiment of the invention, attack-type network traffic data is sorted chronologically to form time-series data. Time intervals are determined, and the time-series data is divided into continuous time periods. Based on a sliding window or other smoothing methods, the attack-type network traffic time-series data is smoothed. Smoothing reduces noise and abrupt changes, resulting in smoother traffic time-series data. Based on statistical analysis and model fitting techniques, Gaussian process analysis is performed on the smoothed traffic time-series data. Gaussian process analysis can be used to describe the growth trend and change characteristics of traffic, obtaining instantaneous traffic growth Gaussian process data. Based on the inflection points and boundaries in the instantaneous traffic growth Gaussian process data, the smoothed traffic time-series data is further refined. Segmentation processing is performed, which divides the time-series smoothed data into multiple traffic time-series segments with different characteristics based on the traffic change trend. Based on parameters such as the mean and variance in the instantaneous traffic growth Gaussian process data, the growth geometric series of each traffic time-series segment is calculated. The growth geometric series describes the growth trend and magnitude of traffic within each segment, yielding segmented traffic geometric growth data. Combining the instantaneous traffic growth Gaussian process data and the segmented traffic geometric growth data, instantaneous geometric multiple growth analysis is performed on the traffic time-series data. This instantaneous geometric multiple growth analysis is used to calculate the traffic growth multiple at each time point, yielding instantaneous geometric multiple growth data.

[0119] This invention, through time-series analysis of attack-type network traffic data, can obtain the changing trends and periodicity of attack-type network traffic. This helps to better understand the dynamic behavior of traffic. By smoothing the time-series data of attack-type network traffic, noise and outliers can be removed, making the data more stable and reliable. By performing Gaussian process analysis on the smoothed time-series traffic data, the probability distribution information of traffic growth can be obtained. This helps to understand the growth of attack-type network traffic and identify abnormal growth events. Based on the Gaussian process data of instantaneous traffic growth, segmenting the smoothed time-series traffic data can more accurately describe the characteristics and trends of traffic changes. Based on the Gaussian process data of instantaneous traffic growth, geometric progression calculation of the growth of segmented time-series traffic can obtain the geometric growth rate of traffic in each time period. This helps to quantify the growth trend and speed of traffic. Based on the Gaussian process data of instantaneous traffic growth and segmented geometric growth data, instantaneous geometric multiple growth analysis of traffic can be performed. This can provide more comprehensive traffic growth indicators, helping analysts understand the growth magnitude and speed of attack-type network traffic.

[0120] Preferably, step S2 includes the following steps:

[0121] Step S21: Perform attack behavior mapping processing on the attack type behavior pattern data based on the firewall attack interception data to obtain attack type behavior mapping data;

[0122] Step S22: Perform attack behavior frequency statistics on the attack type behavior mapping data to obtain attack behavior frequency data;

[0123] Step S23: Analyze the attack heat map distribution of the computer system based on the attack type behavior mapping data and the attack behavior frequency data to obtain attack heat map data;

[0124] Step S24: Perform attack behavior situation awareness on the attack heat map data based on the attack type behavior mapping data to obtain attack behavior situation awareness data;

[0125] Step S25: Perform system anomaly feedback analysis on the computer system based on attack type behavior mapping data and attack behavior situational awareness data to obtain system anomaly feedback data.

[0126] As an example of the present invention, reference is made to... Figure 2 As shown, step S2 in this example includes:

[0127] Step S21: Perform attack behavior mapping processing on the attack type behavior pattern data based on the firewall attack interception data to obtain attack type behavior mapping data;

[0128] In this embodiment of the invention, firewall attack interception data is collected. This data may include attack type, attack source, and attack target. The firewall attack interception data is analyzed and processed to classify and summarize different attack behaviors, forming attack type behavior pattern data. Based on the attack type behavior pattern data, mapping processing is performed to map each attack type to the corresponding attack behavior.

[0129] Step S22: Perform attack behavior frequency statistics on the attack type behavior mapping data to obtain attack behavior frequency data;

[0130] In this embodiment of the invention, by statistically analyzing the attack type behavior mapping data, the frequency of occurrence of each attack behavior is calculated. The statistics may involve setting a time window to determine the calculation range of the frequency, which may be hourly, daily, or other time periods. After the statistics are obtained, the attack behavior frequency data is obtained, and the number of times or frequency of occurrence of each attack behavior within a given time range is recorded.

