Research On VoIP Traffics Modeling And Prediction At Application Layer In Metro Area Network

Complexity and diversity of Internet traffic are constantly growing. Networking researchers become aware of the need to constantly monitor and reevaluate their assumptions in order to ensure that the conceptual models correctly represent reality.Internet traffic today is a complex nonlinear combination of the seasonal time series.Metro area network plays a key role in the Internet, IP-based voice packet transmission (VoIP) telephone services are currently being deployed nationwide in MANs. traffic flows characteristics of MAN are very important for traffic engineering to improve network performance. The design of appropriate and effective traffic flows models is a desirable task for enterprise, university and ISP networkers. The current network traffic measurement research is mainly concentrated on the flow forecasts and analysis based on network layer or transport layer, and a traditional autoregressive integrated moving average (ARIMA) model which can only describe the overall network traffic trends is used, however, VoIP traffic based on application layer aren't always consistent with ARIMA model, because the model has not accurately described the complicated structure of today's Internet-unexpected continuity, anomality and the self-similar characteristics. The main contributions of this paper are as follows:(1) Analyzing the prediction accuracy of the traditional ARIMA model when the network traffics volatile heavily, we find the modeling methodology can be improved efficiency. An improved seasonal ARIMA model is proposed. The volatile points are extracted and rebuilt for prediction so as to improve the prediction accuracy.(2) By using VoIP comprehensive statistical data collected with NetTurbo on an ISP WAN link, we establish a model of improved multiple seasonal ARIMA and predict the VoIP traffics on output link of this MAN. Experimentation shows that based on an improved seasonal ARIMA model, the traffic can be described more accuretely at the valatile circumstances and the accuracy is almost improved by 5%.(3) Putting the improved seasonal ARIMA model into management of VoIP network traffics, we can detect the possible time when the VoIP traffics are beyond the threshold, and pre-take measures to ensure the QoS of VoIP.