Research On Characteristic Analysis And Forecast Of Network Traffic

With the rapid development of network communication technology, the internet begins to carry more and more application services, it brings forward very high demands for network quality of service, traffic control and network management. Analysis and forecast of flow are the foundation of network management and performance analysis: the operating rules and characterizations of the network are obtained through analyzing flow, modeling based on the flow characteristics, can not only forecast flow patterns of behavior, but also be applied to other fields such as congestion control and quality of services. Therefore, starting with the analysis of traffic, the characteristic models and forecast methods of network traffic are studied in this paper.Firstly, the paper summarizes the various approaches to capture true traffic, and comprehensively narrates the current situation about studying flow characteristics at home and abroad; The self-similar characteristics of operational traffic is analyzed in detail: including its mathematical description, the estimative means of Hurst parameter, the cause of self-similar characteristic, the impact on network performance, and the self-similar characteristic of the true network data is validated; At the same time, the periodicity and chaos nature of network traffic are expatiated and the relationship between periodicity and self-similarity is also discussed and explored, it offers the foundation for the following researches. Secondly, the typical forecasting methods and models in various period for the business flow are summarized, the limitation of the traditional business models based on poisson distribution is deeply analyzed; The advantages and disadvantages of primary self-similar models are studied and the latest techniques and theories in traffic forecast domain are deeply discussed; A conclusion that flow forecasting has to adopt the hybrid models to forecast is summed up by means of comparing and analyzing the models; Thirdly, in order to establish the model accord with actual traffic characteristic, including the nature of self-similar, periodicity and chaos, a hybrid network flow forecast model that combined with wavelet technology and time series analysis is proposed and the future behavior trend of true traffic data by using the forecast algorithm of this model is forecasted and analyzed. The simulation results shows that the hybrid model can not only fully describe and characterize the flow properties, but also forecast network traffic behavior accurately and effectively.Finally, the potential applications of forecast algorithm in network management are discussed and analyzed; Shortages of this forecast algorithm are briefly analyzed and the next research target in our future work is pointed out.