| 基于即时编译的动态污点跟踪优化 |
| |
| 引用本文： | 吴泽智,陈性元,杨智,杜学绘.基于即时编译的动态污点跟踪优化[J].软件学报,2017,28(8):2064-2079. |
| |
| 作者姓名： | 吴泽智 陈性元 杨智 杜学绘 |
| |
| 作者单位： | 解放军信息工程大学, 河南 郑州 450001;密码科学技术国家重点实验室, 北京 100094,解放军信息工程大学, 河南 郑州 450001;密码科学技术国家重点实验室, 北京 100094,解放军信息工程大学, 河南 郑州 450001;河南省信息安全重点实验室, 河南 郑州 450001,解放军信息工程大学, 河南 郑州 450001;河南省信息安全重点实验室, 河南 郑州 450001 |
| |
| 基金项目： | 国家自然科学基金（61402437）；国家高技术研究发展计划（863）（2015AA016006，2012AA012704） |
| |
| 摘 要： | 动态污点跟踪技术展现了在移动隐私保护方面强大功能，但存在系统性能较低问题.提出了一个基于即时编译的动态污点传播优化方法.首先，将程序逻辑精确抽象为污点传播逻辑，简化污点传播分析复杂性；然后，提出了一个污点传播框架并证明了在该框架下污点传播分析的正确性和有效性；最后，采用消除，替换和移动等方法将冗余低效的污点传播代码转化为高效等价的污点传播代码.实验结果表明经过优化后单条热路径的污点传播代码节省了38%内存占用和指令执行时间，系统整体性能平均提升了6.8%.
|
| 关 键 词： | 安卓 隐私安全 动态污点跟踪 即时编译 污点传播优化 |
| 收稿时间： | 2015/5/7 0:00:00 |
| 修稿时间： | 2015/12/22 0:00:00 |
Dynamic Taint Tracking Optimization on Just-in-Time Compilation |
| |
| WU Ze-Zhi,CHEN Xing-Yuan,YANG Zhi and DU Xue-Hui.Dynamic Taint Tracking Optimization on Just-in-Time Compilation[J].Journal of Software,2017,28(8):2064-2079. |
| |
| Authors: | WU Ze-Zhi CHEN Xing-Yuan YANG Zhi and DU Xue-Hui |
| |
| Affiliation: | Third College, PLA Information Engineering University, Zhengzhou 450001, China;State Key Laboratory of Cryptology(Beijing Academy of Information Science and Technology), Beijing 100094, China,Third College, PLA Information Engineering University, Zhengzhou 450001, China;State Key Laboratory of Cryptology(Beijing Academy of Information Science and Technology), Beijing 100094, China,Third College, PLA Information Engineering University, Zhengzhou 450001, China;He''nan Province Key Laboratory of Information Security(PLA Information Engineering University), Zhengzhou 450001, China and Third College, PLA Information Engineering University, Zhengzhou 450001, China;He''nan Province Key Laboratory of Information Security(PLA Information Engineering University), Zhengzhou 450001, China |
| |
| Abstract: | Despite the demonstrated usefulness of dynamic taint tracking techniques in mobile privacy security, poor performance achieved by prototypes is a big problem. A novel optimization methodologies for dynamic taint tracking based on just-in-time compile is presented. First, the taint propagation logic is separate from the program logic precisely to simplifying the complexity of the taint propagation analysis; Then, a taint propagation framework is proposed and the taint propagation analysis is proved to be correct and validate; Finally, redundant and inefficient taint propagation codes is transferred to efficient and equivalent codes by adopting the methods of eliminating, replacing and moving. The experiment results show that 38% of memory usage and the time of execution of taint tracking instructions are saved for every single hot trace, and the performance of dynamic taint tracking system is averagely improved 6.8%. |
| |
| Keywords: | Android privacy security dynamic taint tracking just-in-time compiler taint propagation optimization |
|
| 点击此处可从《软件学报》浏览原始摘要信息 |
|
点击此处可从《软件学报》下载全文 |
|
