A computer network security monitoring system and method

By combining multi-source data acquisition and intelligent analysis modules with machine learning models and rule engines, the problem of insufficient cross-dimensional risk monitoring in existing technologies has been solved. This has enabled high-precision risk identification and rapid response, adapting to diverse network environments and reducing false alarm and false negative rates and security losses.

CN123122854APending Publication Date: 2026-11-17张昊
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610249725.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-03
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing network security monitoring technologies are unable to fully capture cross-dimensional security risks, have weak capabilities in identifying unknown attacks, have high false alarm and false negative rates, cannot adapt to cloud-native environments, and are disconnected from early warning and response processes, leading to increased security losses.

Method used

The intelligent analysis module, which combines multi-source data acquisition, machine learning models, and rule engines, enables multi-dimensional monitoring and risk identification, tiered early warning and automatic emergency response, supports dynamic updates and manual intervention, and encrypts data storage to ensure data integrity.

Benefits of technology

It achieves strong multi-dimensional monitoring capabilities, high identification accuracy, high response efficiency, and wide applicability, reducing false alarm and missed alarm rates, shortening risk response time, and improving network security protection capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122854A_ABST
    Figure CN123122854A_ABST
Patent Text Reader

Abstract

This invention discloses a computer network security monitoring system and method, belonging to the field of network security technology. The system includes modules for data acquisition, preprocessing, intelligent analysis, risk warning, emergency response, and storage. The data acquisition module collects data from multiple sources, including network traffic, device logs, and terminal behavior. The preprocessing module performs data cleaning, deduplication, format standardization, and feature extraction. The intelligent analysis module integrates machine learning algorithms and a rule engine to construct a multi-dimensional anomaly identification model, detecting malicious attacks, unauthorized access, and other risks in real time and dynamically optimizing parameters. The risk warning module pushes tiered warnings according to risk levels, the emergency response module automatically performs isolation and interception operations, and the storage module retains data for review. The method corresponds to the system to achieve the above-mentioned full-process monitoring. This invention improves the comprehensiveness and accuracy of monitoring, shortens response time, provides full-chain protection for network security, and is applicable to multiple scenarios such as enterprise LANs and cloud computing platforms, demonstrating significant practicality.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of computer network security technology, and more specifically, to a computer network security monitoring system and method. Background Technology

[0002] With the deepening of digital transformation, computer networks have become a core support for enterprise operations, government processing, and public services. However, their openness and complexity have led to a continuous rise in cybersecurity risks, with frequent security incidents such as malicious attacks, data breaches, and unauthorized access, seriously threatening data privacy and system stability. Especially in scenarios where cloud computing, the Internet of Things, and edge computing are integrated, network boundaries are becoming increasingly blurred, and cross-terminal and cross-platform security risks are becoming more prominent, posing a severe challenge to the adaptability and protection capabilities of traditional monitoring technologies.

[0003] Existing network security monitoring technologies have many limitations: some systems focus on only a single monitoring dimension, such as monitoring only network traffic or device logs, making it difficult to comprehensively capture cross-dimensional security risks. For example, they cannot correlate abnormal terminal behavior with abnormal cloud data access risks. Some rely on fixed rules to identify risks, resulting in weak identification capabilities for unknown and mutated attacks, high false positive and false negative rates, and rule updates lag behind the iteration of attack methods, making it difficult to cope with new threats such as ransomware and APT attacks. At the same time, most systems have a disconnect between early warning and response, requiring manual intervention after a risk is detected, leading to delayed response, increased security losses, and a lack of a sound risk tracing mechanism, making it difficult to accurately locate the source of the attack and its propagation path. In addition, most existing systems are designed for traditional LANs and lack adaptability to cloud-native environments and hybrid cloud architectures. The monitoring scope and accuracy are insufficient to meet the needs of complex network environments, and they cannot achieve unified management and control of distributed nodes.

