分布式深度学习框架下主动式网络安全性能感知系统设计

Design of active network security performance awareness system under distributed deep learning framework

主动式网络在数据传输过程中容易受到黑客的攻击,威胁网络整体安全性能,为此,在分布式深度学习框架下设计主动式网络安全性能感知系统。基于分布式理论加速训练深度学习算法获取最佳参数集合,应用训练好的深度学习算法——卷积自动编码器提取安全性能感知数据的重要特征,以此为基础,分析影响网络安全性能的关键因素。选取感知指标,并应用模糊层次分析法计算其权重数值。构建网络安全性能感知模型,制定网络安全性能等级判定规则,即可确定网络安全性能感知结果。实验数据显示,设计系统应用后主动式网络安全性能感知误差最小值为0.4%,保障了主动式网络安全运行。

In the process of data transmission,active network is vulnerable to attack by hackers,threatening the overall security performance of the network.Therefore,active network security performance awareness system is designed under the framework of distributed deep learning.Based on the distributed theory,accelerate the training of the deep learning algorithm to obtain the best parameter set,and apply the trained deep learning algorithm-convolution automatic encoder to extract the important features of the security performance perception data.On this basis,analyze the key factors affecting the network security performance.Select the perception index,and calculate its weight value using fuzzy analytic hierarchy process.The network security performance perception results can be determined by constructing the network security performance perception model and formulating the network security performance level judgment rules.The experimental data shows that the minimum perceived error of the active network security performance after the application of the design system is 0.4%,which ensures the safe operation of the active network.