A blockchain big data security processing system
By calculating the abnormal coefficients of traffic characteristics, transmission characteristics and structural characteristics in the blockchain big data system, and combining machine learning models to identify malicious traffic, the problem of low accuracy of malicious traffic recognition in the existing technology is solved, and the more Efficient detection and identification of malicious traffic.
Patent Information
- Application Number
- CN202410579272.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-11
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-05-11
AI Technical Summary
In the prior art, when identifying malicious traffic, attackers use camouflage traffic attacks to make the identification accuracy based on historical traffic data feature information low, making it difficult to effectively defend against distributed denial of service attacks in blockchain big data systems.
The traffic feature analysis unit, transmission feature analysis unit and structural feature analysis unit are used to calculate the characteristic anomaly coefficients of traffic data respectively, and combined with machine learning models, a malicious traffic recognition system is established to determine the malicious traffic through a combination of multi-dimensional and different directions.
It improves the detection performance and recognition efficiency of malicious traffic, enhances the ability to adapt to complex environments, improves the accuracy of malicious traffic recognition, and can automatically learn and adapt to new attack methods.
Smart Images

Figure CN118400157B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of blockchain big data security management, and specifically to a blockchain big data security processing system. Background Art
[0002] As is known, blockchain big data security management involves multiple aspects, aiming to ensure the integrity, availability and confidentiality of big data in blockchain systems. Blockchain big data systems are threatened by distributed denial of service attacks, which attempt to consume system resources through a large number of invalid requests, causing the system to fail to operate normally. In order to prevent distributed denial of service attacks, a series of protective measures need to be taken, such as using firewalls, traffic monitoring and restricting IP addresses. For blockchain big data systems, such attacks will affect the stability of nodes, the speed of transaction processing and the availability of the network.
[0003] For example, the patent application with application publication number CN117097506A, application publication date 2023.11.21, and name "Communication Network Traffic Intrusion Detection Method and Detection System Based on Blockchain", its detection method specifically includes: collecting real-time traffic data corresponding to the terminal in the communication network, analyzing and processing the real-time traffic data, and obtaining real-time traffic data feature information; extracting corresponding historical traffic data feature information from the blockchain according to the network browsing log of the terminal; comparing the real-time traffic data feature information with the historical traffic data feature information to determine whether the real-time traffic data is malicious traffic data; analyzing and processing the malicious traffic data to determine the attack status of the malicious traffic data on the terminal, thereby adjusting the working status of the terminal.
[0004] Similar to the above application, it collects the characteristic information of its traffic data and compares it with the characteristic information of historical traffic data to determine whether it is malicious traffic data. However, in order to effectively avoid malicious traffic detection, attackers will actively refer to and imitate the behavior characteristics of historical traffic and launch disguised traffic attacks, resulting in low accuracy in identifying malicious traffic by simply comparing and analyzing it with the characteristic information of historical traffic data. Summary of the invention
[0005] The purpose of the present invention is to provide a blockchain big data security processing system to solve the above-mentioned deficiencies in the prior art.
[0006] In order to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a blockchain big data security processing system, comprising:
[0007] Traffic data collection unit, which is deployed in the blockchain network to collect network traffic data in real time;
[0008] A feature collection unit, which retrieves the network traffic data collected in real time, performs feature analysis on the traffic data, and extracts traffic feature information, transmission feature information, and structural feature information from the traffic data;
[0009] A flow characteristic analysis unit, which is used to analyze the flow characteristic information in the flow data and calculate the flow characteristic anomaly coefficient of the flow data;
[0010] A transmission characteristic analysis unit, which is used to analyze the transmission characteristic information in the flow data and calculate the abnormal coefficient of the flow data transmission characteristic;
[0011] A structural feature analysis unit, which is used to analyze the structural feature information in the flow data and calculate the structural feature anomaly coefficient of the flow data;
[0012] The model creation unit establishes a malicious traffic identification machine learning model, collects a historical training data set of network traffic data in advance, and trains the malicious traffic identification machine learning model based on the historical training data set; imports the currently acquired traffic feature anomaly coefficient, transmission feature anomaly coefficient, and structural feature anomaly coefficient into the malicious traffic identification machine learning model, and identifies malicious traffic through the malicious traffic identification machine learning model.
[0013] As a further description of the above technical solution: the traffic characteristic information in the traffic data includes the number of data packets in the traffic data and the source and destination IP address information corresponding to each data packet;
[0014] The transmission characteristic information in the flow data is the data packet transmission speed in the flow data;
[0015] The structural characteristic information in the traffic data is the data length of the data packets in the traffic data.
