Performance analysis of joint voice-data PRMA over random packet error channels

We consider a joint voice-data packet reservation multiple-access (PRMA) system with transmission errors. Two analysis methods are presented. The first is a combined Markov and equilibrium point analysis. In this method, equilibrium point analysis is used to obtain the equilibrium number of backlogged data terminals, while Markov analysis is used to compute the stationary state probability distribution, assuming the number of backlogged data terminals is equal to the equilibrium value. The second is a Markov analysis using an approximate marginal distribution of backlogged data terminals. System performance measures, namely, voice packet dropping probability, data packet delay, and throughput, are derived in the presence of transmission errors. Computation complexity and accuracy are compared for the two analysis methods. Simulation results are also presented for comparison with analysis results, and good agreement is observed, especially when the packet header error rate is small.