Congestion modeling and synthesis for multiple bottlenecked nodes in packet-switched networks

Traffic management and adjustment is obviously of great importance in any communication network. Without traffic supervision and control, an increase in demand for network resources can lead to a significant decrease in network service quality and efficiency. Since the most important issue in the field of control theory is stability of dynamic systems, control science can be appropriately applied to solve communication network traffic control problems. The initial step is to define a dynamic model for the time-related changes of variables, especially the variables to be controlled such as the queue length and the permitted sending rate. In this paper, multiple bottleneck networks are investigated with some simplifications. Due to the favored analytical model of ATM networks, it has been chosen for applying the proposed method. The impact of changes in parameters and variables in different types of networks are illustrated via simulation so as to identify the requirements and propose appropriate control methods as future works.