IP城域网流量建模研究

【作者】 闫含；

【导师】 高德远；

【作者基本信息】 西北工业大学 ， 计算机应用技术， 2005， 硕士

【摘要】 随着网络技术的迅速发展,互联网的规模日益庞大,互联网所提供的内容也日趋多样化、复杂化,网络流量特征研究的重要性日益突出。本文以西安多媒体城域网为研究对象,对其网络流量进行特征分析及建模研究。 主要的研究结果为以下三点: 1 对原始网络流量进行自相似特性分析,并通过自相关函数和自协方差来检验流量的自相似特性和长程依赖性,发现自协方差C^（m）（k）（其中m为聚集度）是以近似重尾曲线的方式逐渐衰减的。 2 发现IP城域骨干网流量的分布密度可以用Gamma分布进行比较准确的拟合,而其尾部则可以使用重尾分布曲线进行拟合。在本人涉猎的文献中尚未见此类研究报道。 3 在多重分形小波模型（MWM）基础上,采用上述研究结果分别对尺度系数和因子的分布做以修正,建立了网络流量的MWM-G模型,实验结果表明了该模型的有效性。 因此,本文的研究将对有类似网络拓扑结构和流量特点的城域网分析研究带来新的视角。更多还原

【Abstract】 With the rapid development of network technology, the Internet is getting increasingly bulky. The supplies of network and requirements of users are becoming more and more complex and diverse. As a result, it is increasingly prominent on characteristics analysis of network traffic. By targeting XI’AN Broadband Multimedia Network （XI’AN MAN）, we analyze the characteristics and set up a model to the network traffic from XI’AN MAN.Here are the main three results of our research:1 Firstly, we analyze the self-similarity of network traffic. Secondly, checking the self-similarity and long range dependence through the autocorrelation functionand autocovariance. Lastly, we have found that autocovariance C^m（k） （m:aggregation order）, according with the rule of Heavy-Tail function, is decaying gradually.2 Discovering that the distribution density of XI’AN MAN backbone traffic canbe fit accurately with Gamma distribution Ga（a,λ）, and it’s tail can be fit with theHeavy-Tail distribution. Up to now, there still have not been any reports related to the results above which among the documents I dabbled at.3 Based on MWM , we propose a new model（MWM-G） through modifying the distribution of scale coefficient a_j,k and factor A_j,k with the results mentionedabove. The results of the experiment show the effectiveness of this model.Therefore, the research results in this thesis bring us a new view on analysis of MAN with similar network topologic structure and network traffic character.更多还原