【摘要】 网络流量建模是网络规划与性能评价的重要基础,传统的业务模型大多基于泊松模型和马尔可夫排队模型,只具有短程相关性,随着网络业务的不断研究发现,实际网络业务流在很长的时间范围内都具有长程相关性,即一种自相似性。本文采用RMD算法和Fourier变换法对网络流量的自相似模型-FBM模型进行了建模及仿真研究,生成了所需的自相似流量序列。然后分别采用R/S法和方差时间图法对其进行自相似参数检测。结果验证了仿真算法所产生的序列存在着自相似性,并同时对RMD算法和Fourier变换法的优缺点进行了分析。更多还原

【Abstract】 Network traffic models are important basis of network programming and performance evaluation.The conventional models are mostly based on Poisson model and Markovian traffic model,which is only Short-Range Dependence.With the continuous development of network services,studies found that the actual network traffic has a long-range dependence(LRD) now and in a very long time,which is a kind of self-similarity.In this paper,the RMD and Fourier algorithm were adopted to simulate and analyze FBM model,a self-similar model.They generated the necessary sequence of self-similar traffic.Then the article uses R/S method and variance-time method to verify Hurst value of the generated sequence of self-similar traffic in order to verify the self-similarity of the self-similar traffic sequence.The existence of self-similarity is verified by experiments,and the advantage and disadvantage of RMD and Fourier algorithm are analyzed.更多还原