Multi-channel ALOHA and CSMA medium-access protocols: Markovian description and large deviations

Abstract

We consider a multi-channel communication system under ALOHA and CSMA protocols, respectively, in continuous time. We derive probabilistic formulas for the most important quantities: the numbers of sending attempts and the number of successfully delivered messages in a given time interval. We derive (1) explicit formulas for the large-time limiting throughput, (2) introduce an explicit and ergodic Markov chain for a deeper probabilistic analysis, and use this to (3) derive exponential asymptotics for rare events for these quantities in the limit of large time, via large-deviation principles.