| With the rapid development of internet technology and distributed computing, clo udcomputing has become a mature network technology. It integrates a lot of computation resourcesand storage resources together to form a shared virtual computing resource, which provideson-demand services for computer users. There are an increasing number of safety problems withthe popularity of the cloud service. How to strengthen the information security of cloud serviceand how to ensure the internet resource not be invaded or destructed maliciously by hackers havebecome important issues which should be solved in the development of cloud computing.However, cloud computing is a heterogeneous virtual environment, traditional and singleintrusion detection system cannot be directly applied in cloud computing. So, it’s necessary toresearch on the intrusion detection technology in cloud computing environment to guarantee theinformation security of cloud service.In this paper, first of all, we summarize the definition and characteristics of cloudcomputing, analyze the security threats facing the cloud environment. Meanwhile, we reviewmeans of attack in the environment. Then, combining the characteristic of intrusion detectionsystem and the features of cloud computing, we design a kind of model of intrusion detectionsystem based on the BP neural network in cloud computing environment. The research workscan be stated as follows:(1)The paper analyzes safety hidden danger and means of attack in cloud computingenvironment, also research on characteristics of the current intrusion detection technology. wemodels a distributed intrusion detection system which can adapt in cloud computing environmentaccording to the features of cloud computing and basis of security needs. This new technology,which integrates network intrusion system and host intrusion system, adopts multi-agentcoordination mechanism to detect attack behavior in the cloud service. It provides a safeguardfor the currently prevalent cloud computing service.(2) Due to the huge amounts of data needed to handle and varieties of attack behavio r incloud computing environment, Through the research of the BP neural network for its strongself-study and adaptability, we adopts it to the detection module in the model of intrusiondetection for cloud computing. BP neural network could detect the attack behavior timely andsuits to the changing network environment. Moreover, the nonlinear processing capability of BPneural network can handle the complex and random intrusion types.(3) In cloud computing environment, BP neural network need to train the historical data toachieve testing target when we utilize it for intrusion detection, as there are so much complexattack behavior to be handled, the BP neural network needs to spend a lot of training time. Dueto the neural network based on gradient descent method for sample data training, it is easy to fallinto local minimum value and can not achieve the global optimal value, which affects theefficiency of the system detection and the security of cloud service. So, this paper analyzes theartificial bee colony algorithm. According to the global search ability of the artificial bee colony algorithm, we combined it with the BP neural network. And we use the improved artificial beecolony algorithm to optimize the BP neural network, which can improve classified accuracy andthe accuracy of network intrusion detection. Finally, we conduct simulation experiments toverify the high effectiveness of the algorithm. |
