A Cost Assessment Methodology for User-Centric Distributed Massive MIMO Architectures

• Published in: IEEE open journal of the Communications Society Vol. 5; pp. 3517 - 3543
• Main Authors: Fernandes, Andre L. P., Souza, Daynara D., Natalino, Carlos, Tonini, Federico, Cavalcante, Andre M., Monti, Paolo, Costa, Joao C. W. A.
• Format: Journal Article
• Language: English
• Published: New York IEEE 2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: 5G mobile communication; 6G mobile communication; Biological system modeling; Cell-free massive MIMO; Channel estimation; Computational complexity; Computational modeling; Computer architecture; Costs; Distributed processing; Feasibility; feasibility analysis; functional splits; Massive MIMO; Mathematical models; Methodology; Mobile communication systems; network deployment; Optical fiber networks; Power demand; Socioeconomics; Sociotechnical systems; techno-economic assessment; total cost of ownership; User centered design
• ISSN: 2644-125X; 2644-125X
• DOI: 10.1109/OJCOMS.2024.3406374

User-centric (UC) distributed massive multiple-input multiple-output (D-mMIMO), also known as cell-free mMIMO, is a pivotal technology for enabling future mobile communication systems. While UC D-mMIMO intrinsically follows a distributed architecture, its processing can be implemented in a distributed or centralized fashion. This paper proposes a comprehensive cost assessment methodology for UC D-mMIMO, capturing its total cost of ownership and factoring in the deployment configuration, processing implementation, computational demands, and fronthaul signaling. The methodology considers two transmission reception point (TRP) deployment strategies. The first focuses only on supporting user equipment (UE) demands, while the other fulfills these requirements and also actively strives to provide a fairer service among UEs. The proposed methodology is then used to perform a techno-economic assessment of the feasibility of centralized versus distributed processing functional splits while varying key costs and TRP capabilities, like antenna and served UE count. Results suggest that with the TRP deployment that only supports the required UE rate, distributed processing is usually the most feasible option for UE demands of up to 50 Mbps, and centralized processing is more cost-effective in other cases. Additionally, when considering the actively fairer TRP deployment, centralized processing becomes cheaper for any UE demands.