Optimization of Large-Scale Intrusion Detection Algorithms for Digital Information Systems

In recent decades, especially in the past 10 years, network technology has made rapid progress, and also the accompanying network security issues have become more and more prominent. Malicious data flow, especially the identification and detection of high-bandwidth malicious data represented by the botnet DDoS attacks and worms is very important. In this study, the novel large-scale intrusion detection algorithms for digital information systems is proposed. To begin with, the distributed concerns are noted, and determine that core module of the intrusion detection experiment includes the user rule configuration, agent, intrusion detection engine, data collection and intrusion alarm. Then, the novel algorithm is designed. Local risk minimization method in the distributed learning stage to solve the local risk minimization problem. Then, the system enhancement suggestions are provided. Through the considered tests, the attack detection performance is tested.