[0131] Step S23: Analyze the attack heat map distribution of the computer system based on the attack type behavior mapping data and the attack behavior frequency data to obtain attack heat map data;

[0132] In this embodiment of the invention, attack type behavior mapping data and attack behavior frequency data are combined to analyze the attack occurrence in different areas of the computer system. By analyzing the frequency of attack types in different areas, attack heat map data is obtained, which shows the attack density in different areas of the computer system.

[0133] Step S24: Perform attack behavior situation awareness on the attack heat map data based on the attack type behavior mapping data to obtain attack behavior situation awareness data;

[0134] In this embodiment of the invention, attack behavior situation in different regions is analyzed and evaluated based on attack type behavior mapping data, including attack type, attack frequency and attack intensity. Spatial distribution information in attack heat map data is used to determine the situational awareness of attack behavior, such as the concentration and duration of attack activities in a certain region. The obtained attack behavior situational awareness data can be used to predict and adjust the system's security strategies and protection measures.

[0135] Step S25: Perform system anomaly feedback analysis on the computer system based on attack type behavior mapping data and attack behavior situational awareness data to obtain system anomaly feedback data.

[0136] In this embodiment of the invention, attack type behavior mapping data and attack behavior situational awareness data are combined to detect abnormal behaviors and attack events in the computer system, and analyze abnormal feedback data, including the type, frequency and severity of abnormal events, as well as possible attack sources and targets. The abnormal feedback data can help system administrators or security teams take appropriate response measures, fix vulnerabilities and enhance system security.

[0137] This invention maps attack type behavior pattern data to specific behavioral patterns based on firewall attack interception data. This helps to more accurately understand the behavioral characteristics of different types of attacks, providing a useful data foundation for subsequent analysis. Frequency statistics of the attack type behavior mapping data can calculate the frequency of different attack behaviors. This helps to identify which attack behaviors are more common, allowing for more targeted responses and prevention. Based on the attack type behavior mapping data and attack behavior frequency data, attack heat map distribution analysis of computer systems can identify the areas or parts of the system most heavily attacked. This can be used to identify weak points and vulnerable areas in the system, facilitating focused protection and defense. Analyzing attack heat map data based on the attack type behavior mapping data enables attack behavior situational awareness. This means timely understanding of the current attack behavior situation and trends, identifying potential attack threats, and taking timely defensive and response measures. Based on the attack type behavior mapping data and attack behavior situational awareness data, system anomaly feedback analysis can be performed. This helps to identify abnormal behaviors and events in the computer system and provide corresponding feedback information. By analyzing abnormal feedback data, abnormal situations in the system can be quickly identified, providing a basis for further safe operations and interventions.

[0138] Preferably, step S24 includes the following steps:

[0139] Step S241: Capture intrusion attempts on the attack heat map data based on the attack type behavior mapping data to obtain attack type intrusion capture data;

[0140] Step S242: Perform an intrusion progressive strength assessment on the attack type intrusion capture data to obtain intrusion progressive strength data;

[0141] Step S243: Perform biased progressive concealment analysis on the progressive intrusion intensity data to obtain biased progressive intrusion concealment data;

[0142] Step S244: Perform intrusion intensity compensation on the intrusion progressive intensity data based on the intrusion bias progressive concealment data to obtain intrusion progressive intensity compensation data.

[0143] Step S245: Based on the intrusion progressive intensity compensation data and the intrusion bias progressive concealment data, perform attack behavior situational awareness on the attack heat map data to obtain attack behavior situational awareness data.

[0144] As an example of the present invention, reference is made to... Figure 3 As shown, in this example, step S24 includes:

[0145] Step S241: Capture intrusion attempts on the attack heat map data based on the attack type behavior mapping data to obtain attack type intrusion capture data;

[0146] In this embodiment of the invention, attack type behavior mapping data is combined to focus on potential intrusion attempts appearing in attack heat map data, monitor and capture possible intrusion behaviors in the system, such as unauthorized access, malicious scripts or network scanning, record and analyze intrusion attempt data, including intrusion type, attack source, attack target and other information, to form attack type intrusion capture data.

[0147] Step S242: Perform an intrusion progressive strength assessment on the attack type intrusion capture data to obtain intrusion progressive strength data;

[0148] In this embodiment of the invention, attack type intrusion capture data is analyzed, and the progressive strength of each intrusion behavior is evaluated. The progressive strength evaluation can take into account factors such as the complexity of the attack, the time consumption, and the required permissions. After the intrusion progressive strength evaluation is performed, the strength index of each intrusion behavior is obtained, forming intrusion progressive strength data.

