MAC Layer Wireless Multicast: Theory and Approaches

We develop the theory behind wireless MAC layer multicast and obtain throughput optimal transmission policies that provide desired loss and power characteristics. Key idea is to exploit the broadcast nature of wireless channel, i.e, a sender can reach all the receivers in a MAC layer multicast group with a single transmission. Through the broadcast nature of wireless transmissions provides a possible approach to improve the efficiency of the multicast communication, it also imposes critical challenges. A multicast specific challenge is that some but not all the receivers may be ready to receive on account of the interference caused by the transmissions in their neighborhood. The readiness state of a receiver depends on the network load. A transmission policy, which does not transmit until all the receivers are ready may have to wait long and hence it may provide low system throughput. On the other hand, if the sender transmits when only a few receivers are ready, then the transmitted packet will be lost at the receivers that were not ready. This packet loss may not be acceptable. Hence, the policy decision in this case is when should a sender transmit. Further, a sender may achieve required loss characteristics by transmitting a packet several times till sufficient number of receivers receive the packet. But additional power consumption in this case at the sender imposes a limit on the number of such transmissions. We consider a single multicast session in isolation. The impact of the network on the multicast session is modeled by considering random receiver readiness states. In each slot, we assume that a receiver is ready with probability (w.p.) p. Further, the number of arrivals in each slot is assumed to be iid with expectation denoted by λ