An approach to modelling and simulation of complex slotted systems with applications to metropolitan area networks

Presents an approach for the performance analysis of high-speed slotted systems. Our approach is based on a combination of analysis and simulation where a Markovian state-space representation of the system is constructed, which faithfully takes into account the media access control (MAC), the traffic to each node, the position of each node in the network, etc. In large networks, the dimensionality of such a representation makes the exact numerical analysis rather impossible. However, such a representation can be readily used in simulating the complex system very easily. As compared to “blind” simulation of subsystems, our approach is found to be considerably more efficient in both the construction of the simulation program and the run time. We have demonstrated our approach in modelling and performance analysis of a few high-speed slotted systems with very bursty cell arrival processes where the impact of the burst size variability on the performance of each node is carefully examined. As an example, a DQDB network with and without spatial reuse and a slotted buffer insertion bus network are considered