A QoS model for random access systems with multiple request classes

In this paper, we present a simple and realistic two-class QoS model for random access systems. This model provides a general and effective approach to QoS classification by categorizing the available resources into distinct service classes. We present a simple Markov model of the subscriber stations to define the system transitions and state probabilities. Performance metrics based on random access success probabilities, average throughput and acceptance probabilities are investigated. This model provides a useful analytical tool for most contention systems that utilize random access as a mechanism for network entry, resource acquisition and contention-based or best effort data transmission. For example, the IEEE 802.16 bandwidth request and initial ranging mechanisms in the uplink contention channel. For simplicity, we consider only two classes, however the technique can easily be extended to any number of classes.