Solving bandwidth and priority domination problems of DQDB metropolitan area networks

The bandwidth domination problem (BDP) in the Distributed Queue Dual Bus (DQDB) Medium Access Control (MAC) protocol is the problem that the network bandwidth may be dominated by a single node when it is active early and other nodes having heavy-traffic demands are active later. The priority domination problem (PDP) is the problem that the network bandwidth may be dominated by an upstream node with lower priority when it is active early and other nodes with higher priorities are active later. To solve the BDP is to provide a mechanism such that the bandwidth can be shared (balanced) among all active nodes. To solve the PDP is to provide a mechanism such that nodes with higher priority can capture all the bandwidth. In this paper, we propose two adaptive algorithms to solve the BDP. The first one is based on the gradient approach and the second one is based on the proportional approach. We also propose a priority mechanism to solve the PDP. The proposed adaptive algorithms and priority mechanism are evaluated by simulation. From the simulation results, we found that both the proposed adaptive algorithms have a much faster speed to balance the bandwidth than that of standard bandwidth balancing mechanism. The proposed priority mechanism solves the PDP effectively. We also show that if the proposed adaptive algorithms and priority mechanism are combined together, then the nodes with highest priority can use all the bandwidth and the bandwidth can be balanced among these nodes fastly.