基于统计信号处理的网络链路容量和可用带宽联合估计  

Combined estimation of network link capacity and available bandwidth based on statistical signal processing

作　　者：丛林[1] 吕国晗[2] 邓北星[1] 李星[1] 

机构地区：[1]清华大学电子工程系,北京100084 [2]微软亚洲研究院,北京100190

出　　处：《清华大学学报（自然科学版）》2010年第10期1669-1673,共5页Journal of Tsinghua University(Science and Technology)

基　　金：国家"九七三"重点基础研究项目(2007CB310806);国家自然科学基金资助项目(60850003)

摘　　要：网络链路容量和可用带宽参数估计是互联网领域的重要课题。该文将统计信号检测及估计方法应用于此估计问题,提出了扩展自诱导拥塞原理,实现了网络链路容量/可用带宽的联合估计算法,开发了实际测量工具pathWave。NS-2仿真平台下的测量结果验证了扩展自诱导拥塞原理的有效性。Schooner实验床环境下的实验仿真研究表明,与其他传统估计方法以及测量工具相比,pathWave以较低的测量负载和较短的测量时间,有效提高网络链路容量/可用带宽的联合估计精度。Network link capacity（lc） and available bandwidth（av-bw） are of great significance for various network domains.In this paper,the estimation of network lc/av-bw was modeled as the problem of stochastic signal state detection and signal parameter estimation.An approach to measuring network lc/av-bw was developed based on the statistical signal processing theory using the extended self-induced congestion principle.The affectivity of lc/av-bw measurement tool,pathWave,was validated in the NS-2 simulation environment.The pathWave performance was evaluated in a Schooner test-bed by comparing with other representative tools.Analytical,simulative,and experimental results show that the method based on statistical signal processing is more efficient for the combined estimation of lc/av-bw.

关 键 词：网络链路容量 网络可用带宽 统计信号处理 扩展自诱导拥塞原理 

分 类 号：TP393.06[自动化与计算机技术—计算机应用技术]