Maximum delay-constrained source rate over a wireless channel

Providing Quality of Service (QoS) guarantees in the presence of delay-sensitive data streams demands the understanding of the delay's behavior. Moreover, the support of QoS over a wireless channel comes up against to the time-varying nature of the channel. In this work, we accommodate both issues and evaluate the maximum source rate such that certain delay bound <i>D</i>^<i>t</i> can be supported with a violation probability ε. We call this maximum source rate <i>Capacity with Probabilistic Delay Constraint</i> <i>C</i>_<i>D</i>^<i>t</i>,ε. The effective bandwidth theory is a framework widely used to analyze wired networks and constitutes the basis of the work with the necessary adaption for its use in a wireless system. The time-correlated nature of the wireless channel has been modeled with a Finite State Markov Chain (FSMC). As expected, the maximum <i>C</i>_<i>D</i>^<i>t</i>,ε. increases for long allowed delays (<i>D</i>^<i>t</i> ← ∞) and diminishes when the delay constraint is more strict. The expected delay violation probability is compared to simulations in order to validate our results.