【作者】 卢贺；

【导师】 勾学荣；

【作者基本信息】 北京邮电大学 ， 通信与信息系统， 2012， 硕士

【摘要】 随着移动通信技术的发展,3G数据业务逐渐成为移动通信系统中的重要组成部分。由于智能终端的普及以及用户使用行为的改变导致3G数据业务的流量模型与传统的语音业务相比存在显著的不同,而且数据业务模型的研究在3G移动网络的建设和运营中如：话务量预测、资源配置、网络优化等显现出越来越重要的价值。所以,基于不同数据业务特征的业务建模方法以及3G系统中业务流量自相似分析的研究,已经成为一个很有价值的研究领域。本研究课题来自于北邮网络优化研究室与华为技术有限公司合作的“3G移动数据业务建模与话务预测”的项目。本论文主要针对3G网络中分组域数据业务模型以及自相似性流量分析进行研究。论文首先介绍了在现有网络中典型数据业务的业务模型研究现状,总结出一种广义三层模型作为对3G数据业务分析建模的基本方法。然后通过分析QQ的通信机制以及QQ的数据包结构创造性的提出了一种QQ业务模型,并且根据UMTS网络中获取的真实网络数据,对QQ业务模型进行验证,计算并拟合出了相关模型指标分布及分布参数。通过软件需求分析、设计和编码,实现了移动数据业务建模系统。经过大量研究发现分组数据流量存在自相似性和长相关性,本文最后对QQ流量的自相似性进行了分析和验证,并用不同的估计方法给出了QQ流量和QQ包到达率时间序列的Hurst参数值。更多还原

【Abstract】 With the development of mobile communication technology,3G (Third Generation) data services become increasingly important component in mobile communication system. The intelligence of mobile terminal and user behavior of various service make the model of 3G data service significant different from the conventional voice service. The research of data service model is playing more and more important role in the construction and operation of 3G mobile networks such as traffic forecasting, resource allocation, network optimization and so on. The modeling method according to the characteristics of different data services, as well as the self-similarity in traffic analysis in the 3G system, is a valuable research issue.The Research is supported by Beijing University of Posts and Telecommunications Network Optimization Research Center cooperation projects with Huawei Technologies Co., Ltd. "3G Mobile Data Service Modeling and Traffic Forecasting".This paper focused on designing data service model and self-similarity traffic analysis in PS domain of 3G networks. Firstly, this paper introduced the research status and schemes of typical service modeling in the existing networks, and summarized the three-layer model as a basic method of analysis and modeling on the 3G data service. By analyzing the communication mechanism and format of QQ packet, a novel QQ service model is put forward and the characteristics of the parameters and distributions are given based on the real QQ data collected on the UMTS networks. Through software requirement specification, designing and coding, the Mobile Data Service Modeling System is completely implemented. Extensive researches of the network traffic show the packet data traffic has the characteristics of self-similarity and long range dependence. The self-similarity of QQ traffic is investigated and the Hurst parameters are estimated with different methods.更多还原