Dynamic Traffic Competition in Atm Networks

In todays heterogenous multimedia ATM networks, there are multiple network managers, multiple traffic types and every user exploits the network resources according to availability, performance and cost. Thus, there are many decision-makers competing each other, and each one tries to achieve the best for his own good. In order a decision-maker to take the best decisions, he should have accurate information about the network state. However, in todays high speed multimedia ATM networks, the network state changes very fast and abruptly. Thus, a dynamic network model should be considered in the analysis and optimization of these high speed networks. In this paper, we first model the dynamics of ATM networks using both the virtual call and the cell processes in a two-level dynamic model. Then we provide a dynamic game framework for the resource sharing problems in ATM networks with competing traffic. Finally, we formulate the dynamic game problem both as a Nonlinear Complementarity Problem (NCP) and as a Variational Inequality Problem (VIP).