Research On Intrusion Detection Technology Based On Cloud Computing

Intrusion detection system is a security mechanism which dynamically monitor and defensethe intrusion behavior of network and system. but with the explosive growth of network Intrusiondetection data, the traditional intrusion detection systems are faced with the problem of massivedata processing, intrusion detection systems can not meet the needs of real-time and effectiveness.Cloud computing can solve the bottleneck problem of intrusion detection data processing with itscomputing power.So the efficient cloud-based intrusion detection system is the current researchhotspot in intrusion detection areas.The main works of this paper include:(1) Analyse the major security issues of cloud computing, design a model of intrusion detectionsystem CIDS, and describe and design the main function modules and implementation techniques,then demonstrates the reasonableness of CIDS. The model can not only detect the threat of invasionfrom the cloud environment, but also can use the cloud computing technology to analyse andprocess datas.(2) Propose a detection method PPKM based on CIDS detecting module. It takes the principalcomponent analysis PCA as the feature extraction module for data dimensionality reduction, anduse the particle swarm optimization (PSO) to search the K-means clustering centers for globaloptimization. Experiments show that PPKM (PCA+PSO+K-means) improve the analysis engineperformance, the algorithm has a high detection rate and detection effects.(3) In order to meet the cloud computing intrusion detection system real-time requirements, weimplement PPKM on Hadoop cloud platform for parallel processing. Experiments show that,PPKM algorithm significantly reduces training and analyzing time.It verifies the real-time andeffectiveness of proposed cloud-based intrusion detection system.