导 师: 唐屹
学科专业: G0104
授予学位: 硕士
作 者: ;
机构地区: 广州大学
摘 要: 基于fpga技术的网络入侵检测是未来的发展方向,而网络包头的分类是入侵检测系统的关键。 文章首先介绍了fpga技术的基本原理以及其在信息安全方面的应用,接着介绍入侵检测系统以及fpga技术在入侵检测系统中的应用。 分析了几种比较出名的网络包分类算法,包括软件分类方法、tcam分类算法、bv算法、tree bitmap算法以及端口范围分类算法。 在此基础上,文章设计了一个基于fpga技术的入侵检测系统包分类的基本框架图,实现框架图中的各个基本功能模块。在实现过程中,提出了一类结合三态内容可寻址内存(tcam)和普通存储器(ram)的网络包包头分类方案。我们将检测规则编号并位图化,使用ram存储与包头结构相关的规则位图,通过tcam上的数据匹配操作,快速关联待分析的网络数据包与入侵检测规则。文章还讨论了网包头分类方法的优化算法,将优化算法与未优化算法在速度和空间上进行比较。此外,还讨论了对snort的规则库进行整理和规则化的问题。 最后,对所设计的包头分类匹配模块在quartus ii进行仿真评估,将实验结果与已有的一些分类算法进行了比较。结果说明,本设计在匹配速度和更新速度上有优势,但消耗了较多的存储空间. The future direction for IDS /(Intrusion Detection System/) is FPGA-based. Network packet classification pays an important role in IDS. In this paper, we first introduce the FPGA technology and its application in information security, next we introduce IDS and FPGA technology used in IDS. We analyze several famous network packet classification algorithm, including software packet classification、BV algorithm、Tree Bitmap algorithm and port range algorithm. We lay a fame work of network packet classification based on FPGA technology and realize the frame work. During the implementation, we introduce an algorithm for packet classification combining TCAM and normal RAM. We use RAM to store the mapping of the rule set, through TCAM matching and rule set mapping, we can easily get the matching result. We also discuss how to optimize the rule set. In addition, we also discuss how to abstract rule from the Snort rule set and normalize them. Finally, we stimulate the packet classification components on Quartus II, and compare the experimental result with other famous classification algorithm. Result shows that our design is better than other algorithm on matching speed and update speed. but it also have disadvantage on space used and cost.
分 类 号: [TP393.08]
领 域: [自动化与计算机技术] [自动化与计算机技术]