摘要
拟态架构采用多模异构冗余的思想,可有效防御针对系统已知或未知漏洞/后门的攻击。在拟态架构中超时参数是一个重要的参数,既要满足实时性要求,又要使大多数正常结果能够返回,为此提出差分超时参数预测算法,既避免统一定时不够灵活的弊端,又能自适应地解决异构执行体间的差异,并采用SVM回归方法计算超时参数并以此作为执行体超时基准,实验结果表明,该方法对拟态架构中的执行体超时参数预测优秀率为96%,良好率为4%,远远高出KNN和指数回归算法。
Mimic architecture adopts the idea of multi-mode heterogeneous redundancy,which can effectively defend against attacks on known or unknown vulnerabilities/backdoors.In the mimicry architecture,the timeout parameter plays an important role,not only in meeting the requirement of real time,but also in ensuring most results return by the time.Therefore,this paper proposes a differential timeout parameter prediction algorithm,to avoid the inflexibility of the unifying timing,and to solve the difference between the isomeric executors as well.The SVM regression algorithm is used to predict the parameters.The experimental results show that the excellent and good rates are 96%and 4%,far better than KNN and exponential regress algorithm.
作者
宋克
欧阳玲
魏帅
鲁晓彬
SONG Ke;OU Yangling;WEI Shuai;LU Xiaobin(Information Engineering University, Zhengzhou 450001, China;Zhongyuan University of Technology, Zhengzhou 450007, China)
出处
《信息工程大学学报》
2020年第4期470-475,共6页
Journal of Information Engineering University
基金
国家核高基重大专项资助项目(2017ZX01030301)。
关键词
差分超时参数
SVM回归
拟态架构
预测算法
differential time-out parameter
SVM regression
mimic architecture
prediction algorithm
