摘要
根据Intel 80X86体系结构与栈缓冲区溢出的基本特点,从汇编语言程序的视角对缓冲区溢出攻击进行分类,分析了现有的动态防御机制所存在的优缺点,指出现有防御策略所存在的不足.提出一种基于返回地址变换的动态栈缓冲区溢出防御机制,理论分析和实验表明新机制能够极大概率对抗各种缓冲区溢出攻击.该机制能根据所部署系统的安全要求不同采用不同的方案,以满足不同的安全性和不同的效率要求.
In the paper,the basic attack patterns for Intel 80X86 are classified in the viewpoint of assemble language programs.The weak-ness of the existing dynamic buffer-overflow prevention mechanisms is discussed.A new dynamic stack buffer-overflow prevention mechanism based on return-address translation is proposed.The new mechanism is proved to be able to defend multiple patterns attacks with an acceptable performance tradeoff.We present experimental results of both the penetration resistance and the performance impact of the propsed mechanism.With simple modification,the mechanism is suitable for different security and performance needs.
出处
《小型微型计算机系统》
CSCD
北大核心
2012年第5期1012-1017,共6页
Journal of Chinese Computer Systems
基金
国家"八六三"高技术研究发展计划项目(2007AA01Z142)资助