[0016] As a further description of the above technical solution: the flow characteristic analysis unit analyzes the flow characteristic information in the flow data, and calculates the flow characteristic abnormality coefficient of the flow data as follows:
[0017] Count the number of packets with the same source and destination IP address information, the number of packets with the same source IP address information, and the number of packets with the same destination IP address information in the traffic data in turn;
[0018] Calculate the flow characteristic anomaly coefficient Ks of the flow data, where the calculation formula of the flow characteristic anomaly coefficient Ks is:
[0019]
[0020] Where X is the total number of packets in the traffic data, m is the number of packets with the same source and destination IP address information, n is the number of packets with the same source IP address information but different destination IP address information, and i is the number of packets with the same destination IP address information but different source IP address information;
[0021] Indicates the weight coefficient of the anomaly when the source and target IP address information are the same. Indicates the anomaly weight coefficient when the source IP address information is the same but the target IP address information is different. Indicates the anomaly weight coefficient when the target IP address information is the same but the source IP address information is different.
[0022] As a further description of the above technical solution: the transmission characteristic analysis unit analyzes the transmission characteristic information in the flow data and calculates the abnormal coefficient of the flow data transmission characteristic as follows:
[0023] Collect the periodic interval of each data packet in the data flow whose transmission speed exceeds the benchmark value;
[0024] Calculate the abnormal coefficient K of traffic data transmission characteristics based on the periodic interval of each data packet C ;
[0025] The transmission characteristic anomaly is K C The calculation formula is:
[0026]
[0027] Where T jy is the cycle interval time, The average time between each data packet cycle.
[0028] As a further description of the above technical solution: the periodic interval of each data packet whose transmission speed exceeds the reference value in the collected data flow specifically comprises the following steps:
[0029] Collect the transmission speed of each data packet in the traffic data;
[0030] Preset transmission speed benchmark values based on historical data during normal operation
[0031] Compare the currently acquired transmission speed of each data packet with the transmission speed benchmark value, obtain the data packet whose transmission speed exceeds the transmission speed benchmark value, and extract the time node information of the corresponding data packet and record it as T y , where y∈(1, 2, 3…);
[0032] Calculate the corresponding data packet period interval time T based on the extracted data packet time node information Ty jy Cycle interval time Tjy The calculation formula is: jy =T y -T y-1 , which is to extract the interval time between adjacent time nodes in the corresponding data packet.
[0033] As a further description of the above technical solution: the structural feature analysis unit analyzes the structural feature information in the flow data and calculates the structural feature anomaly coefficient of the flow data as follows:
[0034] Preset the packet length threshold based on historical data during normal operation
[0035] Collect the data length of each data packet in the traffic data, compare the data length of each data packet currently obtained with the length threshold, and obtain the data packet whose data length exceeds the data length threshold;
[0036] Collect the IP address information of each data packet in the traffic data and count the number of different IP address information;
[0037] Collect IP address information of data packets with data length and length threshold, and count the number of different IP address information;
[0038] The abnormal coefficient of flow data structure characteristics is calculated through integrated analysis.
[0039] As a further description of the above technical solution: the flow data structure characteristic abnormal coefficient Kg is calculated by integration analysis, wherein the calculation formula of the structure characteristic abnormal coefficient Kg is:
[0040]
[0041] Among them, X is the total number of data packets in the traffic data, Q is the number of data packets whose data length exceeds the data length threshold, F is the number of different IP address information of the data packets in the traffic data; S is the number of different IP address information of the data packets whose data length exceeds the data length threshold.
[0042] As a further description of the above technical solution: the historical training data set includes several groups of training data collected from the historical traffic data of the blockchain, and each group of training data includes feature data and label data;
[0043] The characteristic data includes a flow characteristic anomaly coefficient, a transmission characteristic anomaly coefficient and a structural characteristic anomaly coefficient;
[0044] The label data is collected when each set of traffic data training data is collected. Through manual review and analysis, it is determined whether each set of traffic data is malicious traffic. If the traffic data is normal traffic, the label data is marked as 1. On the contrary, if the traffic data is malicious traffic, the label data is marked as 0.
