Research and Application of Mobile Agent Invasion Detection Model Based on Data Mining

In the information era, the network has become not only a primary means of data exchange and business transaction but also attackers' major target. Therefore, an invasion detection technology which can discover and report non-authorized access or anomalies of system came into being. Traditional invasion detection system is inefficient and unsatisfactory in detecting ability, this article proposed a kind of mobile agent invasion detection model based on data mining algorithm, applied separately relating rule algorithm to exceptional detection and misuse detection, carried on analysis to new unknown behavior that existed in border zone of two kinds of detection pattern with cluster algorithm. However mobile agent can shift automatically to target system and collect only information relative to invasion, the system log is no longer transmitted to the server. The communication load of system is reduced enormously, and working efficiency is increased, The validity of this model has been proven by simulation experiment.