摘要
在深入分析基于树结构和计算几何学点定位算法而衍生的主要分类算法原理的基础上 ,建立了算法性能评估仿真测试平台 .平台包括网络流的软件仿真、算法的软件实现和算法性能参数测试统计三个主要部分 .通过对各算法的时间性能和空间性能及其随分类规则数和分类域数的变化趋势的测量 ,从数量上把握各算法的主要特点及其适用性 ,从而可以将各种算法有机地结合在一起胜任速度越来越高的网络应用需求 ,同时也是进一步研究其它算法的基础 .
After investigating the variety of algorithms based on tries structure and computational geometry, the software testing platform for their performance evaluation was implemented. The platform consists of network traffic simulation, algorithms and output parameters statistics. With the main performance curves, including time complexity curves and space complexity curves presented in the paper, the behaviors of algorithms can be recognized deeply. This is necessary as a good starting point to do further research on the new algorithms for packet classification and it is also possible to satisfy the most current application requirements by the combination of the present algorithms.
基金
中国科学院回国择优基金资助项目
关键词
包分类算法
仿真测试
时间性能曲线
空间性能曲线
packet classification
simulated testing
time complexity curve
space complexity curve
