Network attack event prediction and optimal active defense strategy selection system
Through the distributed dense network system and HMM model, combined with social engineering knowledge base and attack classification standards, the types of network attacks are predicted and optimal defense strategies are adopted, which solves the problem of difficult description of the real state and untimely defense of the network attack and defense process in the existing technology. Problem, achieving efficient cyber attack prediction and defense.
Patent Information
- Application Number
- CN202210049295.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-17
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-01-17
AI Technical Summary
The existing cyber attack and defense process is difficult to describe the real state of the network. Traditional security detection tools cannot be applied to super-large network environments and cannot effectively prevent cyber attacks. They usually can only defend after being attacked.
A distributed dense network system is used to combine the HMM model to capture and classify network attack events, match features through social engineering knowledge base and attack classification standards, predict possible attack types, and use the optimal active defense strategy to select the system to adopt before attack Defensive measures.
It realizes efficient and reliable cyber attack prediction and defense, and can adopt optimal strategies before attacks, significantly improving the security and defense capabilities of the cyberspace.
Smart Images

Figure CN114553481B_ABST
Abstract
Description
Technical Field
[0001] The implementation of the present invention involves the field of network attack and defense and the protection of network space security, mainly an HMM prediction system and an optimal active defense strategy selection system. Background Art
[0002] In today's information age, the rapid development of network information technology has not only promoted social production and change, created a new space for human life, but also greatly changed the way humans understand and transform the world. At present, network information technology is widely used in various fields of human society, and the network environment has developed from a simple Internet to a ubiquitous cyberspace. With its close integration with the real world, cyberspace has shown unprecedented vitality and vigor, and has strongly promoted the development and progress of society in the information age. However, cyberspace is like a double-edged sword. While promoting social and economic prosperity and national prosperity, it also faces serious security threats and challenges. The current network security situation is becoming increasingly severe. To this end, we have the need and responsibility to face the many challenges in the field of cyberspace security, enhance network defense and deterrence capabilities, and ensure the security of cyberspace. The rapid popularization of the Internet and mobile Internet has made it easier for the general public to access the Internet. While enjoying the convenience brought by the Internet, people are prone to great harm caused by cyber attacks due to their lack of network security awareness.
[0003] The existing network attack and defense process is difficult to describe the real state of the network. The HMM prediction method can construct the real state of the network based on the known attack and defense state, which can solve this problem more effectively. Secondly, traditional security detection tools are generally deployed at specific locations at the entrances and exits of enterprises and business systems, and cannot be applied to the current ultra-large-scale network environment. In addition, each detection point is independent of each other and has exclusive data. The attack event is usually evaluated based on the detection result of a certain detection point, and the detection effect of the attack event is not good. In addition, the current response to network attacks is often to defend after the attack, which cannot effectively ensure network security. The optimal strategy selection system for active defense can effectively defend according to the attacks that occur in the network to ensure the smooth and safe operation of the network environment. Summary of the invention
[0004] The present invention provides a system for obtaining network attack events using a distributed dense network system, and then using the HMM prediction method and the optimal defense strategy selection system. For attacks that may be suffered by the network in the future, the HMM method is used to reasonably predict the possible attacks, and then the optimal strategy selection system for active defense is used to achieve efficient and reliable prediction of attacks suffered by the network, and the optimal strategy is used to ensure the security of the network space before the network is attacked. In the process of the network being attacked, the method of the HMM model is more in line with the actual situation of network attack and defense, and has good practical guiding significance. The active defense method defends before the attack, can greatly defend the attacks suffered by the network, and has an advantage that cannot be ignored for practical applications.
[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions:
[0006] In the first aspect, the present invention adopts a method for detecting and classifying network attack events, mainly using a distributed dense network system, which is composed of multiple dense network systems. It includes:
[0007] Capture cyber attack events;
[0008] According to the social engineering knowledge base and the existing attack classification standards, feature matching is performed on the network attack events;
[0009] The optional network attack events are divided into:
[0010] If the attacker features included in the event features belong to the target attacker feature set included in the social engineering knowledge base, the network attack event is determined to be a social engineering attack event; otherwise, the next step of matching is performed.
[0011] According to the existing attack classification standards, network attack incidents are divided into active attack incidents and passive attack incidents according to the attack characteristics.
[0012] Matching cyber attack events;
[0013] The network attack events are classified into social engineering attack events, active attack events such as (message tampering, forgery, denial of service) and passive attack events such as (traffic analysis, eavesdropping) according to the matching results.