[0004] Therefore, there is an urgent need for a network security monitoring system and method with multi-dimensional monitoring, intelligent identification, and rapid response capabilities to improve the initiative and effectiveness of network security protection and adapt to the security needs of diversified network architectures. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing network security monitoring systems and provide a computer network security monitoring system and method that enables multi-source data fusion monitoring, intelligent and accurate risk identification, hierarchical early warning and automatic emergency response, reduces false alarm and missed alarm rates, shortens response time, and provides full-chain protection for network security.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A computer network security monitoring system includes a data acquisition module, a data preprocessing module, an intelligent analysis module, a risk warning module, an emergency response module, and a storage module; The data acquisition module adopts a distributed deployment and uses network sniffing technology, API interface calls and log capture tools to collect network traffic data, device logs, user behavior data and cloud resource operation records from multiple sources in real time, and supports dynamic adjustment of the collection frequency; The data preprocessing module is used to clean, standardize and convert the format of the collected data and extract features to generate a standardized feature dataset. The intelligent analysis module integrates machine learning models and rule engines to comprehensively analyze feature datasets, determine risk types and levels, and dynamically update identification rules and model parameters. The risk warning module is preset with three levels of risk standards: high, medium and low. It generates warning information based on the judgment results and pushes it to designated maintenance personnel through multiple channels. The emergency response module is linked with the risk warning module, automatically executes response operations according to a preset hierarchical strategy, supports manual intervention and adjustment, and records response logs. The storage module uses a distributed database to store all process data and has data encryption, access control, and fast retrieval functions.

[0007] Furthermore, the data acquisition module supports wired, wireless, and hybrid network environments, and the acquired data covers the network layer, transmission layer, application layer, terminal, and user behavior dimensions.

[0008] Furthermore, the data preprocessing module uses the Pandas library for data cleaning, standardizes the data using JSON format, and extracts data features using the TF-IDF algorithm and statistical analysis.

[0009] Furthermore, the machine learning model in the intelligent analysis module is a random forest model, which is trained and optimized through security event samples. The rule engine integrates industry security standards and custom rules, and the analysis results of the two are combined to determine the risk according to preset weights.

[0010] Furthermore, the risk warning module's push channels include SMS, email, system pop-ups, WeChat Work, and DingTalk interfaces, prioritizing the push of high-level risk information.

[0011] Furthermore, the emergency response module's tiered strategy is as follows: for low-risk situations, log marking and enhanced traffic monitoring are implemented; for medium-risk situations, access permission blocking, terminal isolation, and traffic interception are implemented; and for high-risk situations, core system switching, network partition isolation, and data backup are implemented.

[0012] Furthermore, the storage module uses a Hadoop distributed database, and the data encryption uses the AES-256 encryption algorithm, supporting fast retrieval by time, risk type, and device number.

[0013] A computer network security monitoring method, implemented based on the system described in any one of claims 1-7, includes the following steps: S1: The data acquisition module collects network and terminal-related data from multiple sources in real time and transmits it to the data preprocessing module; S2: The data preprocessing module cleans, standardizes, and extracts features from the collected data to generate a standardized feature dataset. S3: The intelligent analysis module calls machine learning models and rule engines to comprehensively analyze feature datasets and determine risk types and levels; S4: The risk warning module generates warning information based on the risk level and pushes it to maintenance personnel through multiple channels; S5: The emergency response module automatically executes response operations according to preset strategies, records the response process, and supports manual intervention; S6: The storage module encrypts and stores all process data for review, traceability, and system optimization.

[0014] Furthermore, in step S1, under abnormal circumstances, the data acquisition module automatically increases the acquisition frequency, with the highest acquisition frequency not less than 10 times / second.

[0015] Furthermore, in step S3, the machine learning model is periodically iterated and optimized using stored historical data, and the security rules of the rule engine are updated, with the optimization cycle not exceeding one month.

