摘要
现有网络层析成像的研究大多假设链路状态在测量周期内保持不变,因此难以捕获网络链路状态参数的时变特征。打破传统链路丢包率估计方法对链路状态平稳的假设,提出一种基于时空相关性的网络链路时变丢包率估计方法。该方法使用状态转移矩阵描述链路丢包率的时空相关性并进行估计,然后利用最小二乘法修正先验估计结果,以获得链路时变丢包率估计结果。NS-2仿真结果验证了提出的方法能有效追踪链路丢包率的变化,且优于平稳链路丢包率估计方法。
Most existing works of network tomography assumed that link states remained constant during measurement period, with the result that the time-varying characteristics of link state parameters could not be captured. This paper presented a tem- poral and spatial correlation based time-varying network link loss rate estimation method by releasing the stationary link state assumption. It used the state transition matrix to describe the temporal and spatial correlation of link loss rates, and estimated the link loss rates. It applied the least square algorithm to revise the prior estimates in order to obtain the time-varying link loss rates. NS-2 simulation results show that the proposed method is capable of tracking the variation of link loss rates effectively, and is superior stationary link loss rate estimation method.
出处
《计算机应用研究》
CSCD
北大核心
2013年第2期557-559,581,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(60872033)
关键词
网络层析成像
非平稳
时空相关性
丢包率
network tomography
non-stationary
temporal and spatial correlation
loss rates
