We explore how next-generation wireless systems can exploit sectorization—the ability of each infrastructure node to activate multiple directional antennas—to reshape the network and deliver fiber-like speeds. Our work begins by defining a comprehensive sectorized network model that captures how each node’s sectors interact to form multi-hop paths, and how interference is influenced by different sector configurations.
Within this model, we compute the capacity region, which represents all the traffic demands (or flows) the network can support without overloading. By carefully tuning each node’s sectorization, we can increase total throughput, reduce congestion, and route traffic more efficiently. We formalize this tuning as an optimization problem, where given a target network flow, we find the best way to define the sectors. We then delve deeper into how dynamically reconfiguring these sectors can change the effective network topology—unlocking even more capacity when traffic patterns vary over time.
A central contribution is our Even Homogeneous Sectorization approach, which constructs a bipartite structure in the effective network graph. This structure significantly simplifies capacity calculations, speeds up routing decisions, and makes distributed algorithms easier to implement. By endowing the network graph with bipartite properties, we show how to quickly determine stable flows, use backpressure-based routing to handle fluctuating traffic, and achieve strong performance in simulations.
Promponas, P., Chen, T., & Tassiulas, L. Optimizing Sectorized Wireless Networks: Model, Analysis, and Algorithm. In 24th International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing (MobiHoc '23).