[0016] Compared with the prior art, the present invention has the following advantages: 1. Strong multi-dimensional monitoring capabilities: It adopts a multi-source data collection method, covering the network layer, transmission layer, application layer, terminal and user behavior, to comprehensively capture security risks and avoid the limitations of single-dimensional monitoring; 2. High identification accuracy: It integrates machine learning models and rule engines, which can quickly identify known risks and accurately capture unknown and mutated attacks. It also reduces false positives and false negatives through dynamic parameter optimization. 3. High response efficiency: It realizes the linkage between early warning and response, and automatically executes response operations at different levels, which significantly shortens the risk response time, reduces the losses caused by security incidents, and supports manual intervention to improve the flexibility of response; 4. Wide applicability: Supports various network environments such as wired, wireless, and cloud computing, and can be adapted to the network security monitoring needs of different industries such as enterprises, government, and finance, with strong scalability. Attached Figure Description

[0017] Figure 1 This is a structural block diagram of the computer network security monitoring system of the present invention; Figure 2 This is a flowchart of the computer network security monitoring method of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] This invention provides a computer network security monitoring system, comprising: Data acquisition module: Employing a distributed deployment approach, this system utilizes network sniffing technology, API calls, and log capture tools to collect real-time network layer traffic data, transport layer session data, application layer access logs, terminal device operation logs, user behavior data, and cloud resource operation records from multiple sources. It supports data collection in wired, wireless, and hybrid network environments, with dynamically adjustable collection frequency. The network sniffing technology is compatible with mainstream protocols such as TCP / IP, HTTP, and HTTPS, accurately capturing key information such as source and destination IP addresses, port numbers, and data payloads of data packets. The API interface supports integration with mainstream cloud service providers (Alibaba Cloud, Tencent Cloud, Huawei Cloud) and third-party security devices, enabling seamless collection from heterogeneous data sources. A built-in data verification mechanism ensures the integrity of the collected data and prevents data loss due to network fluctuations. Data preprocessing module: The received multi-source data is cleaned to remove redundant, noisy, and invalid data; heterogeneous data undergoes format standardization conversion to unify data encoding and field definitions; key features in the data, including traffic features, behavioral features, and log features, are extracted using feature extraction algorithms to generate a standardized feature dataset to support subsequent analysis; specifically, data cleaning uses an outlier detection algorithm (Z-score algorithm) to identify and remove data that deviates from the normal range, and a data deduplication algorithm to remove duplicate records; the feature extraction stage combines statistical and semantic features, with traffic features covering average packet size, transmission rate, connection duration, etc., behavioral features including user login frequency, operation permission changes, file transfer paths, etc., and log features are extracted in a structured manner through text segmentation and feature encoding to improve the efficiency of subsequent analysis; Intelligent Analysis Module: It features a built-in machine learning model and rule engine. The machine learning model is iteratively optimized through training samples, enabling it to identify unknown attacks, such as malicious code propagation, DDoS attacks, SQL injection, and brute-force attacks. The rule engine integrates industry security standards and custom rules to quickly match and identify known risks. The module merges the analysis results from both, comprehensively determines the risk level through a weighted allocation algorithm, and dynamically updates the identification rules and model parameters to reduce the false positive rate. The machine learning model adopts a fusion architecture of random forest and neural network, balancing identification accuracy and computational efficiency. It absorbs new security event samples in real time through an online learning mechanism, enabling dynamic iteration of model parameters. The rule engine supports visual rule configuration, allowing operations and maintenance personnel to add and modify rules according to business needs. Rule matching uses a multi-pattern matching algorithm to improve the identification speed of known risks, and the weights of both can be dynamically adjusted according to actual application scenarios to balance the identification priority of unknown and known risks. Risk warning module: The system pre-defines high, medium, and low risk levels. Based on the intelligent analysis module's assessment, it generates corresponding warning messages, including the risk type, location, scope of impact, potential losses, and preliminary handling suggestions. Warning messages are pushed through multiple channels such as SMS, email, system pop-ups, and WeChat / DingTalk interfaces, prioritizing high-level risks to designated maintenance personnel. The module incorporates a built-in warning priority ranking mechanism; when multiple risk events occur simultaneously, they are ranked according to both risk level and the importance of their impact on business operations, ensuring that high-level risks related to core business are handled first. It also supports traceability of warning information, recording push time, recipients, and read status for easy maintenance management and accountability. Emergency Response Module: In conjunction with the early warning module, a tiered handling strategy is preset. For low-risk situations, log marking and enhanced traffic monitoring are automatically executed. For medium-risk situations, temporary access restrictions, abnormal terminal isolation, and suspicious traffic interception are automatically executed. For high-risk situations, while pushing out early warnings, core business system switching, network partition isolation, and data backup are automatically executed. Operation and maintenance personnel can manually intervene to adjust the handling strategy and record the entire process handling log. For temporary blocking operations, the blocking duration and unblocking conditions can be set, and the blocking time can be manually unblocked in advance or extended according to the progress of risk handling. The core business system switching adopts a hot standby switching mechanism to ensure that the business interruption time does not exceed 1 second. Data backup adopts a combination of incremental backup and full backup, which saves storage resources and ensures data recoverability. A visual operation interface is provided for manual intervention, which supports real-time adjustment of handling strategies and simultaneous recording of operation logs. Storage module: The system employs a distributed database to store collected raw data, preprocessed feature data, analysis results, early warning information, and handling logs, supporting efficient reading and writing of massive amounts of data and long-term retention. It features data encryption and access control functions to ensure the security and integrity of stored data, while also supporting rapid data retrieval by time, risk type, device number, and other dimensions, providing support for security incident review and tracing. The database adopts a master-slave architecture, with the master node responsible for data writing and the slave nodes responsible for data reading and backup, improving data access efficiency and disaster recovery capabilities. Data encryption uses a dual mechanism of transmission encryption (SSL / TLS protocol) and storage encryption (AES-256 algorithm). Access control is based on the RBAC permission model, refining access permissions for different roles to prevent data leakage. It also supports data lifecycle management, automatically cleaning up expired and redundant data and optimizing storage resource usage.

