大规模网络中IP流流速分析

Analysis of IP Flows Rate in Large-scale Networks

作为Internet流量模型研究的重要组成部分,IP流流速反映了各种不同应用类型的流量在网络中时实际负载的贡献情况。通过分析和寻找对IP流平均流速产生主要影响的若干关键因子,可以为基于IP流的流量模型提供必要的条件。本文首先分别对IP流的三种主要构成部分:TCP流、UDP流和ICMP流的平均流速从协议分析的角度进行建模,从平均流速模型参数分析中,获得在不同阶段对决定IP流平均流速的若干主要影响因子;然后,使用采集自各种不同时间、不同应用背景和不同负载的大规模高速网络的TRACE作为研究对象,通过试验分析的方法,对所选取的TRACE中不同协议类型的IP流平均流速进行统计分析,检验这些因素在实际网络中在不同流长的情况下对IP流平均流速的贡献,从而验证了本文所提出的IP流平均流速影响因子的可靠性。

As one of most important components of Internet traffic modeling, the IP flow rate can be used to describe the load distribution of different applications in the network And it is essential condition for IP flow model that some key factors could be found which influenced the flow rate heavily. Firstly, this paper modeling the IP flow rate based on protocol analysis for the three main components： TCP flows, UDP flows and ICMP flows. From the analysis of parameters of those models, the flows rate are found out to be determined by some influence factors. And then, those TRACEs which come from the networks with different time, different areas and different payloads are chosen as study objects. IP flow rate of the different protocol IP flows in these traces are analyzed using statistical method. The influence factors are verified their efficiency for the IP flows rate with different length in the actual networks, and the reliability of the IP flows rate influence factors which are introduced by this paper is also verified.