A Simulation and Prediction Model for Internet Traffic and QoS Based on 1-Step Markov-Chain

It is very difficult to predict long-term network traffic behavior in a large-scale network such as Internet. Due to the competition among ISPs and Internet exchanges (IXs), Internet users always experience different kinds of QoS, that effectively affect the network traffic. Accordingly, this paper designs a traffic model in the IX environment based on 1-step Markov Chain. The model simulates the IX-competition behavior. IX providers´ market share will be diluted through market competition. This paper considers the ISP business relationship with respect to the current Internet situation in most countries of the world. Routing designs and service binding strategies are also proposed in the study. The routing strategies vary with different competing environments. The model is thus meant to lead competitive capacity allocation strategies.