Hekaton: Efficient and Practical Large-Scale MIMO

Abstract

Large-scale multiuser MIMO (MU-MIMO) systems have the potential for multi-fold scaling of network capacity. The research community has recognized this theoretical potential and developed architectures [1,2] with large numbers of RF chains. Unfortunately, building the hardware with a large number of RF chains is challenging in practice. CSI data transport and computational overhead of MU-MIMO beamforming can also become prohibitive under large network scale. Furthermore, it is difficult to physically append extra RF chains on existing communication equipments to support such large-scale MU-MIMO architectures. In this paper, we present Hekaton, a novel large-scale MU-MIMO framework that combines legacy MU-MIMO beamforming with phased-array antennas. The core of Hekaton is a two-level beamforming architecture. First, the phased-array antennas steer spatial beams toward each downlink user to reduce channel correlation and suppress the cross-talk interference in the RF domain (for beamforming gain), then we adopt legacy digital beamforming to eliminate the interference between downlink data streams (for spatial multiplexing gain). In this way, Hekaton realizes a good fraction of potential large-scale MU-MIMO gains even under the limited RF chain number on existing communication equipments. We evaluate the performance of Hekaton through over-the-air testbed built over the WARPv3 platform and trace-driven emulation. In the evaluations, Hekaton can improve single-cell throughput by up to 2.5X over conventional MU-MIMO with a single antenna per RF chain, while using the same transmit power.