检测移动Agent系统中恶意攻击的改进方案

对一种检测恶意Agents攻击行为的异常检测模型作出了改进.主要通过分析移动Agent的移动模式及在不同主机上停留时间,以此来判断移动Agent的行为,并且把正常Agent的移动轨迹进行分类,并预测该Agent的下一目的地.最后,通过实验,对提出的算法进行评估,并分析实验结果.

A Reinforcement in Detecting Anomalous Agents in Mobile Agent System

Despite many benefits that mobile Agents can bring out, its users face security problem that is from malicious Agents and hosts. Though there exist some measures to protect mobile Agents from some attacks to some degree, mobile Agents could still be assaulted illegally while they are roaming from one host to another. In order to address this problem further, this paper improves an anomaly detection model for detecting malicious Agents. In particular, we focus on mobile Agent's activities by measuring its movement pattern and residence time on hosts in the original step. Moreover, we classify the normal Agent tracks and predict the next Agent's target. Finally we do some experiments to evaluate the prediction algorithm and analyze the results.