基于OPNET的BotNet最优步长传播仿真  

Simulation of optimal step size propagation of BotNet based on OPNET

作　　者：杨云[1] 胡谷雨[1] 李华波[1] 罗隽[2] 

机构地区：[1]解放军理工大学指挥自动化学院,江苏南京210007 [2]海军指挥学院,江苏南京211800

出　　处：《解放军理工大学学报（自然科学版）》2012年第4期383-387,共5页Journal of PLA University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60603029)

摘　　要：僵尸网络(BotNet)主要采用蠕虫扫描的方式进行传播,传统的蠕虫传播策略主要在扩散效率和扫描准确性上进行折衷,共性缺点是存在对同一主机重复扫描和网络之间交叉扫描的严重问题,会对互联网产生严重的流量冲击,降低BotNet的隐蔽性。为解决这一共性问题,根据互联网的无尺度特性,在对比传统网络蠕虫传播策略优缺点的基础上,结合前向神经网络现有的BP学习算法对BotNet传播进行了分析,从理论上提出一种计算BP近似最优步长的算法,并通过OPNET建立传播攻击模型进行了仿真验证。结果表明,该算法有效地提高了BotNet在无尺度网络中的传播性能。BotNet mainly uses worm scan strategies to propagate.Traditional strategies always make compromise design between diffusion efficiency and scan accuracy,and their general defects focus on the repeat scanning process,which imports serious Internet traffic problem and decreases the invisibility of BotNet.To solve the problem,a comparative analysis of the advantages and disadvantages of the traditional propagation strategies was made,the existing BP feedforward neural network learning algorithm was used to analyze the spreading of BotNet,and an algorithm proposed to calculate the approximate optimal step size of BP.An OPNET simulation shows that the algorithm can effectively improve the performance of the spreading of BotNet in scale-free network.

关 键 词：僵尸网络 神经网络 BP算法 OPNET 

分 类 号：TP393.4[自动化与计算机技术—计算机应用技术]