[0045] As a further description of the above technical solution: Training a malicious traffic identification machine learning model based on a historical training data set is specifically:
[0046] Divide the historical training data set into 80% training set and 20% test set;
[0047] Construct a malicious traffic identification machine learning model, and input the training set as input data into the malicious traffic identification machine learning model for training, so as to obtain a trained malicious traffic identification machine learning model;
[0048] The test set is used to verify the accuracy of the trained malicious traffic identification machine learning model, and the malicious traffic identification machine learning model that meets the preset accuracy is output.
[0049] As a further description of the above technical solution: it also includes an alarm processing module. When the malicious traffic identification machine learning model identifies malicious traffic, the alarm processing module is triggered. The alarm module issues an alarm in time, quickly responds to malicious traffic attacks and takes protective measures.
[0050] In the above technical solution, the blockchain big data security processing system provided by the present invention has the following beneficial effects:
[0051] Compared with the traditional method of identifying malicious traffic by comparing and analyzing the characteristic information of historical traffic data, this blockchain big data security processing system can realize the targeted design of efficient feature extraction and analysis methods in the face of complex and diverse blockchain comprehensive environments. It can also comprehensively improve the detection performance, recognition efficiency and generalization ability of malicious traffic by combining the abnormal coefficient of traffic characteristics, abnormal coefficient of transmission characteristics and abnormal coefficient of structural characteristics in multiple dimensions and different directions. At the same time, by using machine learning to train the model and identify the normal and malicious traffic of traffic data, it can automatically learn and adapt to new malicious traffic attack methods, further improving the accuracy of malicious traffic identification. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.
[0053] Figure 1 A schematic diagram of a blockchain big data security processing system provided by an embodiment of the present invention; DETAILED DESCRIPTION
[0054] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.
[0055] See also Figure 1 , an embodiment of the present invention provides a technical solution: a blockchain big data security processing system, comprising:
[0056] The traffic data collection unit is deployed in the blockchain network to collect network traffic data in real time. Specifically, the traffic data collection unit is a traffic capture device and a data preprocessing module deployed in the network. The data preprocessing module performs data cleaning and data conversion on the collected data, removes noise and invalid data, and converts the original data into a format suitable for analysis according to subsequent analysis requirements;
[0057] A feature collection unit retrieves the network traffic data collected in real time, performs feature analysis on the traffic data, and extracts feature information from the traffic data. Specifically, the feature collection unit extracts useful features from the traffic data, which can help subsequent analysis to distinguish normal traffic from malicious traffic. The extracted traffic data feature information includes traffic feature information, transmission feature information, and structural feature information.
[0058] A flow characteristic analysis unit, which is used to analyze the flow characteristic information in the flow data and calculate the flow characteristic anomaly coefficient of the flow data;
[0059] A transmission characteristic analysis unit, which is used to analyze the transmission characteristic information in the flow data and calculate the abnormal coefficient of the flow data transmission characteristic;
[0060] A structural feature analysis unit, which is used to analyze the structural feature information in the flow data and calculate the structural feature anomaly coefficient of the flow data;
[0061] The larger the traffic characteristic anomaly coefficient, the transmission characteristic anomaly coefficient, and the structural characteristic anomaly coefficient, the greater the probability that the corresponding data traffic is malicious traffic.
[0062] The model creation unit establishes a malicious traffic identification machine learning model, collects a historical training data set of network traffic data in advance, and trains the malicious traffic identification machine learning model based on the historical training data set; imports the currently acquired traffic feature anomaly coefficient, transmission feature anomaly coefficient, and structural feature anomaly coefficient into the malicious traffic identification machine learning model, and identifies malicious traffic through the malicious traffic identification machine learning model.
[0063] In another embodiment provided by the present invention, the traffic characteristic information in the traffic data includes the number of data packets in the traffic data and the source and destination IP address information corresponding to each data packet;
[0064] The transmission characteristic information in the traffic data is the transmission speed of the data packets in the traffic data;
[0065] The structural characteristic information in the traffic data is the data length of the data packets in the traffic data.
