A Novel Traffic Model for the Analysis of Network Traffic Characteristic
Lingbo Pei, Ming Chen · 2007
According to some discoveries and analysis results of network measurement for network traffic in recent years, much literature has proven that traffic present the nonstationary Poisson characteristic at sub-second time scales in IP backbone, which not accord with self- similarity characteristic for last decade. By analyzing Internet structure, a novel weighted mapping model is put forward, which gives a reasonable explanation for the new traffic characteristic. The model relates chaos impact factors with traffic characteristic, and transfers the traffic with self-similarity in access network to that with Poisson characteristic. Finally, the validity of the model is verified by a lot of simulation experiments on NS2.