IP网络流量分析预测工具的设计与实现

发布时间：2018-03-03 05:03

本文选题：互联网流量　切入点：流量矩阵　出处：《华中科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着IP网络的飞速发展和网络应用的日益丰富，运营商对于网络流量的精细化管理的需求愈加迫切。一方面，需要处理大量的采自于现实网络的流量数据，以获知网络链路的利用率等状态信息；另一方面，也需要对网络流量的增长趋势进行预测，以便提前展开网络规划等工作。为了应对这些需求，本文设计了一套IP骨干网络流量的分析预测工具。该网络流量分析预测工具由数据预处理、节点流量分析、流量矩阵分析、节点流量预测以及流量矩阵预测等五部分构成。该工具以现网实际采集的流量矩阵数据为主要分析数据，SNMP测量节点流量为辅助分析数据。在节点流量分析部分，提供了流量分解、流量建模、流量对比以及网络链路负载计算等功能；在节点流量预测部分，，提供了ARIMA预测、多元回归预测等功能；在矩阵流量分析部分，提供了对流量矩阵关键元素、扇出项等进行观察和分析的功能；在矩阵流量预测部分，可以把节点流量的预测结果分摊到各流向上，对节点出入流量、链路负载利用率以及流量流向分配比例等进行可视化展示。本文实现的流量分析预测工具，已通过某运营商提供的现网实际采集流量数据的测试。可支持实际流量数据的导入，可对网络局部或者整体进行分析与预测，并输出图形化报表。该工具能对网络维护、网络规划提供科学的参考，起到积极有益的作用。
[Abstract]:With the rapid development of IP network and the increasingly rich network applications, the operators need to manage network traffic more and more urgently. On the one hand, a large number of traffic data collected from real networks need to be processed. In order to obtain status information such as the utilization rate of network links; on the other hand, it is also necessary to predict the growth trend of network traffic in order to carry out work such as network planning ahead of time. This paper designs a set of IP backbone network traffic analysis and prediction tools. The network traffic analysis and prediction tool consists of data preprocessing, node traffic analysis, traffic matrix analysis, The tool is composed of five parts: node flow prediction and traffic matrix prediction. This tool mainly analyzes the traffic matrix data collected by the network, and the SNMP traffic measurement node traffic is used as the auxiliary analysis data. In the node flow analysis part, The functions of traffic decomposition, traffic modeling, traffic comparison and network link load calculation are provided. In the node traffic prediction section, the functions of ARIMA prediction, multiple regression prediction, and matrix traffic analysis are provided. It provides the function of observing and analyzing the key elements of the flow matrix, fan out items, etc. In the forecasting part of the flow of the matrix, the forecast results of node flow can be apportioned to each flow direction, and the node flow in and out can be divided into and out. Link load utilization and flow flow distribution ratio are visualized. The traffic analysis and prediction tool realized in this paper has been tested by the actual traffic data collected by a network provided by a certain operator. It can support the introduction of actual traffic data and can be used to analyze and forecast the local or whole network. The tool can provide a scientific reference for network maintenance and network planning and play a positive and beneficial role.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【相似文献】