[0066] In another embodiment provided by the present invention, the flow characteristic analysis unit analyzes the flow characteristic information in the flow data, and calculates the flow characteristic anomaly coefficient of the flow data as follows:
[0067] Count the number of packets with the same source and destination IP address information, the number of packets with the same source IP address information, and the number of packets with the same destination IP address information in the traffic data in turn;
[0068] Calculate the flow characteristic anomaly coefficient Ks of the flow data, where the calculation formula of the flow characteristic anomaly coefficient Ks is:
[0069]
[0070] Where X is the total number of packets in the traffic data, m is the number of packets with the same source and destination IP address information, n is the number of packets with the same source IP address information but different destination IP address information, and i is the number of packets with the same destination IP address information but different source IP address information;
[0071] Indicates the anomaly weight coefficient when the source and target IP address information are the same. Indicates the anomaly weight coefficient when the source IP address information is the same but the target IP address information is different. Indicates the anomaly weight coefficient when the target IP address information is the same but the source IP address information is different. Optional It should be noted that malicious traffic usually transmits data packets with the same source IP address and destination IP address. By statistically analyzing the source and destination IP address information of the data packets in the collected traffic data, the traffic characteristic anomaly coefficient Ks is calculated, and the traffic characteristic anomaly coefficient Ks is used to represent the probability that the traffic data is malicious traffic.
[0072] In another embodiment provided by the present invention, preferably, the transmission characteristic analysis unit analyzes the transmission characteristic information in the flow data, and calculates the flow data transmission characteristic abnormality coefficient specifically as follows:
[0073] Collect the periodic interval of each data packet in the data flow whose transmission speed exceeds the benchmark value;
[0074] Calculate the abnormal coefficient K of traffic data transmission characteristics based on the periodic interval of each data packet C ;
[0075] The transmission characteristic anomaly is K C The calculation formula is:
[0076]
[0077] Where T jy is the cycle interval time, The average value of the periodic interval of each data packet. The transmission characteristic anomaly coefficient is ≤1. When the transmission characteristic anomaly coefficient is equal to 1, it means that the data packets with a transmission speed exceeding the transmission speed benchmark value appear periodically, which is obviously malicious traffic.
[0078] The step of collecting the periodic interval of each data packet whose transmission speed exceeds the reference value in the data flow specifically includes the following steps:
[0079] Collect the transmission speed of each data packet in the traffic data;
[0080] Preset transmission speed benchmark values based on historical data during normal operation
[0081] Compare the currently acquired transmission speed of each data packet with the transmission speed benchmark value, obtain the data packet whose transmission speed exceeds the transmission speed benchmark value, and extract the time node information of the corresponding data packet and record it as T y , where y∈(1, 2, 3, ...);
[0082] Calculate the corresponding data packet period interval time T based on the extracted data packet time node information Ty jy Cycle interval time T jy The calculation formula is: jy =T y -T y-1 , which is to extract the interval time between adjacent time nodes in the corresponding data packet.
[0083] It should be noted that the blockchain network layer is a dynamic and open network environment. In order to increase the success rate of attacks, attackers use dynamic traffic obfuscation and covert traffic penetration to evade existing detection methods. For example, malicious requests are periodically sent to target nodes through puppet machines to occupy network resources. The malicious traffic is transmitted at a fast speed, which causes network congestion and affects normal network operation. The transmission speed benchmark value is determined based on historical data, and the data packets whose transmission speed exceeds the transmission speed benchmark value are determined by comparison and analysis. The corresponding data packet cycle interval is determined, and the volatility change of the interval time is calculated. The smaller the volatility change, the larger the transmission characteristic anomaly coefficient, that is, the probability that the traffic data is malicious traffic is indicated by the transmission characteristic anomaly coefficient.
[0084] In another embodiment provided by the present invention, the structural feature analysis unit analyzes the structural feature information in the flow data, and calculates the structural feature anomaly coefficient of the flow data as follows:
[0085] Preset the packet length threshold based on historical data during normal operation
[0086] Collect the data length of each data packet in the traffic data, compare the data length of each data packet currently obtained with the length threshold, and obtain the data packet whose data length exceeds the data length threshold;
[0087] Collect the IP address information of each data packet in the traffic data and count the number of different IP address information;
[0088] Collect IP address information of data packets with data length and length threshold, and count the number of different IP address information;
[0089] The abnormal coefficient of flow data structure characteristics is calculated through integrated analysis.
[0090] Specifically, the flow data structure characteristic abnormal coefficient Kg is calculated by integration analysis, wherein the calculation formula of the structure characteristic abnormal coefficient Kg is:
[0091]
[0092] Among them, X is the total number of data packets in the traffic data, Q is the number of data packets whose data length exceeds the data length threshold, F is the number of different IP address information of the data packets in the traffic data; S is the number of different IP address information of the data packets whose data length exceeds the data length threshold.