[0149] Step S243: Perform biased progressive concealment analysis on the progressive intensity data of intrusion to obtain biased progressive concealment data of intrusion.

[0150] In this embodiment of the invention, intrusion progressive intensity data is used to analyze whether intrusion behavior exhibits a tendency toward progressiveness. Progressiveness refers to the trend of intrusion behavior gradually increasing or changing direction. By analyzing the trends and change patterns in the intrusion progressive intensity data, intrusion progressiveness concealment data is obtained.

[0151] Step S244: Perform intrusion intensity compensation on the intrusion progressive intensity data based on the intrusion bias progressive concealment data to obtain intrusion progressive intensity compensation data.

[0152] In this embodiment of the invention, the intrusion progressive intensity data is compensated based on the intrusion bias progressive concealment data to make up for the errors or deviations given by bias and progressiveness. The intrusion progressive intensity data is adjusted according to the information in the intrusion bias progressive concealment data to ensure that the evaluation results are more accurate and reliable. The obtained intrusion progressive intensity compensation data can better describe the changing trend and intensity of intrusion behavior and provide more accurate analysis basis.

[0153] Step S245: Based on the intrusion progressive intensity compensation data and the intrusion bias progressive concealment data, perform attack behavior situational awareness on the attack heat map data to obtain attack behavior situational awareness data.

[0154] In this embodiment of the invention, the attack heat map data is further analyzed by combining the intrusion progressive intensity compensation data and the intrusion bias progressive concealment data to obtain the situational awareness of the attack behavior. By considering the correlation between progressive intensity, bias and attack area, the trend, distribution and possible evolution of the attack behavior are identified. The obtained attack behavior situational awareness data can help system administrators or security teams better understand and respond to current and potential attack threats.

[0155] This invention analyzes and captures attack heatmap data based on attack type behavior mapping data, enabling the detection and identification of intrusion attempts within a system. This facilitates the timely capture and recording of detailed information on malicious intrusion behavior, providing valuable data support for subsequent analysis and countermeasures. Analyzing and evaluating the captured intrusion data by attack type allows for the determination of the progressive intensity of the intrusion behavior. This helps assess the threat and severity of the intrusion, providing security teams with quantitative reference indicators to better understand the severity of the intrusion and implement appropriate defense and response measures. Biased progressive concealment analysis of intrusion intensity data identifies whether there is a trend towards progressive concealment in intrusion behavior. This helps uncover potential covert methods and progressive attack strategies employed by attackers, leading to better identification and response to these potential threats. Compensation adjustments are made to the intrusion progressive intensity data based on the biased progressive concealment characteristics, providing more accurate intrusion intensity information. By adjusting compensation data, the threat level of intrusion behavior can be better reflected, providing more accurate data support for subsequent defense and response decisions. Analyzing attack heatmap data using intrusion asymptotic intensity compensation data and intrusion bias asymptotic concealment data enables attack behavior situational awareness. This means that it is possible to understand the development trend and changes of attack behavior in real time, identify high-risk areas and attack trends in the system, and help to take appropriate security measures and defense strategies in a timely manner to ensure system security.

[0156] Preferably, biased progressive concealment analysis of intrusion progressive intensity data includes the following steps:

[0157] Multi-level neural network learning is performed on the intrusion progressive intensity data based on the feedforward neural network to obtain intrusion progressive neural network data.

[0158] Intrusion bias data is obtained by capturing intrusion bias intrusion intensity data based on intrusion progressive neural network data.

[0159] Intrusion logic strategy deduction is performed on intrusion bias data and intrusion progressive intensity data to obtain intrusion logic deduction strategy data;

[0160] Intrusion logic deduction strategy data is used to simulate intrusion combination methods to obtain intrusion method combination simulation data.

[0161] Based on the intrusion logic deduction strategy data, intrusion method combination simulation data, and intrusion bias data, biased progressive concealment analysis is performed to obtain intrusion biased progressive concealment data.

