Wireless Network Topology Control: Adjustable Resiliency and Network Traffic Delivery

Abstract

We consider the problem of wireless topology control with adjustable k-resilience and we evaluate related unicast and multicast network traffic delivery statistics across a series of initial simulation experiments. Our main goal is to examine design tradeoffs between resilient topology control, transmission power requirements, and resultant network throughput characteristics. We also present newly developed extensions to the network simulator, NS3, supporting multicast traffic experimentation. Our results for the random geometric networks studied demonstrate that moderate levels of $k$-resilience result in improvements in unicast traffic delivery while managing the growth of transmission power requirements across the network. These same trends were also observed for multicast experiments with connected dominating set (CDS) based forwarding, but traffic delivery did not improve for higher $k$ connectivities in the case of basic multicast flooding. Overall, we conclude that topology control edge connectivity resilience adds significant traffic delivery improvements in addition to providing other benefits, but the additional cost for establishing topological resiliency needs to be carefully considered against other defined system constraints. To address transmit power increases associated with resiliency and to improve traffic capacity in lower $k$ network topologies, further research is planned in applying network directivity components and multi-channel network architectures to the problem space.