[0093] Among them, when the attacker carries out a malicious traffic attack, the attacker will control a certain number of puppet machines to send data packets with abnormal structures, such as data packets with abnormal data length. By collecting the ratio data of data packets with abnormal data length to the total data packets, and comparing the ratio of the number of different IP address information corresponding to the data packets with abnormal data length to the total number of different IP address information, it is obvious that the attacker controls a certain number of puppet machines, which will lead to smaller differences in their IP address information. That is, the larger the structural characteristic abnormality coefficient Kg is, the greater the probability that the corresponding data traffic is malicious traffic.
[0094] In another embodiment provided by the present invention, the historical training data set includes several groups of training data collected from the historical traffic data of the blockchain, and each group of training data includes feature data and label data;
[0095] The characteristic data includes a flow characteristic anomaly coefficient, a transmission characteristic anomaly coefficient and a structural characteristic anomaly coefficient; wherein the flow characteristic anomaly coefficient, the transmission characteristic anomaly coefficient and the structural characteristic anomaly coefficient of the training data are calculated in the same manner as the flow characteristic anomaly coefficient, the transmission characteristic anomaly coefficient and the structural characteristic anomaly coefficient of the flow data in the above-mentioned calculation method;
[0096] When collecting training data for each set of traffic data, the label data is manually reviewed and analyzed to determine whether each set of traffic data is malicious traffic. If the traffic data is normal traffic, the label data is marked as 1. On the contrary, if the traffic data is malicious traffic, the label data is marked as 0.
[0097] The specific training of malicious traffic identification machine learning models based on historical training data sets is as follows:
[0098] Divide the historical training data set into 80% training set and 20% test set;
[0099] Construct a malicious traffic identification machine learning model, and input the training set as input data into the malicious traffic identification machine learning model for training, so as to obtain a trained malicious traffic identification machine learning model;
[0100] The test set is used to verify the accuracy of the trained malicious traffic identification machine learning model, and the malicious traffic identification machine learning model that meets the preset accuracy is output.
[0101] This blockchain big data security processing system realizes the targeted design of efficient feature extraction and analysis methods in the face of complex and diverse blockchain comprehensive environments, and comprehensively improves the detection performance, recognition efficiency and generalization ability of malicious traffic through combined judgment and identification of multiple dimensions and different directions of traffic feature anomaly coefficients, transmission feature anomaly coefficients and structural feature anomaly coefficients. At the same time, by using machine learning to train models and identify normal and malicious traffic in traffic data, it can automatically learn and adapt to new malicious traffic attack methods, further improving the accuracy of malicious traffic identification.
[0102] In another embodiment provided by the present invention, an alarm processing module is also included. When the malicious traffic identification machine learning model identifies malicious traffic, the alarm processing module is triggered. The alarm module issues an alarm in a timely manner, quickly responds to malicious traffic attacks and takes protective measures.
[0103] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.
Claims
1. A blockchain big data security processing system, characterized in that: include: Traffic data collection unit, which is deployed in the blockchain network to collect network traffic data in real time; A feature collection unit, which retrieves the network traffic data collected in real time, performs feature analysis on the traffic data, and extracts traffic feature information, transmission feature information, and structural feature information from the traffic data; A flow characteristic analysis unit, which is used to analyze the flow characteristic information in the flow data and calculate the flow characteristic anomaly coefficient of the flow data; A transmission characteristic analysis unit, which is used to analyze the transmission characteristic information in the flow data and calculate the abnormal coefficient of the flow data transmission characteristic; A structural feature analysis unit, which is used to analyze the structural feature information in the flow data and calculate the structural feature anomaly coefficient of the flow data; The model creation unit establishes a malicious traffic identification machine learning model, collects a historical training data set of network traffic data in advance, and trains the malicious traffic identification machine learning model based on the historical training data set; imports the currently acquired traffic feature anomaly coefficient, transmission feature anomaly coefficient, and structure feature anomaly coefficient into the malicious traffic identification machine learning model, and identifies malicious traffic through the malicious traffic identification machine learning model; The transmission characteristic analysis unit analyzes the transmission characteristic information in the flow data and calculates the abnormal coefficient of the flow data transmission characteristic as follows: Collect the periodic interval of each data packet in the data flow whose transmission speed exceeds the benchmark value; Calculate the abnormal coefficient of traffic data transmission characteristics based on the periodic interval of each data packet ; in Transmission characteristic anomaly coefficient The calculation formula is: ; in is the cycle interval time, The average value of the interval time between each data packet cycle; The collecting of the periodic intervals of each data packet in the data flow whose transmission speed exceeds the reference value specifically comprises the following steps: Collect the transmission speed of each data packet in the traffic data; Preset transmission speed benchmark values based on historical data during normal operation Compare the currently acquired transmission speed of each data packet with the transmission speed benchmark value, obtain the data packet whose transmission speed exceeds the transmission speed benchmark value, and extract the time node information of the corresponding data packet and record it as , where y∈(1, 2, 3, ...); Calculate the corresponding data packet cycle interval time based on the extracted data packet time node information Ty Cycle interval time The calculation formula is: , which is to extract the interval time between adjacent time nodes in the corresponding data packet.
