融合稀疏注意力机制在DDoS攻击检测中的应用
DDoS attack detection model based on sparse attention mechanism
王博 1万良 1叶金贤 1刘明盛 1孙菡迪1
作者信息
- 1. 贵州大学计算机科学与技术学院公共大数据国家重点实验室,贵州贵阳 550025
- 折叠
摘要
针对现有的DDoS(distributed denial of service)攻击检测模型面临大量数据时,呈现出检测效率低的问题.为适应当前网络环境,通过研究DDoS攻击检测模型、提取流量特征、计算攻击密度,提出一种基于融合稀疏注意力机制的DDoS攻击检测模型GVBNet(global variable block net),使用攻击密度自适应计算稀疏注意力.利用信息熵以及信息增益分析提取攻击流量的连续字节作为特征向量,通过构建基于GVBNet的网络模型在两种数据集上进行训练.实验结果表明,该方法具有良好的识别效果、检测速度以及抗干扰能力,在不同的环境下具有应用价值.
Abstract
The existing DDoS(distributed denial of service)attack detection model presents the problem of low detection efficiency when facing a large amount of data.To adapt to the current network environment,the DDoS attack detection model GVBNet(global variable block net)based on the fused sparse attention mechanism was proposed by studying the DDoS attack detection model,extracting the traffic features,and calculating the attack density,and the attack density was used to calculate the sparse attention adaptively.The continuous bytes of attack traffic were extracted as feature vectors using information entropy and infor-mation gain analysis,and the network model based on GVBNet constructed was trained on two data sets.Results of experiments show that the method has good recognition effects,detection speed,and anti-interference capability,has application value in dif-ferent environments.
关键词
分布式拒绝服务攻击/稀疏注意力机制/攻击密度/信息熵/信息增益/模型优化/攻击检测Key words
distributed denial-of-service attack/sparse attention mechanism/attack density/information entropy/information gain/model optimization/attack detection引用本文复制引用
基金项目
国家自然科学基金(62062020)
国家自然科学基金(62262004)
贵州省教育厅自然科学研究项目(黔教科2007015号)
出版年
2024
NETL
NSTL
维普
万方数据