面向实时网络状态适配的动态网络安全防御方法及系统

By constructing state vector sequences and risk trend prediction models, the optimal defense strategy is dynamically selected, solving the problems of lag and static configuration in traditional network defense systems. This achieves real-time performance and intelligent adaptation of the network defense system, thereby improving the defense effectiveness.

CN121239489BActive Publication Date: 2026-07-17GUANGDONG ZHUOYUE ZHIYUN INFORMATION ENGINEERING CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUANGDONG ZHUOYUE ZHIYUN INFORMATION ENGINEERING CO LTD
Filing Date
2025-11-12
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Traditional network defense systems lack real-time perception mechanisms and adaptive adjustment capabilities, making it difficult to cope with advanced persistent threats, lateral movement attacks, and dynamic changes in network status. This results in a disconnect between defense actions and the network risk situation, and a lack of quantitative description of defense effectiveness and proactive strategy capabilities.

Method used

By collecting key network indicators to construct a state vector sequence, using a time series prediction model with an embedded structured attention mechanism to predict attack risk trends, and combining the current state with the risk vector, the optimal strategy path is dynamically selected in the defense action graph, and standardized control commands are generated to achieve real-time network state adaptation.

Benefits of technology

It enables real-time quantitative representation of network status and forward-looking identification of potential attacks, dynamically selects the optimal defense strategy, improves the real-time performance, controllability and intelligent adaptability of the network defense system, and avoids defense redundancy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121239489B_ABST
    Figure CN121239489B_ABST
Patent Text Reader

Abstract

本发明提出面向实时网络状态适配的动态网络安全防御方法及系统，方法包括：通过采集网络流量速率、异常连接比、拓扑变动频率和协议分布偏离度等关键指标，构建标准化的状态向量序列；基于该序列，利用嵌入结构化注意机制的时间序列预测模型，前瞻性预测未来多个时间窗口的攻击风险趋势；结合当前状态与风险趋势，在预定义的防御动作图谱中动态选择最优策略路径，并将其映射为设备可执行的标准化控制命令，按优先级调度下发执行。本发明实现了从状态感知、风险预测到策略生成与执行的闭环自适应防御，显著提升了网络安全系统的实时性、智能性与工程可部署性。
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Network security dynamic early warning method and system based on knowledge graph

    CN119788344A

  • Network space security situation awareness detection analysis system and method

    CN120710721A