Network Traffic Modeling And Prediction Based On FARIMA Model With Alpha Stable Distribution

With the rapid popularity of the Internet,the Internet has become an indispensable part of our lives.The exploration of the network has also become a topic of common concern to contemporary researchers,and the key to network exploration is to predict and study network traffic.Conventional network traffic prediction methods are based on the modeling of network traffic characteristics and analysis of the network traffic itself.And most of the existing network traffic model establishment and prediction methods do not take into account the heavy-tail characteristics and self-similar characteristics of network traffic.Therefore,based on the theory of Alpha stable distribution,this study conducts self-similarity research on network traffic,and uses the improved FARIMA model to model and predict network traffic data.Based on the introduction of the three basic theories of self-similarity,long correlation and heavy tail characteristics,this thesis analyzes the relationship between self-similarity and long correlation and heavy tail distribution.The reasons for the self-similarity of network traffic and the effect of self-similarity on network performance are expounded.At the same time,the Hurst parameter estimation algorithm is used to conduct long-term research on actual network traffic.Based on the Alpha stable distribution theory,the estimation algorithm of its characteristic index ? is used to verify the heavy-tail characteristics of the actual network traffic.The advantages and disadvantages of the existing self-similar network traffic models are compared,and the FARIMA model is selected to model and predict the network traffic data through comparative analysis.In order to improve the prediction accuracy of the model,the Alpha stable distribution is used to describe the actual network traffic,and the FARIMA model parameter estimation method is improved.The Alpha stable distribution combined with the FARIMA model parameter value method is applied to network traffic modeling and prediction with heavy-tail characteristics.Complete the modeling and simulation of FARIMA model by simulating network traffic data.Finally,the improved FARIMA model is used to model and predict the real network data of Bell Labs.The prediction results show that the FARIMA model based on Alpha stable distribution can effectively improve the prediction accuracy of network traffic data with heavy-tailed distribution characteristics.Network traffic analysis provides a reliable analysis tool and is of great significance to the development of network services.