ColorCast: Deterministic broadcast in powerline networks with uncertainties

This paper initiates the study of broadcast in a powerline communication network, where nodes communicate by local broadcasts in the grid, and where the quality of the communication links is subject to uncertainty. We first show that state-of-the-art broadcast algorithms for known topologies fail to broadcast messages in such a challenging environment, even when the link quality uncertainty is small. We then present a deterministic algorithm COLORCAST that distributes a message to all nodes in the network in time Θ(n). The algorithm is based on graph coloring and avoids collisions while guaranteeing a high parallelism. In particular, COLORCAST strictly outperforms existing deterministic broadcast algorithms for unknown topologies in the sense that its time complexity is asymptotically lower than the best possible runtime for the unknown setting. Our formal analysis is complemented with a simulation study on real grid topologies, which confirms the benefits of COLORCAST compared to state-of-the-art protocols.