摘要
传统的连续时延分布估计往往需要假设时延满足某种分布,估计精度受制于假设分布与实际时延分布的相关性。Gianni Antichi等提出了一种链路时延累积量估计的方法,无需假设时延满足某种分布,但需要内部节点的协作。针对上述问题提出一种完全依靠端到端测量的链路时延累积量估计方法,根据端到端的时延构建端到端时延累积量与链路时延累计量的方程,最终利用最优化方法计算出链路时延累积量的最优解。ns-2仿真结果验证了该方法的有效性。
The traditional methods for continuous delay inference usually under the assumption of the delay meet certain distribution,the estimation accuracy subjects to the dependence between assumptions and the actual delay distribution.Gianni Antichi,et al have proposed a method to estimate link delay cumulant,without assumption of the delay meet some kinds of distribution,but requires the collaboration between the internal nodes of network.Aiming at this issue,this paper proposes a method,which completely relies on end-to-end measurements,to estimate delay cumulant.The method firstly constructs equations about the end-to-end delay cumulant and the link delay cumulant by utilizing the end-to-end measured delay,and then uses optimization method to resolve the equation to get the optimal solution of the link delay cumulant.NS-2 simulation results validate the effectiveness of the method.
出处
《计算机工程与应用》
CSCD
2012年第3期91-94,208,共5页
Computer Engineering and Applications
基金
现代通信国家重点实验室基金项目(No.9140C1104040904)
教育部"新世纪优秀人才支持计划"(No.NCET-07-0148)
关键词
单播
端到端测量
时延累积量
背靠背包
层析成像
unicast
end-to-end measurement
delay cumulant
back to back packets
tomography