Research And Implementation Of Distributed Network Traffic Data Analysis And Management

With the rapid development of Internet, network management has become increasingly important. Traffic monitoring and statistical analysis are the bases of the entire network management. In order to achieve integrated management of the network, it is necessary to collect and manage the network traffic information, obtain accurate traffic data. It is required to analyze network traffic information and accurately judge the health status and future development trend of the network.Most of traffic analysis methods are based on single or local links. However, nowadays network researchers facing more important question is to model and analyze all the links, such as estimation of traffic matrices. Most of the previous approaches infer traffic matrix from either SNMP link loads or sampled NetFlow records. In this work, we design novel inference techniques that, by statistically correlating SNMP link loads and sampled NetFlow records, allow for much more accurate estimation of traffic matrices than obtainable from either information source alone. And we design techniques that, by comparing notes between the above two information sources, identify and remove dirty data, and therefore allow for accurate estimation with the cleaned data. We also present an approach to traffic matrix estimation problem with the single data source. Given the network topology and the link traffic, it utilizes the expectation maximization algorithm to compute the traffic demand of the pair of origin and destination. We evaluated the presented several algorithms using real experimental data. Experiment results show that the proposed method is more accurate than tomogravity algorithm and generalized gravity model method.According to the requirements of the integrated network management technique, we design the architecture of the network traffic situation management system. Furthermore, we designed and realized the prototype. There are five function modules of the system being designed in detail, including traffic collection, data reception, data transmission, traffic analysis, data storage and management. Besides, traffic analysis structure based on OD flow is brought forward in this thesis. And it has completed traffic data analysis with the corresponding algorithms, which involves determining key links and bottleneck nodes in the network, detecting elephant flows and abnormal flows. Traffic database is designed to manage large-scale raw traffic data and the analysis data. Finally, the function and efficiency of the prototype are tested using real-sampled network traffic data. The running results and the performance analysis are presented.In summary, this work deals with distributed network traffic data analysis and management, which provides a basis for future research and development.