Research And Implementation Of Instruction Detection System Based On Cloud Architecture

IDS(Intrusion Detection System)is a kind of security system which monitoring networktransmission, System log, and user activity. IDS were used to discover illegal users’ access,suspicious Intrusion behavior. Different with other passive network security system, IDS is akind of active safety technology.In view of the overload, unsupported of multicomputer conjunction analysis and maintenance of huge rule database of traditional IDS, this paper proposes a cloud architecture of intrusion detection system, improved the IDS in both architecture and algorithm.Two parts were contained of cloud architecture IDS: cloud site and use site. The network data was captured and preprocessed in user site and the intermediate results was send to thecloud as samples. User site contains data capture module, data analysis module, data preprocess module and data synthesize module. Data detection server cluster, cloud monitor server cluster, database server cluster composed the cloud site.Data were captured and preprocessed as sample space. KSISVM algorithm was used toanalysis the samples in cloud site to generate the detection result.KDDCUP99data were the research object in the simulation experiment; Experimentalresults show the prediction accuracy, sample select ability of KSISVM improves the IDS accuracy. And the concurrencies of KSISVM processes conjunction analysis reduce the time cost quickly with stable accuracy. The KS-ISVM is suitable for cloud architecture IDS.