[0014] Then matching is performed based on the features, which are further subdivided into message tampering, forgery, denial of service traffic analysis, and eavesdropping.
[0015] The detection rule set includes at least: message tampering, forgery, denial of service traffic analysis, and eavesdropping.
[0016] According to the type of match, the database content is updated in real time to better perform the next match.
[0017] In a second aspect, the present invention provides a capability of predicting network attack events based on an HMM model, which is applied to a prediction system, including:
[0018] Capture module, used to capture network attack events;
[0019] A matching module, used to perform feature matching on the network attack events according to a social engineering knowledge base and existing attack classification standards;
[0020] The classification module classifies the network attack events according to the matching results, and divides them into social engineering attack events, active attack events such as (message tampering, forgery, denial of service) and passive attack events such as (traffic analysis, eavesdropping).
[0021] The prediction module predicts the types of network attack events that may occur based on these classified network attack events and the known state set.
[0022] In a third aspect, the present invention also provides an optimal active defense strategy selection system based on HMM, characterized in that it includes: a dense network arrangement system, a controller, an event analysis system, a data processing system, a data storage system, a strategy selection system, and an operator system;
[0023] A honeynet system is used to execute the method for detecting network attack events as described in any one of claims 1 and send the detected network attack events to a controller; the controller sends instructions to the system after processing, and then arranges a distributed honeynet system.
[0024] The controller is used to store the network attack events sent by each honeynet system into the data storage system, and send event analysis instructions corresponding to each social engineering attack event and the existing attack classification standards to the event analysis system; and send instructions to the strategy selection system to select the optimal defense strategy.
[0025] The event analysis system is used to perform correlation analysis on each social engineering attack event that matches the current event analysis instruction and the existing attack classification standard, and update the social engineering knowledge base and the existing attack classification standard included in the data storage system according to the correlation analysis results. The event types in the storage and system are classified into social engineering attack events, active attack events such as (tampering with messages, forgery, denial of service) and passive attack events such as (traffic analysis, eavesdropping).
[0026] The data processing system is used to systematically process the network attack events stored in the data storage system. According to the existing data types, they are divided into observation sequence - O, state sequence - I, initial state probability - π, state transition matrix - A, observation probability matrix - B, and then the HMM model is used for modeling and analysis. Instructions are sent to the operator system for calculation.
[0027] The operator system is used to receive instructions issued by the data processing system for operation. According to the operation results, the hidden state set can be predicted, that is, the type of event attack that may occur. Send instructions to the controller. The hidden state set and the existing attack event state set constitute all the states that are attacked in the network. We take active defense methods based on this state set to ensure network security.
[0028] The strategy selection system is used to receive instructions from the control system. The hidden state set and the existing attack event state set are composed of all the states that have been attacked in the network. We select the best strategy for defense based on this state set to ensure network security. The strategy selection system includes effective protection methods such as the deployment of dense network systems and firewalls.
[0029] The distributed dense network system is composed of multiple dense network systems.
[0030] In a fourth aspect, the present invention provides a strategy selection system, the strategy selection system comprising:
[0031] Systems consisting of multiple processors or multiple memories;
[0032] A storage device for storing a plurality of programs or a program set consisting of a plurality of programs;
[0033] When multiple programs are executed by a processor, the processor set implements the optimal strategy selection method provided by the present invention.