2. A blockchain big data security processing system according to claim 1, characterized in that: The traffic characteristic information in the traffic data includes the number of data packets in the traffic data and the source and destination IP address information corresponding to each data packet; The transmission characteristic information in the flow data is the data packet transmission speed in the flow data; The structural characteristic information in the traffic data is the data length of the data packets in the traffic data.
3. A blockchain big data security processing system according to claim 1, characterized in that: The flow characteristic analysis unit analyzes the flow characteristic information in the flow data and calculates the flow characteristic abnormality coefficient of the flow data as follows: Count the number of packets with the same source and destination IP address information, the number of packets with the same source IP address information, and the number of packets with the same destination IP address information in the traffic data in turn; Calculate the flow characteristic anomaly coefficient Ks of the flow data, where the calculation formula of the flow characteristic anomaly coefficient Ks is: ; Where X is the total number of packets in the traffic data, m is the number of packets with the same source and destination IP address information, n is the number of packets with the same source IP address information but different destination IP address information, and i is the number of packets with the same destination IP address information but different source IP address information; Indicates the anomaly weight coefficient when the source and target IP address information are the same. Indicates the anomaly weight coefficient when the source IP address information is the same but the target IP address information is different. Indicates the anomaly weight coefficient when the target IP address information is the same but the source IP address information is different.
4. A blockchain big data security processing system according to claim 1, characterized in that: The structural feature analysis unit analyzes the structural feature information in the flow data and calculates the structural feature anomaly coefficient of the flow data as follows: Preset the packet length threshold based on historical data during normal operation Collect the data length of each data packet in the traffic data, compare the data length of each data packet currently obtained with the length threshold, and obtain the data packet whose data length exceeds the data length threshold; Collect the IP address information of each data packet in the traffic data and count the number of different IP address information; Collect IP address information of data packets with data length and length threshold, and count the number of different IP address information; The abnormal coefficient of structural characteristics of traffic data is calculated through integration analysis.
5. A blockchain big data security processing system according to claim 4, characterized in that: The flow data structure characteristic abnormal coefficient Kg is calculated by integration analysis, where the calculation formula of the structure characteristic abnormal coefficient Kg is: ; Among them, X is the total number of data packets in the traffic data, Q is the number of data packets whose data length exceeds the data length threshold, F is the number of different IP address information of the data packets in the traffic data; S is the number of different IP address information of the data packets whose data length exceeds the data length threshold.
6. A blockchain big data security processing system according to claim 1, characterized in that: The historical training data set includes several sets of training data collected from the historical traffic data of the blockchain, and each set of training data includes feature data and label data; The characteristic data includes a flow characteristic anomaly coefficient, a transmission characteristic anomaly coefficient and a structural characteristic anomaly coefficient; The label data is collected when each set of traffic data training data is collected. Through manual review and analysis, it is determined whether each set of traffic data is malicious traffic. If the traffic data is normal traffic, the label data is marked as 1. On the contrary, if the traffic data is malicious traffic, the label data is marked as 0.
7. A blockchain big data security processing system according to claim 6, characterized in that: The specific training of malicious traffic identification machine learning model based on historical training data set is as follows: Divide the historical training data set into 80% training set and 20% test set; Construct a malicious traffic identification machine learning model, and input the training set as input data into the malicious traffic identification machine learning model for training, so as to obtain a trained malicious traffic identification machine learning model; The test set is used to verify the accuracy of the trained malicious traffic identification machine learning model, and the malicious traffic identification machine learning model that meets the preset accuracy is output.
8. A blockchain big data security processing system according to claim 1, characterized in that: It also includes an alarm processing module. When the malicious traffic identification machine learning model identifies malicious traffic, the alarm processing module is triggered. The alarm module issues an alarm in a timely manner, quickly responds to malicious traffic attacks, and takes protective measures.