[0162] In this embodiment of the invention, based on a feedforward neural network model, intrusion asymptotic intensity data is used as input to train the neural network model. During training, the weights and biases of the neural network are adjusted using a backpropagation algorithm to minimize the error between the predicted and actual values. After training, a learned neural network model is obtained, namely, intrusion asymptotic neural network data. Intrusion bias is captured based on the intrusion asymptotic intensity data. The intrusion asymptotic intensity data is analyzed and judged to capture the bias characteristics of intrusion behaviors, resulting in intrusion bias data. This data reflects the degree of bias of each intrusion behavior and indicates the specific direction in which the intrusion behavior tends. Based on the intrusion bias data, the bias characteristics and trends of each intrusion behavior are analyzed. Inference algorithms or logical deduction techniques are used to determine the intrusion asymptotic intensity data. By combining intrusion bias data with intrusion logic strategy data, possible intrusion logic strategies are deduced, resulting in intrusion logic deduction strategy data. This data describes the correlation between intrusion behaviors and possible execution strategies. Based on this intrusion logic deduction strategy data, different combinations of intrusion methods are simulated, considering the synergistic effects and sequence between intrusion behaviors. Possible combinations of intrusion methods are simulated, resulting in intrusion method combination simulation data. This data shows the possible combinations of intrusion behaviors and their execution sequence. By combining the intrusion logic deduction strategy data, intrusion method combination simulation data, and intrusion bias data, comprehensive analysis and calculations are performed to identify the gradual and biased characteristics of intrusion behaviors, as well as the possible covert methods and strategies employed. This yields intrusion bias gradual covert data, which reveals the hidden trends and gradual changes in intrusion behaviors, contributing to further security analysis and defense strategy formulation.

[0163] This invention is based on multi-level neural network learning using feedforward neural networks. By using feedforward neural networks, multi-level neural network learning can be performed on intrusion asymptotic intensity data. This means that the system automatically extracts and learns relevant features and patterns from the input data (intrusion asymptotic intensity) through the training and learning process of the neural network, thereby obtaining intrusion asymptotic neural network data. This pre-trained intrusion asymptotic neural network data is then applied to the intrusion asymptotic intensity data. Through the predictive and classification capabilities of the neural network model, intrusion bias is captured and judged in the intrusion asymptotic intensity data. This step helps determine the bias of intrusion behavior, i.e., attack trends or attacker tendencies. Based on the intrusion bias data, further analysis and deduction are performed on the intrusion asymptotic intensity data. This includes inferring the logical strategies behind the intrusion behavior, the methods, means, and strategies that attackers may employ, as well as possible attack paths and evolution processes. In this step, intrusion logic deduction strategy data is used to simulate intrusion combination methods. This means simulating and predicting the likelihood of intrusions exhibiting a bias towards gradual, covert behavior to understand the different combinations of methods and techniques attackers might employ. In the final step, a biased gradual covert analysis is conducted, comprehensively considering intrusion logic deduction strategy data, intrusion method combination simulation data, and intrusion bias data. This analysis aims to identify and understand the gradual covert strategies and potential threats that attackers might employ, providing a more in-depth security risk assessment and effective defense measures.

[0164] Preferably, step S25 includes the following steps:

[0165] Step S251: Simulate abnormal behavior state of the computer system based on attack type behavior mapping data and attack behavior situation awareness data to obtain abnormal behavior state simulation data; wherein the abnormal behavior state simulation data includes CPU abnormal behavior state data and memory abnormal behavior state data.

[0166] Step S252: Analyze the abnormal clock frequency fluctuations in the CPU abnormal behavior status data to obtain abnormal clock frequency fluctuation data;

[0167] Step S253: Perform read rate loss simulation calculation on the memory abnormal behavior state data to obtain memory read rate loss data;

[0168] Step S254: Based on the abnormal fluctuation data of clock frequency, evaluate the memory access failure probability of memory read rate loss data to obtain memory access failure probability data;

[0169] Step S255: Perform system anomaly feedback analysis on the computer system based on clock frequency abnormal fluctuation data, memory read rate loss data, and memory access failure probability data to obtain system anomaly feedback data.

