摘要
漏洞挖掘是解决当前工业控制系统安全问题的有效途径.本文分析了现有工业控制平台漏洞挖掘的问题,提出了一种改进的Fuzzing架构.通过引入置信度的概念对测试用例进行量化,将其作为分类器的输入进行分类,从而预先筛选出可能有效的测试用例,实现了减少输入空间、增加命中率的目的.基于该架构设计的针对工业控制系统的一种通用漏洞挖掘框架,实现了集畸形数据构造、测试目标监控和测试结果管理为一体,并同时支持多目标、多协议、多平台的扩展.通过对某款工业控制器进行实际测试,证明了该架构的可行性与高效性.
Vulnerability detection is an effective way to solve security problem of current industrial control system.By analyzing the difficulties of vulnerability detection in the existing industrial control platform,this paper proposes an improved fuzzing framework that introduces the concept of confidence to quantify the test cases as a classifier input,and thus pre-screens potential test cases,so as to reduce input space and enhance hit rate.Based on this architecture design for industrial control systems,the generic framework for vulnerability mining combines with malformed data structure,test target monitoring,and test results management,and supports multi-target,multi-protocol,multiplatform extensions.Finally,experimental results on an industrial controller have shown the feasibility and effectiveness of the method.
出处
《武汉大学学报(理学版)》
CAS
CSCD
北大核心
2013年第5期411-415,共5页
Journal of Wuhan University:Natural Science Edition
基金
国家自然科学基金(61173138
61272452)
国家自然科学基金青年基金(61003268)
湖北省重点新产品新工艺研究开发项目(2012BAA03004)资助项目