[0034] The technical solution implemented by the present invention is that the dense network system is used to obtain network attack events; according to the social engineering library and the existing attack classification standards, the network attack events are matched with characteristics; according to the matching results, the captured network attack event states are classified into social engineering attack events, active attack events such as (tampering with messages, forgery, denial of service) and passive attack events such as (traffic analysis, eavesdropping). The problem that the traditional classification monitoring method in the prior art cannot effectively classify network attack events is solved, and high-efficiency and high-precision detection and classification of network attack events are achieved. Then, according to the attack events, the data processing system is used to divide the states of the obtained network attack events, and according to the existing data types, they are divided into observation sequence-O, state sequence-I, initial state probability-π, state transfer matrix-A, observation probability matrix-B, and then a HMM model is constructed, and the operator system is used to solve the state. According to the operation results, the hidden state set can be predicted, that is, the event attack type may occur. Send instructions to the controller. The hidden state set and the existing attack event state set constitute all the states that are attacked in the network. We take active defense methods according to this state set to defend and ensure network security. The hidden state set and the existing attack event state set constitute all the states that are attacked in the network. We select the best strategy to defend and ensure network security based on this state set. The strategy selection system includes effective protection methods such as dense network system and firewall. It can explain and prevent network attack events more clearly and has practical significance. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] Figure 1 It is a flow chart of a method for detecting and classifying network attack events in Embodiment 1 of the present invention;
[0036] Figure 2 It is a comprehensive flow chart of the prediction method based on the HMM model and the strategy selection in the second embodiment of the present invention;
[0037] Figure 3 is a flow chart of each module in the second embodiment of the present invention;
[0038] Figure 4 is a schematic diagram of the structure of the computing system in the third embodiment of the present invention;
[0039] Figure 5 is a schematic diagram of the structure of a distributed dense network system in Embodiment 4 of the present invention; DETAILED DESCRIPTION
[0040] In order to more clearly illustrate the existing technical solutions, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. The specific embodiments described herein are only used to explain the present invention, but not to limit the present invention. It should also be noted that, for the convenience of description, only the parts related to the present invention are shown in the accompanying drawings, rather than all structures. For those of ordinary skill in the art, relevant work can be completed based on these drawings.
[0041] Embodiment 1
[0042] like Figure 1 As shown, a flow chart of a method for detecting and classifying network attack events in Embodiment 1 of the present invention;
[0043] This embodiment can be applied to the detection of network attack events. Then the network attack events can be classified according to the classification criteria, and finally the database system can be updated in real time. This module can be implemented by combining software and hardware. Figure 1 As shown, the steps are as follows:
[0044] Step 101: Obtain network attack events.
[0045] The distributed honeynet system is essentially a combination of multiple honeynet systems. The honeynet system is a technology for inducing deception by attacking. By placing some hosts and information as bait, it can induce the attacker to attack it, thereby obtaining network attack events, capturing network attack events for analysis, and understanding the methods used by the attacker. The honeynet system can induce network attacks and consume the resources of network attacks, thereby protecting network security.
[0046] In an embodiment of the present invention, a distributed dense network system is composed of multiple dense network systems. Multiple distributed dense network systems are deployed in a large area, and multiple dense network systems are deployed in different areas, which together constitute a distributed dense network system, thereby expanding the detection range of network events and improving the ability to detect network attack events.
[0047] In the embodiment of the present invention, the dense network system is a real computer system, and the dense network system is arranged in a virtual machine of a computer at a designated location to prevent the dense network system from being damaged and causing damage to the real computer.
[0048] Step 102: Perform feature matching on network attack events according to the social engineering database and attack classification standards.
[0049] Step 103 classifies network attack events into active attack events, passive attack events and social engineering attack events according to the matching results.
[0050] Step 104 updates the database in real time according to the characteristics of the event, so that the next match can be faster.
[0051] In the technical solution of the example of the present invention, the dense network system captures network attack events, performs feature matching on network attack events according to the social engineering database and attack classification standards, and classifies network attack events into active attack events, passive attack events and social engineering attack events according to the matching results, and updates the database system in real time according to the features of the events. Such a multi-cycle system can solve the problem of insufficient existing network attack detection, and the database system can be updated in real time, which can more quickly classify network attack events and provide a good foundation for the next step of efficiently and accurately predicting the attacks of network events.
[0052] Embodiment 2
[0053] Figure 2 This is a schematic diagram of the flow structure of the detection method based on HMM and the optimal strategy selection system based on active defense in the second implementation case of the present invention. This embodiment can be implemented by combining software and hardware, assembled into an overall system, and deployed where needed. Figure 2 As shown:
[0054] Flowchart of an example of the present invention, a dense network system is deployed throughout the network to obtain network attack events;
[0055] According to the social engineering database and existing attack classification standards, feature matching is performed on network attack events;
[0056] According to the matching results, the captured network attack event status is classified into social engineering attack events, active attack events such as (tampering with messages, forgery, denial of service) and passive attack events such as (traffic analysis, eavesdropping). This solves the problem that the traditional classification monitoring method in the prior art cannot effectively classify network attack events, and achieves high-efficiency and high-precision detection and classification of network attack events.
[0057] Then, based on the attack events, the data processing system is used to divide the acquired network attack events into states. According to the existing data types, they are divided into observation sequence - O, state sequence - I, initial state probability - π, state transfer matrix - A, observation probability matrix - B, and then an HMM model is constructed. The operator system is used to solve the state. Based on the calculation results, the hidden state set can be predicted, that is, the possible event attack type.