[0170] In this embodiment of the invention, attack type behavior mapping data is collected and analyzed. This data indicates abnormal system behavior caused by different attack types. Attack behavior situational awareness data is used to simulate the computer system's abnormal behavior state under attack, generating abnormal behavior state simulation data, including CPU abnormal behavior state data and memory abnormal behavior state data. This data reflects the abnormal performance of the system under different attack types. Analysis of CPU abnormal behavior state data focuses primarily on abnormal fluctuations in CPU clock frequency. Statistical methods, pattern recognition algorithms, or other related technologies are used to analyze the fluctuation characteristics of CPU clock frequency under abnormal states, obtaining clock frequency abnormal fluctuation data to describe the changes in CPU clock frequency under abnormal conditions. Analysis of memory abnormal behavior state data focuses on changes in memory read rate under abnormal conditions, simulating and calculating the loss of memory read rate. This calculation may use models, simulations, or experimental methods to obtain memory read rate loss data, which reflects the system's memory read rate loss under abnormal conditions. By taking the degree of rate reduction and combining clock frequency anomaly data and memory read rate loss data, a memory access failure probability assessment is performed. Based on known memory performance parameters, historical data, or models, the probability of memory access failure can be estimated, resulting in memory access failure probability data. This data describes the probability of memory access failure occurring in the system under abnormal conditions. By combining clock frequency anomaly data, memory read rate loss data, and memory access failure probability data, system anomaly feedback analysis is performed to analyze the specific performance of the system under abnormal conditions, such as performance degradation, crashes, or data inconsistencies, resulting in system anomaly feedback data. This data provides information about the expected performance and response of the system when subjected to abnormal behavior.

[0171] This invention simulates abnormal behavior states of a computer system to obtain abnormal behavior state simulation data. This data includes CPU and memory abnormal behavior state data, which helps simulate and understand the behavior patterns and states of a computer system under attack or abnormal conditions. It provides a detailed description and visualization of abnormal computer system behavior, aiding in further analysis and identification of potential threats. It helps detect and identify abnormal fluctuations in CPU clock frequency, potentially indicating an attack or malfunction, and provides a quantitative assessment of CPU abnormal behavior, supporting subsequent analysis and response. It helps assess the extent of memory read rate loss, potentially indicating a potential attack or malfunction, and provides a quantitative assessment of memory abnormal behavior, supporting subsequent analysis and response. It helps assess the potential risks of memory access failures and provides a quantitative assessment of related probabilities, understanding the likelihood of memory access failures. This helps optimize system configuration and enhance system stability. It provides a comprehensive analysis and assessment of abnormal computer system conditions, helping to alert system administrators or security teams to potential attacks or malfunctions in the computer system and enabling timely and appropriate response measures.

[0172] Preferably, the memory access failure probability assessment based on memory read rate loss data includes the following:

[0173] The abnormal frequency fluctuation data is decomposed into abnormal frequency fluctuation trends to obtain abnormal frequency fluctuation trend decomposition data.

[0174] Based on the abnormal frequency fluctuation trend decomposition data, memory read rate loss state space modeling is performed on the memory read rate loss data to obtain the read rate loss state space model.

[0175] Based on the read rate loss state space model, nonlinear relationships are captured in the memory read rate loss data to obtain nonlinear relationship data of read loss.

[0176] Based on the nonlinear relationship data of the loss, nonparametric statistical probability calculation is performed on the memory read rate loss data to obtain the read rate nonparametric probability data;

[0177] The memory access failure probability is evaluated based on the nonlinear relationship data of read loss and the nonparametric probability data of read rate, and the memory access failure probability data is obtained.

[0178] In this embodiment of the invention, for clock frequency anomaly fluctuation data, appropriate signal processing techniques are used to decompose the anomaly frequency fluctuation trend. The decomposition can employ classical time-frequency analysis methods (such as wavelet analysis, fast Fourier transform, etc.) to extract the clock frequency anomaly fluctuation trend, obtaining anomaly frequency fluctuation trend decomposition data. This data reflects the changes in different frequency components of the clock frequency anomaly. Using the anomaly frequency fluctuation trend decomposition data, a state-space model of memory read rate loss is established. By observing the relationship between the anomaly frequency fluctuation trend decomposition data and the memory read rate loss data, the state space of memory read rate loss is defined. Based on the state-space model, the memory read rate loss is modeled to capture its dynamic changes. Using the read rate loss state-space model, nonlinear relationships in the memory read rate loss data are captured. Appropriate mathematical methods or machine learning techniques are used to detect nonlinear relationships in the memory read rate loss data, extracting nonlinear relationship data to describe the complex interaction between different states of memory read rate loss. Using the read loss nonlinear relationship data, nonparametric statistical probability calculations are performed on the memory read rate loss data, using statistical methods (such as kernel density estimation, Monte Carlo methods, etc.). (e.g., Carlo simulation) to calculate the nonparametric probability density function of memory read rate loss data under different states, obtaining read rate nonparametric probability data. This data provides probability distribution information for different memory read rate loss situations. Combining read loss nonlinear relationship data and read rate nonparametric probability data, memory access failure probability assessment is performed. Based on existing memory access failure probability models or statistical methods, the probability of memory access failure is inferred, obtaining memory access failure probability data. This data describes the probability of memory access failure occurring under different memory read rate loss situations.

