Multicast Outage Probability and Transmission Capacity of Multihop Wireless Networks

TL;DR

This paper develops a tractable model for multicast wireless networks, analyzing multicast transmission capacity considering outage, retransmissions, and tessellation, providing insights into optimizing multicast performance in multihop networks.

Contribution

It introduces a new multicast model with a capacity metric that accounts for outage and retransmissions, and derives the capacity scaling laws under various network regimes.

Findings

Multicast capacity scales as Θ(ρ k^{x} log(k) v^{y}) under certain conditions.

Proper retransmission strategies significantly improve multicast transmission capacity.

Tessellation of multicast clusters reduces outage and enhances capacity.

Abstract

Multicast transmission, wherein the same packet must be delivered to multiple receivers, is an important aspect of sensor and tactical networks and has several distinctive traits as opposed to more commonly studied unicast networks. Specially, these include (i) identical packets must be delivered successfully to several nodes, (ii) outage at any receiver requires the packet to be retransmitted at least to that receiver, and (iii) the multicast rate is dominated by the receiver with the weakest link in order to minimize outage and retransmission. A first contribution of this paper is the development of a tractable multicast model and throughput metric that captures each of these key traits in a multicast wireless network. We utilize a Poisson cluster process (PCP) consisting of a distinct Poisson point process (PPP) for the transmitters and receivers, and then define the multicast transmission capacity (MTC) as the maximum achievable multicast rate per transmission attempt times the maximum intensity of multicast clusters under decoding delay and multicast outage constraints. A multicast cluster is a contiguous area over which a packet is multicasted, and to reduce outage it can be tessellated into $v$ smaller regions of multicast. The second contribution of the paper is the analysis of several key aspects of this model, for which we develop the following main result. Assuming $\tau/v$ transmission attempts are allowed for each tessellated region in a multicast cluster, we show that the MTC is $\Theta(\rho k^{x}\log(k)v^{y})$ where $\rho$, $x$ and $y$ are functions of $\tau$ and $v$ depending on the network size and intensity, and $k$ is the average number of the intended receivers in a cluster. We derive $\{\rho, x, y\}$ for a number of regimes of interest, and also show that an appropriate number of retransmissions can significantly enhance the MTC.