Architecture and Abstractions for Environment and Traffic Aware System-Level Coordination of Wireless Networks: The Downlink Case

Two ways to substantially enhance wireless broadband capacity are full frequency reuse and smaller cells, both of which result in operational regimes that are highly dynamic and interference limited. This paper presents a system-level approach to interference management, that has reasonable backhaul communication and computation requirements. The basis for the approach is clustering and aggregation of measurements of the spatial diversity in sensitivity to interference associated with average user populations. This enables the system to exchange information and optimize coordinated transmission schedules using only coarse grained data. The paper explores various ways of optimizing such schedules: from a static, decoupled version to a dynamic version capturing user-level scheduling, fluctuating loads and inter-cell interference that couples base stations' performance. Based on extensive system-level simulations, we demonstrate reductions in file transfer delay ranging from 20-80%, from light to heavy loads, as compared to a simple baseline not unlike those in the field today. This improvement is achieved while providing more uniform coverage, and reducing base station power consumption by up to 45%.