云环境下基于SOM神经网络的入侵检测方法研究 Research on Intrusion Detection Method Based on SOM Neural Network in Cloud Environment

云安全已成为云计算发展过程中面临的重要挑战，基于云计算的入侵检测系统将成为云安全体系的重要组成部分。根据云计算特点和安全需求，设计了一种适合云环境的入侵检测系统模型，在入侵检测算法中引入SOM自组织特征映射神经网络算法，对SOM网络连接权值随机初始化可能导致的训练失败问题，采用基于模拟退火的微粒群算法对其进行优化，通过仿真实验验证优化算法可有效提高入侵检测性能。 Cloud security has become an important challenge in the development of cloud computing. The intrusion detection system based on cloud computing will be an important part of the cloud security system. According to the characteristics and security requirements of cloud computing, an intrusion detection system model is designed for the cloud environment, and the SOM self-organiz- ing feature map neural network algorithm is introduced into the intrusion detection algorithm. The random initialization of SOM network connection weights may lead to the failure of the training, so the particle swarm optimization algorithm based on simulated annealing is used to optimize the SOM neural network algorithm. The simulation experiment results show that the optimization algorithm can effectively improve the performance of intrusion detection.