George Ntavazlis Katsaros

The general tendency to deliver Open Radio Access Network (Open-RAN) solutions by means of software-based, or even cloud-native, realizations drives the development community to fully capitalize on software architectures, even for the computationally demanding 5G physical layer (PHY) processing. However, software solutions are typically orders of magnitude less efficient than dedicated hardware in terms of power consumption and processing speed. Consequently, realizing highly-efficient, massive multiple-input multiple-output (mMIMO) solutions in software, while exploiting the wide 5G transmission bandwidths, becomes extremely challenging and requires the massive parallelization of the PHY processing tasks. In this work, for the first time, we show that massively parallel software solutions are capable of meeting the processing requirements of 5G New Radio (NR), still, with a significant increase in the corresponding power consumption. In this context, we quantify this power consumption overhead, both in terms of Watts and carbon emissions, as a function of the concurrently transmitted information streams, of the base-station antennas, and of the utilized bandwidth. We show that the computational power consumption of such PHY processing is no longer negligible and that, for mMIMO solutions supporting a large number of information streams, it can become comparable to the power consumption of the Radio Frequency (RF) chains. Finally, we discuss how a shift towards non-linear PHY processing can significantly boost energy efficiency, and we further highlight the importance of energy-aware digital signal processing design in future PHY processing architectures.

Multi-user multiple-input, multiple-output (MU-MIMO) designs can substantially increase the achievable throughput and connectivity capabilities of wireless systems. However, existing MU-MIMO deployments typically employ linear processing that, despite its practical benefits, can leave capacity and connectivity gains unexploited. On the other hand, traditional non-linear processing solutions (e.g., sphere decoders) promise improved throughput and connectivity capabilities, but can be impractical in terms of processing complexity and latency, and with questionable practical benefits that have not been validated in actual system realizations. At the same time, emerging new Open Radio Access Network (Open-RAN) designs call for physical layer (PHY) processing solutions that are also practical in terms of realization, even when implemented purely on software. This work demonstrates the gains that our highly efficient, massively parallelizable, non-linear processing (MPNL) framework can provide, both in the uplink and downlink, when running in real-time and over-the-air, using our new 5G-New Radio (5G-NR) and Open-RAN compliant, software-based PHY. We showcase that our MPNL framework can provide substantial throughput and connectivity gains, compared to traditional, linear approaches, including increased throughput, the ability to halve the number of base-station antennas without any performance loss compared to linear approaches, as well as the ability to support a much larger number of users than base-station antennas, without the need for any traditional Non-Orthogonal Multiple Access (NOMA) techniques, and with overloading factors that can be up to 300%.