[0179] This invention first decomposes the abnormal frequency fluctuation data into anomaly frequency fluctuation trends. This step aims to detect and extract abnormal frequency fluctuation trends for subsequent analysis. Using the decomposed data, state-space modeling is performed on the memory read rate loss data. This process maps the memory read rate loss data into a state-space model for further analysis and processing. Based on the read rate loss state-space model, the nonlinear relationships of the memory read rate loss data are captured. The purpose of this step is to represent the nonlinear relationships as a data model to better understand and analyze the characteristics of memory read rate loss. Based on the nonlinear relationship data of read loss, nonparametric statistical methods are used to calculate the probability of memory read rate loss. This process can estimate the probability distribution of memory read rate loss based on the actual data distribution to further assess the likelihood of memory access failures. Combining the nonlinear relationship data of read loss and the nonparametric probability data of read rate, a memory access failure probability assessment is performed. This step calculates the probability of memory access failures based on the nonlinear relationships and probability data, used to determine the likelihood and risk of memory access failures.

[0180] Preferably, step S3 includes the following steps:

[0181] Step S31: Design intrusion detection rules based on attack behavior situational awareness data and system anomaly feedback data to obtain intrusion detection rule data;

[0182] Step S32: Encode the intrusion detection rule data to obtain intrusion detection rule encoded data;

[0183] Step S33: Based on the random forest algorithm, construct a computer system security verification model using intrusion detection rule encoding data, attack behavior situational awareness data, and system anomaly feedback data to obtain the system security verification model.

[0184] In this embodiment of the invention, attack behavior situational awareness data and system anomaly feedback data are collected. This data may include network traffic logs, event logs, and system performance indicators. Existing attack behaviors and system anomalies are analyzed to identify common attack patterns and abnormal behaviors. Based on the analysis results, intrusion detection rules are designed. These rules describe the characteristics of typical attack patterns and abnormal behaviors. The intrusion detection rules should be efficient and accurate, capable of detecting potential intrusion behaviors in real-time or offline. The designed intrusion detection rules are then encoded for subsequent model building and computation. Different encoding methods can be used, such as converting the rules into binary vectors or using one-hot encoding, to meet the requirements of the random forest algorithm. After encoding, the rules can be... The encoded data of intrusion detection rules is obtained, which presents a discrete representation of rule features. The encoded data of intrusion detection rules, attack behavior situational awareness data, and system anomaly feedback data are integrated into a dataset. The random forest algorithm is used to construct a security verification model for the computer system. This model can determine whether there is intrusion behavior or system anomaly based on the input data. When constructing the model, the encoded data of intrusion detection rules is used as the feature vector, and the attack behavior situational awareness data and system anomaly feedback data are used as training samples. The random forest algorithm uses decision tree ensemble to effectively handle nonlinear relationships and complex interactions between features. After construction, the system security verification model can be used to monitor and detect intrusion behavior and anomalies in the computer system in real time.

[0185] This invention combines intrusion detection rule encoding data, attack behavior situational awareness data, and system anomaly feedback data to obtain comprehensive and multi-dimensional security verification data. Such integrated data analysis can improve system security and identify and predict potential threats and vulnerabilities. The random forest algorithm, a machine learning algorithm, performs classification, regression, and anomaly detection by ensembling multiple decision trees. Building a computer system security verification model based on the random forest algorithm leverages its highly scalable, efficient, and robust characteristics to provide accurate prediction and evaluation results. By constructing this system security verification model, a comprehensive security assessment and verification of the system can be performed. This model can identify potential security vulnerabilities and threats and provide corresponding defense suggestions and measures. This enhances system security and effectively responds to various attacks and intrusions.

