报文分类算法可扩展性标准评测系统  

Standard scalability evaluation system of packet classification algorithms

作　　者：周粳迪[1] 程东年[1] 刘勤让[1] 

机构地区：[1]国家数字交换系统工程技术研究中心,河南郑州450002

出　　处：《计算机工程与设计》2009年第18期4141-4145,4150,共6页Computer Engineering and Design

基　　金：国家973重点基础研究发展计划基金项目(2007CB307102)

摘　　要：针对报文分类算法没有标准可扩展性评测工具的问题,基于计算机体系结构测试领域的benchmark方法开发出基于benchmark的报文分类算法可扩展性评测系统。该系统利用benchmark参数文件引导生成规则库,允许用户利用高层输入参数控制规则库和Trace的产生,并能够实现对被测算法可扩展性能指标的实时监测。最后利用该评测系统对典型报文分类算法在不同模拟场景下进行仿真评测。仿真结果表明,该系统能够准确评测算法的可扩展性,为研究报文分类算法的可扩展性提供了标准评测工具。According to a problem that there hadn＇ tbeen a standard scalability evaluation tools for packet classification algorithms, a scalability evaluation system is developed for packet classification based on the benchmark of computer architecture evaluation field, the system utilized benchmark parameter files to guide the generation of rule set, permitted users using high-level input parameters to control the generation of rule set and trace, and achieved real-time monitoring the scalable performance of evaluated algorithms. Finally the system is utilized to evaluate the classic packet classification algorithms in different simulated scenes, the result indicated that the system could evaluate the scalability of all algorithms exactly, a standard evaluation tool is provided for researching the scalability ofpacket classification algorithms.

关 键 词：报文分类 可扩展性 复杂度 评测系统 BENCHMARK 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]