Delay Analysis of Interacting Queues with an Approximate Model.

Abstract

An approximate model of coupled Markov chains is proposed and analyzed for a slotted ALOHA system with a finite number of buffered nodea. This model differs from earlier ones in that it attempts to capture the interdependence between the nodes. The analytical results lead to a set of equations that, when solved numerically, yield the average packet delay. Comparison between computational and simulation results for a small number of nodes show excellent agreement for most throughput values, except for values near saturation. Numerical comparisons for a two-node system show that a non-symmetric loading of the system provides better delay-throughput performance than a symmetric one.