[0186] This invention utilizes a computer firewall to extract attack interception data from a computer system, enabling timely identification of malicious attacks and protecting the computer system's security. Attack type identification based on firewall attack interception data determines the specific attack type, helping security teams better understand and respond to different types of attacks. Attack behavior pattern analysis of attack type data reveals attacker behavior patterns and methods, aiding in predicting future potential attacks and improving system security and defense capabilities. Attack behavior mapping processing of attack type behavior pattern data associates attack behaviors with specific systems or resources, facilitating a better understanding of attacker goals and intentions. Attack behavior situational awareness based on attack type behavior mapping data allows for real-time monitoring and analysis of attacker behavior dynamics, rapid identification and response to potential threats, and improved system security. Furthermore, system anomaly feedback analysis of the computer system based on attack type behavior mapping data and attack behavior situational awareness data can identify… This invention addresses anomalies in the system, helping to promptly identify and fix potential vulnerabilities or security risks, improving system stability and reliability. Intrusion detection rule design based on system anomaly feedback data can extract patterns and rules related to known intrusion behaviors, helping the system to promptly detect and intercept intrusion attempts, thus improving system security. Building a computer system security verification model based on the random forest algorithm for intrusion detection rule data can establish an effective model for judging and evaluating the system's security status and risk level. System security risk detection based on the system security verification model can assess existing security risks and potential threats, helping system administrators or security teams to take timely measures to address and prevent risks, improving system security. Analyzing and processing system security risk data can identify configuration vulnerabilities, weaknesses, or insecure settings in the system, and provide corresponding configuration optimization suggestions, helping system administrators improve system configuration and reduce the probability and impact of system attacks. Therefore, this invention provides a computer system security model verification method that solves the problem of traditional computer system security model verification methods failing to accurately recreate attack scenarios and potential attack characteristics, thus hindering accurate model verification of computer systems. This invention accurately recreates attack scenarios and potential attack characteristics, enabling accurate model verification of computer systems.

[0187] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0188] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A computer system security model verification method, characterized in that: The following steps are involved: Step S1: extracting attack interception data of a computer system through a computer firewall to obtain firewall attack interception data; Attack type identification is performed based on firewall attack interception data to obtain attack type data; attack behavior pattern analysis is performed on the attack type data to obtain attack type behavior pattern data; Step S2: performing attack behavior mapping processing on the attack type behavior pattern data to obtain attack type behavior mapping data; Perform attack behavior situation awareness according to the attack type behavior mapping data to obtain attack behavior situation awareness data; Performing system abnormality feedback analysis on the computer system according to the attack type behavior mapping data and the attack behavior situation awareness data to obtain system abnormality feedback data; wherein step S2 includes: Step S21: performing attack behavior mapping processing on the attack type behavior pattern data according to the firewall attack interception data to obtain attack type behavior mapping data; Step S22: performing attack behavior frequency statistics on the attack type behavior mapping data to obtain attack behavior frequency data; Step S23: performing attack thermal area distribution analysis on the computer system according to the attack type behavior mapping data and the attack behavior frequency data to obtain attack thermal area data; Step S24: performing attack behavior situation awareness on the attack thermal area data according to the attack type behavior mapping data to obtain attack behavior situation awareness data; wherein step S24 includes: Step S241: performing intrusion attempt capture on the attack thermal area data according to the attack type behavior mapping data to obtain attack type intrusion capture data; Step S242: performing an intrusion progressive intensity assessment on the attack type intrusion capture data to obtain intrusion progressive intensity data; Step S243: performing a biased progressive concealed analysis on the invasion progressive intensity data to obtain the invasion biased progressive concealed data; wherein performing a biased progressive concealed analysis on the invasion progressive intensity data includes: Based on the feedforward neural network, multi-level neural network learning is performed on the invasion progressive intensity data to obtain invasion progressive neural network data; According to the invasion progressive neural network data, the invasion bias is captured on the invasion progressive intensity data to obtain the invasion bias data; Perform invasion logic strategy deduction on invasion progressive intensity data according to invasion bias data to obtain invasion logic deduction strategy data; Perform intrusion combination means simulation on the intrusion logic deduction strategy data to obtain intrusion means combination simulation data; According to the intrusion logic deduction strategy data, the intrusion means combination simulation data and the intrusion bias data, the biased progressive concealment analysis is carried out to obtain the intrusion biased progressive concealment data; Step S244: performing intrusion intensity compensation on the intrusion progressive intensity data according to the intrusion bias progressive concealment data to obtain intrusion progressive intensity compensation data; Step S245: performing attack behavior situation awareness on the attack thermal area data according to the invasion progressive intensity compensation data and the invasion bias progressive concealment data to obtain attack behavior situation awareness data; Step S25: performing system abnormality feedback analysis on the computer system according to the attack type behavior mapping data and the attack behavior situation awareness data to obtain system abnormality feedback data; wherein step S25 includes: Step S251: simulating the abnormal behavior state of the computer system according to the attack type behavior mapping data and the attack behavior situation awareness data to obtain abnormal behavior state simulation data; wherein the abnormal behavior state simulation data includes CPU abnormal behavior state data and memory abnormal behavior state data; Step S252: performing clock frequency abnormal fluctuation analysis on the CPU abnormal behavior state data to obtain clock frequency abnormal fluctuation data; Step S253: performing a read rate loss simulation calculation on the memory abnormal behavior state data to obtain memory read rate loss data; Step S254: performing a memory access failure probability assessment on the memory read rate loss data according to the clock frequency abnormal fluctuation data to obtain memory access failure probability data; Step S255: performing system abnormality feedback analysis on the computer system according to the clock frequency abnormal fluctuation data, the memory read rate loss data, and the memory access failure probability data to obtain system abnormality feedback data; Step S3: Design intrusion detection rules according to the system abnormality feedback data to obtain intrusion detection rule data; construct a computer system security verification model for the intrusion detection rule data based on the random forest algorithm to obtain a system security verification model; Step S4: Perform system security risk detection according to the system security verification model to obtain system security risk data; perform system configuration optimization on the system security risk data to obtain system configuration optimization data.