[0020] This invention also provides a computer network security monitoring method, based on the above system, comprising the following steps: S1: During the data acquisition phase, the data acquisition module collects relevant data from the network and terminals in real time from multiple sources according to a preset frequency and range, and transmits it to the data preprocessing module. S2: Data preprocessing stage, the preprocessing module cleans, standardizes, and extracts features from the collected data to generate a standardized feature dataset; S3: Intelligent Analysis Stage. The intelligent analysis module calls machine learning models and rule engines to conduct comprehensive analysis of the feature dataset and determine the risk type and level. S4: Risk warning phase, the risk warning module generates warning information based on the risk level and pushes it to maintenance personnel through multiple channels; S5: During the emergency response phase, the emergency response module automatically executes the corresponding response operations according to the preset strategy, records the response process, and supports manual intervention. S6: During the data retention phase, the storage module encrypts and stores all data throughout the process for subsequent review, tracing, and system optimization.

[0021] The present invention will be further described in detail below with reference to specific embodiments, so that those skilled in the art can understand it.

[0022] This embodiment provides a computer network security monitoring system and method, applied to an enterprise LAN security monitoring scenario. The enterprise LAN includes 500 terminal devices, 10 network switches, 5 servers, and cloud storage resources. The system deployment adopts a distributed architecture. The data acquisition module deploys acquisition nodes on each switch, server, and core terminal. The core terminals are selected from devices in key departments such as enterprise R&D and finance to ensure full collection of core business data. TCP / UDP traffic data is collected through network sniffing tools, server logs and cloud resource operation records are called through API interfaces, and user behavior and device operation logs are collected through terminal agent programs. The collection frequency is set to 1 time / second, supporting high-frequency collection (up to 10 times / second) under abnormal conditions. The acquisition nodes have the function of resuming interrupted transmission. When the network is interrupted and restored, the missed data can be automatically re-collected to ensure the continuity of the collected data.

