Generalized framework of modeling communication networks

Network modeling is the first step of performance analysis and simulation of communication networks. Good models should find a trade-off between precise description of real networks and simplification of analysis or simulation. In this article we suggest a generalized framework, that is to model communication networks from three different perspectives: topology, protocol and traffic. Hierarchical modeling is very effective in modeling the topology of communication networks, especially large-scale networks, since it can simplify the solution greatly and avoid being trapped in modeling complicated network details from the beginning of topology modeling. And as with protocol modeling, we can follow three steps: layering, division and formalization. Traffic models are at the hearts of performance evaluation of communications networks. Good traffic models should capture closely the statistical properties of the traffic in real networks. Actually, the three factors involved in communication networks are dependent on one another. Moreover, we should find a trade-off between the precision and efficiency when we try to model a real network according to what we care about mostly in modeling it.