[0058] Send instructions to the controller. The hidden state set and the existing attack event state set constitute all the states that have been attacked in the network. We take active defense methods based on this state set to defend and ensure network security. The hidden state set and the existing attack event state set constitute all the states that have been attacked in the network. We take the best strategy to defend and ensure network security based on this state set. The strategy selection system includes effective protection methods such as the deployment of dense network systems and firewalls. It can explain and prevent network attack events more clearly, which has practical significance.
[0059] Figure 3 It is the overall system flow module diagram: it is divided into six modules, namely acquisition module, matching module, classification module, data processing module, HMM prediction module, and strategy selection module.
[0060] An acquisition module 301 is used to acquire network attack events;
[0061] The matching module 302 performs feature matching on network attack events according to the social engineering database and attack classification standards;
[0062] The classification module 303 classifies the network attack events into active attack events, passive attack events and social engineering attack events according to the matching results;
[0063] The data processing module 304 performs data modeling according to the event characteristics after classification, divides the obtained network attack events into states, and divides them into observation sequence - O state sequence - I initial state probability - π state transfer matrix - A observation probability matrix - B according to the existing data type, and then constructs an HMM model.
[0064] The HMM prediction module 305 uses the operator system to perform state solving of the HMM model, and according to the operation result, it can predict the hidden state set, that is, the event attack type that may occur.
[0065] Strategy selection module 306, the hidden state set and the existing attack event state set constitute all the states of the network that are attacked. We select the best strategy for defense based on this state set to ensure network security. The strategy selection system includes effective protection methods such as the deployment of a dense network system and a firewall.
[0066] Embodiment 3
[0067] Figure 4 This is a diagram of the connection between the data processing module system, the database system and the dense network system in the third embodiment of the present invention. This embodiment describes the composition of the data processing module in detail.
[0068] The data storage system 401 is used to store network attack events.
[0069] The event analysis system 402 is used to perform correlation analysis on the types of network attack events that are obtained, and to update the features corresponding to the social engineering database and event types in the database system in real time according to the analysis results.
[0070] The prediction and analysis system 403 predicts possible network attack events according to the designation issued by the controller.
[0071] The operator system 404 is used to receive instructions issued by the data processing system for operation. According to the operation results, the hidden state set, that is, the event attack type that may occur, can be predicted, and then instructions are issued to the controller.
[0072] The distributed dense network system 405 is deployed throughout the network to obtain network attack events.
[0073] Database system 406 includes a social engineering library and existing attack classification standards.
[0074] Data processing system 407, data processing system, is used to systematically process the network attack events stored in the data storage system, and divide them into observation sequence - O state sequence - I initial state probability - π state transition matrix - A observation probability matrix - B, and then use the HMM model to model and analyze. Send instructions to the operator system for calculation
[0075] The controller 408 is used to store the network attack events acquired by the dense network system 405 into the data storage system 401, and send instructions to the event analysis system and the data processing system.
[0076] In the example of the present invention, the event analysis system 401 adds active attack, passive attack, and social engineering attack tags to various network attack events in the data storage system 401.
[0077] In the example of the present invention, the controller 408 is the control center of the entire module, connecting various modules and sending or receiving instructions to various systems.
[0078] Embodiment 4
[0079] Figure 5 It is an overall structural diagram of a distributed dense network system in Embodiment 4 of the present invention. Figure 5 The implementation method shown is the composition of a dense network system, which provides event acquisition for the entire module. The dense network system is composed of multiple such systems, and multiple such dense network systems are deployed in the entire network to form a distributed dense network system.
[0080] like Figure 5 As shown, the dense network system 501 is presented in the form of various components. The dense network system includes: multiple processing units 505, storage 509, memory unit 510, controller 507, various I / O interfaces 504, external devices 502, display device 503, network adapter 506, and database system 508.
[0081] The network system includes a variety of computer system readable media, which can be any available media that can be accessed by the network system, including volatile and non-volatile media, removable and non-removable media.
[0082] The honeynet system can communicate with one or more external devices 502 (e.g., keyboards, pointing devices, etc.), and can also communicate with one or more devices that enable users to communicate with the honeynet system devices, and any device that enables the honeynet system to communicate with one or more other computing devices (e.g., network cards). Such communication can be carried out through input / output (I / O) interface 504.
