一种端到端可用带宽的测量方法  

End-to-end Available Bandwidth Measurement Method

作　　者：王芳[1] 钱德沛[1] 许大炜[1] 杜磊[1] 

机构地区：[1]北京航空航天大学计算机学院,北京100083

出　　处：《计算机工程》2006年第7期126-128,共3页Computer Engineering

基　　金：国家"973"计划基金资助项目(G1999032710);国家"863"计划基金重点资助项目(2001AA112120)

摘　　要：重点研究了网络端到端可用带宽的测量方法,分析了IGI和PTR算法的原理和局限性,将算法从单跳模式扩展到多跳网络,利用延时变更的概念,分析了探针包序列间隔变化与背景流量的关系,以此估计背景流量,并运用“相等区间”的方法确定最佳测量点,提高了可用带宽测量的准确性。The article focuses on Internet end-to-end available bandwidth measurement. This paper discusses the principle and limitations of the measuring algorithms IGI/PTR. Then the single-hop model of the algorithm IGI is expanded to multiple-hop model. The concept of ＂delay variation＂ is used to help understanding the relationship between the change of packet train gap and the cross traffic, therefore to estimate the cross traffic throughput. The method of ＂equality interval＂ is adopted to find optimal turning point. This method increases the accuracy of available bandwidth measurement.

关 键 词：可用带宽 带宽测量 主动测量 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]