基于属性数据的系统调用过滤方法  

Attribute-based methods for system call filtering

作　　者：郭航[1,2] 连一峰[1] 

机构地区：[1]中国科学院软件研究所,北京100190 [2]中国科学院大学,北京100049

出　　处：《计算机工程与设计》2014年第8期2621-2627,共7页Computer Engineering and Design

基　　金：国家科技支撑计划基金项目(2012BAK26B01);国家自然科学基金项目(61100226);北京市自然科学基金项目(4122085)

摘　　要：针对现有系统调用过滤方法的局限性,对如何有效准确地精简系统调用日志进行研究,分析系统调用日志中涉及网络攻击的重要系统调用信息,提出一种基于属性数据的系统调用过滤方法。通过追踪和分析系统调用的属性数据,引入系统调用依赖规则,在确保准确性的前提下,对系统调用日志进行合理有效地精简、过滤;在此基础上,实现一个名为"系统调用分离器"的过滤工具。通过实验验证了该方法及工具的有效性和适用性。Due to the limitations of the existing methods for system call filtering, an issue that how to filter system call logs more efficiently was studied, and the important information of system calls relating to network attacks was analyzed. Therefore an attribute-based method for system call filtering was proposed. Through tracing and analyzing all the needed attributes of system calls, system call dependency rules were used. The huge system call logs were filtered in a more efficient and effective way on the premise of ensuring the filtering accuracy. Based on this, a system call filtering tool called Separator was implemented and tested in experiments.

关 键 词：系统调用 系统对象 属性数据 系统调用依赖规则 网络攻击 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]