[0083] The present invention provides a prediction method based on HMM and a system for selecting an optimal defense strategy. When the current network is attacked, the HMM method is used to reasonably predict possible attacks, and then the optimal strategy selection system based on active defense is used to achieve efficient and accurate prediction of attacks suffered by the network, and the optimal strategy is used to ensure the security of the network space before the network is attacked. In the process of the network being attacked, the method of using the HMM model is more in line with the actual situation of network attack and defense, and has good practical guiding significance. The active defense method, which defends before the attack, can greatly reduce the attacks suffered by the network, and has an advantage that cannot be ignored for practical applications.
[0084] Finally, it is important to note that the above are only preferred embodiments of the present invention and are only used to illustrate the technical principles used in the present invention. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the attached claims. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of protection of the present invention.
Claims
1. A method for predicting network attack events and selecting the optimal active defense strategy, which uses network attack event prediction and strategy selection, including: Capture cyber attack events; According to the type of network attack, the captured network attack events are matched with the existing network attack event types; Based on the matching results, it is determined whether the captured network attack event is a specific type of attack. The optimal strategy selection method is finally selected based on the event prediction; The prediction method using HMM (Hidden Markov Model) can construct the real state of the network based on the known attack and defense state; These include: Capture module, used to capture network attack events; A matching module, used for performing feature matching on the network attack events according to a social engineering knowledge base and existing attack classification standards; The classification module classifies the network attack events according to the matching results, and classifies them into social engineering attack events, active attack events and passive attack events; A data processing module, used for systematically processing the network attack events stored in the data storage system; The prediction module predicts the types of possible network attack events based on the known state set. The strategy selection module is used to receive instructions from the control system. The hidden state set and the existing attack event state set constitute all the states that have been attacked in the network. We select the optimal strategy for defense based on this state set.
2. The method for predicting network attack events and selecting the optimal active defense strategy according to claim 1, characterized in that: According to the social engineering database and the existing attack classification standards; according to the matching results, determine whether the detected network attack event is a specific type of attack; according to the match, classify the status of the captured network attack event into social engineering attack events, active attack events and passive attack events.
3. According to the method for predicting network attack events and selecting the optimal active defense strategy as described in claim 2, its main feature is that after the network attack events are classified using existing classification standards, the database system is updated in real time with newly emerging features.
4. A network attack event prediction and optimal active defense strategy selection system using the method described in claim 1 or 3, characterized in that: The detection and classification of network attack incidents mainly use a distributed dense network system, which is composed of multiple dense network systems.
5. The network attack event prediction and optimal active defense strategy selection system according to claim 4 comprises: Arrange dense network system, controller, event analysis system, data processing system, data storage system, strategy selection system, and operator system; The dense network system is used to execute the method for detecting network attack events according to any one of claims 1 or 3, and send the detected network attack events to the controller; The controller is used to store the network attack events sent by each honeynet system in the data storage system, and send event analysis instructions corresponding to each social engineering attack event and the existing attack classification standard to the event analysis system; And send instructions to the strategy selection system to select the optimal defense strategy; The event analysis system is used to perform correlation analysis on each social engineering attack event that matches the current event analysis instruction and the existing attack classification standard, and update the social engineering knowledge base and the existing attack classification standard included in the data storage system according to the correlation analysis result; classify the event types stored in the system into social engineering attack events, active attack events and passive attack events; The data processing system is used to perform systematic data processing on the network attack events stored in the data storage system, and divide them into observation sequence - O, state sequence - I, initial state probability - π, state transition matrix - A, observation probability matrix - B according to the existing data types, and then use the HMM model for modeling analysis; The operator system is used to receive instructions issued by the data processing system for operation. According to the operation results, the hidden state set can be predicted, that is, the event attack type that may occur; the instructions are sent to the controller. The hidden state set and the existing attack event state set constitute all the states that are attacked in the network. We take active defense methods to defend according to this state set; The strategy selection system is used to receive instructions from the control system. The hidden state set and the existing attack event state set constitute all the states that have been attacked in the network. We select the optimal strategy for defense based on this state set; the strategy selection system includes effective protection methods such as the deployment of a dense network system, firewalls, etc.
6. The network attack event prediction and optimal active defense strategy selection system according to claim 5 is characterized in that: The strategy selection system issues instructions through the control system based on the state set calculated by the operator system. Each event has a corresponding optimal strategy selection, and strategy selection is performed for defense before future attack events occur.
Citation Information
Patent Citations
Network security accident classification and predicting method and system
CN107241352A
Threat detection method based on event sequence
CN110535878A
Social engineering attack event detection method, device and system
CN111859374A
