Capacity of Wireless Networks with Heterogeneous Traffic

We study the scaling laws for wireless ad hoc network in which the distribution of nodes in the network is homogeneous but the traffic is heterogeneous. More specifically, we consider the case in which a node is the sink to k sources sending different information, while the rest of the nodes are part of unicast communications with a uniform assignment of source-destination pairs. We prove that the capacity of these heterogeneous networks is ¿(n/T_max), where T_max and n denote the maximum traffic for a cell and the number of nodes in the network, respectively. Equivalently, our derivations reveal that, when n - k ¿ constant, the network capacity is equal to ¿(¿(n/(log n))) for k = O(¿(n log n)) and equal to ¿ (n/k) for k = ¿(¿(n log n)). Furthermore, the network capacity is ¿(1) when n - k = constant. These results demonstrate that the capacity of a heterogeneous network is dominated by the maximum congestion in any area of the network.