2. The computer system security model verification method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: extracting attack interception data of the computer system through the computer firewall to obtain firewall attack interception data; Step S12: performing attack target analysis on the firewall attack interception data to obtain attack target data; Step S13: Identify the attack type according to the firewall attack interception data and the attack target data to obtain attack type data; Step S14: Analyze the attack type data for attack behavior patterns to obtain attack type behavior pattern data.

3. The computer system security model verification method according to claim 2, characterized in that: Step S14 includes the following steps: Step S141: extracting interception logs of firewall attack interception data to obtain interception log data; Step S142: performing attack type log matching on the attack type data according to the interception log data to obtain attack type log matching data; Step S143: extracting attack type network traffic from the attack type log matching data to obtain attack type network traffic data; Step S144: performing an instantaneous geometric growth analysis of traffic according to the attack type network traffic data to obtain instantaneous geometric growth data of traffic; Step S145: performing traffic potential growth pattern recognition on the traffic instantaneous geometric multiple growth data to obtain traffic potential growth pattern data; Step S146: performing attack behavior pattern analysis on the attack type data according to the traffic potential growth pattern data to obtain attack type behavior pattern data.

4. The computer system security model verification method according to claim 3, characterized in that: The analysis of the instantaneous geometric growth of traffic based on the attack type network traffic data includes the following steps: Perform time series analysis on the attack type network traffic data to obtain the attack type network traffic time series data; Performing time series smoothing processing on the attack type network traffic data according to the attack type network traffic time series data to obtain traffic time series smoothing data; Perform instantaneous traffic growth Gaussian process analysis on attack type network traffic data based on traffic time series smoothing data to obtain instantaneous traffic growth Gaussian process data; According to the instantaneous flow growth Gaussian process data, the flow time series smoothing data is processed into flow time series segments to obtain flow time series segmented data; Based on the instantaneous flow growth Gaussian process data, the flow time series segmented data is calculated to increase geometric series, and the segmented flow geometric growth data is obtained; According to the Gaussian process data of instantaneous flow growth and the segmented flow geometric growth data, the instantaneous geometric multiple growth analysis of flow is performed to obtain the instantaneous geometric multiple growth data of flow.

5. The computer system security model verification method according to claim 1, characterized in that: The memory access failure probability assessment for memory read rate loss data includes the following: Decomposing the abnormal frequency fluctuation trend of the clock frequency abnormal fluctuation data to obtain abnormal frequency fluctuation trend decomposition data; According to the abnormal frequency fluctuation trend decomposition data, the memory read rate loss data is modeled in the memory read rate loss state space to obtain the read rate loss state space model; Capturing the nonlinear relationship of memory read rate loss data according to the read rate loss state space model to obtain read loss nonlinear relationship data; Perform non-parametric statistical probability calculation on the memory read rate loss data according to the non-linear relationship data of the read loss, and obtain the non-parametric probability data of the read rate; The memory access failure probability is evaluated based on the read loss nonlinear relationship data and the read rate nonparametric probability data to obtain the memory access failure probability data.

6. The computer system security model verification method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: Design intrusion detection rules according to the attack behavior situation awareness data and the system abnormality feedback data to obtain intrusion detection rule data; Step S32: Encoding the intrusion detection rule data to obtain intrusion detection rule encoding data; Step S33: construct a computer system security verification model based on the intrusion detection rule encoding data, attack behavior situation awareness data and system abnormality feedback data based on the random forest algorithm to obtain a system security verification model.