摘要
网络时延是重要的网络性能指标,是网络服务质量测量与分析领域的重要目标之一。在网络拓扑已知且稳定及链路性能时空独立性的假设前提下,给出了网络链路时延估计模型和端时延数据采集方法,提出了一种低复杂度的网络链路时延估计方法。首先应用伪似然估计,然后确定可解的探测单元,通过限制平均采样精度和探测单元链路数的增加来显著降低计算复杂度,解决了计算复杂度过高的链路时延估计求解问题。最后利用基于NS2的仿真研究验证了时延估计方法的有效性和准确性。
Network delay is one of the important network performance parameters,and it is one of the important objectives in the study field of network service quality measurement and analysis.Based on the assumptions that network topology structure was known and stable,and the link performance was temporally and spatially independent,the network link delay estimation model and end-to–end delay data acquisition were presented.An approach to network internal link delay estimation with low computation complexity was proposed.Firstly,Pseudo Likelihood Estimation（PLE） was adopted.Secondly,estimation units with definite solution were determined.By means of controlling the increase of average sampling precision and estimation unit links,the computation complexity of link delay estimation was significantly lowered.This approach can solve the problem of high computation complexity of network link delay estimation.Finally simulation study was performed on NS2 platform.The simulation results show that the approach is effective and accurate.
出处
《计算机仿真》
CSCD
北大核心
2011年第12期107-110,248,共5页
Computer Simulation
基金
国家自然科学基金项目(61172165)
广东省自然科学基金项目(S2011010000697)
广东省高等职业院校珠江学者岗位计划资助项目(2011)
深圳市基础研究计划-杰出青年基金项目(JC201005280613A)
关键词
网络链路时延
时延估计模型
伪似然估计
计算复杂度
Network link delay
Delay estimation model
Pseudo likelihood estimation（PLE）
Computation complexity
