Performance improvements for network-wide broadcast with instantaneous network information

Abstract

We study the performance of network-wide broadcasting as a function of the information implicitly available at nodes from neighbourhood transmissions. We term this set of instantaneous information as network information. Our discussion is focused on stateless broadcasting algorithms in which nodes decide on their forwarding behaviour based on the available network information. While stateless broadcasting schemes in the existing literature use various design guidelines that take advantage of specific aspects of the information, we develop a unified analytical model by characterizing the information available during different stages of broadcasting. Thus, our results are applicable to all stateless algorithms. We analyze broadcasting performance in terms of the transmission probability and redundancy of transmissions. Subsequently, we use our results to obtain insights on the feasibility conditions governing algorithm design depending on the network density and costs. While the first part of the work considers ideal channel conditions modeled as a unit disk graph (UDG), we subsequently enhance the model using a quasi-unit disk graph model (QUDG) to understand the effect of dynamic channel conditions.