高性能网包分类理论与算法综述

发布时间：2018-03-22 07:19

本文选题：网包分类　切入点：计算几何　出处：《计算机学报》2013年02期 　论文类型：期刊论文

【摘要】：随着IP网络架构的不断演进以及网络业务和安全需求的不断增长,高性能网包分类在下一代交换机、路由器、防火墙等网络基础设备中有着越来越广的应用.网包分类算法作为高性能网包分类的核心技术,具有重要的研究价值和实践意义.文中从理论分析和算法设计两方面介绍了高性能网包分类的最新研究成果.在理论分析层面,依据计算几何理论对网包分类问题的数学解法及复杂度进行了归纳,总结了网包分类算法的理论依据及性能评价方法.在算法设计层面,对具有影响力的网包分类算法按照不同的研究方向进行了归类和介绍,并结合自身研究成果对不同类别的算法设计思路行了深入分析.作者在多核网络处理器平台以及FPGA平台上实现了几类具有代表性的网包分类算法,并通过真实的网络流量测试比较了不同类型算法在不同系统平台上的实际性能.最后,作者总结并展望了高性能网包分类的下一步发展方向.
[Abstract]:With the increasing of the evolution of IP network architecture, network services and security requirements, high performance Packet Classification in the next generation of switches, routers, there are more and more widely used firewall network infrastructure. The core technology of network packet classification algorithm for high performance packet classification, has the important research value and the practical significance. This paper from two aspects of theory analysis and algorithm design, this paper introduces the latest research results of high performance packet classification. In the theoretical analysis, based on the mathematical method for calculating geometric theory of packet classification problem and complexity are summarized, summarizes the theory and method of performance evaluation of packet classification algorithm in the algorithm design level, the influential network packet classification algorithm according to the different research directions were classified and introduced, and combined with the algorithm design of different categories of their own research ideas For an in-depth analysis. The author on multi-core network processors and FPGA platform to implement network packet classification algorithm for several kinds of representative, and compare the actual performance of different types of algorithms in different system on the platform of the network traffic by real test. Finally, the author summarizes and prospects the next development direction of high performance packet classification.

【作者单位】：清华大学自动化系;清华大学信息技术研究院;清华信息科学与技术国家实验室(筹);
【基金】：清华信息科学与技术国家实验室(筹)学科交叉基金北京市共建项目专项资助
【分类号】：TP393.08

【相似文献】