电子科技university本科毕业论文模版.docxVIP

摘要 PAGE \* MERGEFORMATI 摘要 Snort是一个开源的网络入侵防御和检测系统（NIDS），其利用模式匹配算法以及用户自定义的规则集来检测潜在的非法入侵。在近几年网络流量激增与规则集不断增大的背景下，Snort这个基于CPU的单线程程序的处理能力是不够的。本文介绍了一个优化版本的Snort的设计与实现，它利用OpenCL将Snort中的模式匹配算法改进到GPU上运行。本项目的测试部分在Freescale的iMX6Q-SabreLite ARM开发板上进行，并且由于OpenCL良好的可移植性，程序也可以在x86的平台上进行运行测试。测试结果显示，改进版的Snort相对原版的Snort，将可应对的最大网络流量提高了两倍。 关键字：Snort，网络入侵检测系统，模式匹配，OpenCL，GPU   Abstract II Abstract Snort is an open source network intrusion prevention and detecting system (NIDS) using pattern-matching algorithms and user-defined rule set to detect potential illegal break-ins. As Snort is a single-thread CPU-oriented program, it is not enough when the network stream throughput increases rapidly and the rule set gets quite larger in recent years. This paper introduces a design and implementation of an optimized version of Snort based on GPU-oriented pattern-matching algorithms using OpenCL. The project is tested on a Freescale iMX6Q-SabreLite ARM board, and due to the portability of OpenCL, it also works on x86 platforms. The test results show the performance of the optimized Snort approximately improves the maximum network stream throughput that Snort could cope with by a factor of two, compared with the original version. Key Words: Snort，NIDS，patterns matching，OpenCL，GPU   PAGE \* MERGEFORMAT10 目 录  TOC \o 1-3 \h \z \u  HYPERLINK \l _Toc356767931 第1章 引言  PAGEREF _Toc356767931 \h 2  HYPERLINK \l _Toc356767932 1.1 模版使用说明  PAGEREF _Toc356767932 \h 2  HYPERLINK \l _Toc356767933 1.2 课题背景  PAGEREF _Toc356767933 \h 2  HYPERLINK \l _Toc356767934 1.3 课题的价值及意义  PAGEREF _Toc356767934 \h 2  HYPERLINK \l _Toc356767935 1.4 课题的国内外研究现状  PAGEREF _Toc356767935 \h 2  HYPERLINK \l _Toc356767936 1.5 课题难点、重点、核心问题及方向  PAGEREF _Toc356767936 \h 2  HYPERLINK \l _Toc356767937 第2章 相关概念与技术  PAGEREF _Toc356767937 \h 2  HYPERLINK \l _Toc356767938 2.1 概述  PAGEREF _Toc356767938 \h 2  HYPERLINK \l _Toc356767939 第3章 设计与实现  PAGEREF _Toc356767939 \h 2