[0023] The data preprocessing module is implemented using the Python programming language and runs on the Ubuntu 20.04 operating system. Data cleaning is performed using the Pandas library to remove null values, duplicate data, and data with incorrect formatting. For semi-structured log data, regular expressions are used to extract key fields to ensure data structure. JSON format is used for data standardization, unifying the definitions of fields such as "attack type," "device IP," and "operation time." Operation time is uniformly formatted in UTC to avoid data corruption caused by time zone differences. The TF-IDF algorithm is used to extract log text features, and statistical analysis is used to extract traffic features (packet size, transmission frequency). After generating the feature dataset, Min-Max normalization is used to map the feature values ​​to the [0,1] interval, improving the training efficiency and recognition accuracy of the machine learning model. The intelligent analysis module incorporates a machine learning model based on random forests, trained with 100,000 security event samples, covering typical attack samples from multiple industries including finance, government, and enterprises. The model contains 100 decision trees with a depth of 15 layers, achieving an accuracy rate of 98.5%. False positives are mainly concentrated in low-traffic mutation attack scenarios. The rule engine integrates OWASP Top 10 security risk rules and enterprise-defined rules (such as prohibiting remote access to core servers outside of working hours and limiting the number of failed login attempts per IP address to no more than 5 per day). The analysis results of the two rules are weighted at 60% and 40% respectively, comprehensively determining the risk level. When a terminal is detected attempting to brute-force the server password outside of working hours (10:00 PM to 7:00 AM the next day) and fails to log in 8 times within 10 minutes, both the model and the rule engine trigger risk identification, comprehensively determining it as a medium-risk level and marking the terminal as a key monitoring target. The risk warning module immediately generates a warning message, including the risk type (brute force attack), terminal IP, server address, risk level (medium), number of attack attempts, and current terminal operating status. This message is pushed to the operations and maintenance personnel group via WeChat, and a system pop-up alert is simultaneously displayed to on-duty personnel. The pop-up is set to automatically call the personnel if the message is not read within 30 seconds. The emergency response module automatically performs a temporary ban on the terminal IP (ban duration 2 hours) and restricts server access permissions. Simultaneously, it mirrors the terminal's traffic to the analysis server for in-depth analysis and records the handling logs. After receiving the warning, operations and maintenance personnel can view the terminal's historical operation trajectory, installed software list, and network connection records through the system's visual interface. If a malicious attack is confirmed, the ban duration can be extended to 24 hours, and the physical location and network link of the attacking terminal can be traced using the source tracing tool. If it is an employee's mistake, the ban can be manually lifted after verifying the employee's identity, and the employee's terminal can be added to the trusted list to reduce the probability of subsequent false alarms. The storage module uses a Hadoop distributed database with a storage capacity of 10TB. The cluster consists of 5 nodes: 3 primary nodes and 2 backup nodes, ensuring no data loss in case of hardware failure. All data is encrypted using the AES-256 encryption algorithm, with the key kept by a designated person and changed regularly. Hierarchical access permissions are set for maintenance personnel: only administrators can view the full data, while ordinary maintenance personnel can only access logs and alerts within their scope of responsibility. During system operation, the machine learning model is iteratively optimized regularly (once a month) using stored historical data, and the security rules of the rule engine are updated. In addition, log audits and system self-checks are performed weekly to identify and investigate anomalies in the operation of data collection nodes, databases, and analysis modules, continuously improving monitoring capabilities and system stability. In this embodiment, the system operates stably and can accurately identify security risks such as brute-force attacks, SQL injection, and malicious code propagation, with a false alarm rate of less than 1.5% and a response time of no more than 3 seconds for medium- and high-level risks. After three months of trial operation, it successfully intercepted 12 malicious attack events, including 3 APT attacks, 5 brute-force attacks, and 4 SQL injection attacks, effectively preventing the risk of core data leakage and system paralysis. At the same time, the system supports integration with the enterprise's existing security management platform, enabling synchronous sharing of monitoring data and handling results without the need for additional reconstruction of the existing security architecture, reducing the enterprise's deployment and learning costs, fully adapting to the enterprise's actual operation and maintenance needs, and effectively ensuring the secure and stable operation of the enterprise's local area network.

