A Discrete-Time Queuing Analysis of the Wireless ATM Multiplexing System

In this paper we analyze a wireless channel model which is subject to service interruption either because the channel is not available or the server is serving to the other users. The wireless channel is modeled by a Markov chain with two states corresponding to high and low error states, respectively. It is assumed that the channel is slotted in time and no transmission is possible during the high-error state. The traffic generated by a mobile user is modeled as the superposition of independent two states On-Off Markov sources, which are statistically independent of the server. A source generates packets only during an On state each of which fits to a channel slot. A discrete-time queuing analysis derives the probability generating function (PGF) of the queue length under the assumption of an infinite buffer. From the PGF, mean queue length, mean delay and approximate finite buffer overflow probabilities are calculated.