基于神经网络的DDoS防护绩效评估

发布时间：2018-11-11 00:01

【摘要】：面对日益严重的分布式拒绝服务(distributed denial of service,DDoS)攻击威胁和众多防护措施,需要防护绩效评估方法指导防护措施的选择.现有绩效评估方法通过对比防护措施部署前后的攻击效果进行评估,需对防护措施进行卸载及重新部署,实施成本高.针对这种不足,首先建立了防护绩效评估模型(defence evaluation model,DEM),该模型从用户感受角度进行指标选取,减少了传统方式下测评过程需要的指标数量,降低了数据获取的难度.利用神经网络良好的泛化能力,将其引入DDoS防护绩效评估过程;在计算已部署防护措施攻击效果的同时,预测得到未部署防护措施时的攻击效果,减少了测量次数.使用网络仿真程序SSFNet模拟典型攻击场景进行实验,验证了提出的评估方法以及神经网络的预测能力.
[Abstract]:In the face of the increasingly serious threat of distributed denial of service (distributed denial of service,DDoS) attacks and numerous protective measures, the selection of protective measures needs to be guided by the methods of performance evaluation. The existing performance evaluation methods evaluate the effectiveness of the attack before and after the deployment of protective measures. It is necessary to uninstall and redeploy the protective measures and the cost of implementation is high. In order to solve this problem, first of all, the model of protection performance evaluation (defence evaluation model,DEM) is established. The model selects indicators from the perspective of user perception, which reduces the number of indicators needed in the traditional evaluation process and reduces the difficulty of data acquisition. Using the good generalization ability of the neural network, it is introduced into the DDoS protection performance evaluation process, while calculating the attack effect of the deployed protective measures, predicting the attack effect when the protective measures are not deployed, and reducing the measurement times. The network simulation program SSFNet is used to simulate the typical attack scene. The proposed evaluation method and the prediction ability of the neural network are verified.
【作者单位】：可信计算与信息保障实验室(中国科学院软件研究所);信息网络安全公安部重点实验室(公安部第三研究所);
【基金】：国家“八六三”高技术研究发展计划基金项目(SQ2013GX02D01211,2011AA01A203) 国家自然科学基金项目(61100226,61303248) 北京市自然科学基金项目(4122085) “十二五”国家科技支撑计划基金项目(2012BAK26B01)
【分类号】：TP393.08

【相似文献】