[0024] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A computer network security monitoring system, characterized in that: It includes a data acquisition module, a data preprocessing module, an intelligent analysis module, a risk warning module, an emergency response module, and a storage module; The data acquisition module adopts a distributed deployment and uses network sniffing technology, API interface calls and log capture tools to collect network traffic data, device logs, user behavior data and cloud resource operation records from multiple sources in real time, and supports dynamic adjustment of the collection frequency; The data preprocessing module is used to clean, standardize and convert the format of the collected data and extract features to generate a standardized feature dataset. The intelligent analysis module integrates machine learning models and rule engines to comprehensively analyze feature datasets, determine risk types and levels, and dynamically update identification rules and model parameters. The risk warning module is preset with three levels of risk standards: high, medium and low. It generates warning information based on the judgment results and pushes it to designated maintenance personnel through multiple channels. The emergency response module is linked with the risk warning module, automatically executes response operations according to a preset hierarchical strategy, supports manual intervention and adjustment, and records response logs. The storage module uses a distributed database to store all process data and has data encryption, access control, and fast retrieval functions.

2. The computer network security monitoring system according to claim 1, characterized in that, The data acquisition module supports wired, wireless, and hybrid network environments, and the acquired data covers the network layer, transmission layer, application layer, terminal, and user behavior dimensions.

3. The computer network security monitoring system according to claim 1, characterized in that, The data preprocessing module uses the Pandas library for data cleaning, standardizes the data using JSON format, and extracts data features using the TF-IDF algorithm and statistical analysis.

4. The computer network security monitoring system according to claim 1, characterized in that, The machine learning model in the intelligent analysis module is a random forest model, which is trained and optimized through security event samples. The rule engine integrates industry security standards and custom rules, and the analysis results of the two are combined to determine the risk according to preset weights.

5. The computer network security monitoring system according to claim 1, characterized in that, The risk warning module pushes information through SMS, email, system pop-ups, WeChat Work, and DingTalk interfaces, prioritizing the push of high-level risk information.

6. The computer network security monitoring system according to claim 1, characterized in that, The emergency response module employs a tiered strategy: for low-risk situations, log marking and enhanced traffic monitoring are implemented; for medium-risk situations, access permissions are blocked, terminals are isolated, and traffic is intercepted; for high-risk situations, core system switching, network partitioning and isolation, and data backup are implemented.

7. The computer network security monitoring system according to claim 1, characterized in that, The storage module uses a Hadoop distributed database, and the data encryption uses the AES-256 encryption algorithm, supporting fast retrieval by time, risk type, and device number.

8. A computer network security monitoring method, implemented based on the system described in any one of claims 1-7, characterized in that, Includes the following steps: S1: The data acquisition module collects network and terminal-related data from multiple sources in real time and transmits it to the data preprocessing module; S2: The data preprocessing module cleans, standardizes, and extracts features from the collected data to generate a standardized feature dataset. S3: The intelligent analysis module calls machine learning models and rule engines to comprehensively analyze feature datasets and determine risk types and levels; S4: The risk warning module generates warning information based on the risk level and pushes it to maintenance personnel through multiple channels; S5: The emergency response module automatically executes response operations according to preset strategies, records the response process, and supports manual intervention; S6: The storage module encrypts and stores all process data for review, traceability, and system optimization.

9. The computer network security monitoring method according to claim 8, characterized in that, In step S1, under abnormal circumstances, the data acquisition module automatically increases the acquisition frequency, with the highest acquisition frequency not less than 10 times / second.

10. The computer network security monitoring method according to claim 8, characterized in that, In step S3, the machine learning model is periodically iterated and optimized using stored historical data, and the security rules of the rule engine are updated. The optimization cycle does not exceed one month.