Cost Analysis of 5G Fronthaul Networks through Functional Splits at the PHY Layer in a Capacity and Cost Limited Scenario

Abstract

The expected growth in carried traffic and the added complexity of different deployment approaches (distributed vs. centralized), showing different requirements, make transport network one of the main design challenges in 5G. One of those challenges is posed by the Centralized Radio Access Network (CRAN) paradigm, whereby different functionalities of the base station are split between a central unit and a remote unit, both connected by a fronthaul/midhaul network. When this centralization includes physical layer functions, stringent capacity and delay constraints are imposed on the fronthaul, thus making its design and deployment more challenging and costly. At this point, the anticipated capacity requirements for fronthaul links are enormous and, as of today, no single technology can support such requirements. Hence, the complex transport network will be heterogeneous in nature. There is consensus in following two approaches to tackle the fronthaul challenge: i) building a heterogeneous network by combining different technologies; and ii) employing different functional splits, which have the potential to reduce the capacity requirements on fronthaul links. Hence, it is important that we exploit different potential technologies and a functional split approach for 5G fronthaul networks design. As our contribution, we show how intelligently selected functional splits at physical layer can be utilized to serve the radio access networks in a capacity-limited scenario. From a different point of view, we also propose maximizing the centralization by means of a heterogeneous combination of functional splits in a budget-limited scenario. Results presented in this paper show that the combination of functional splits has the potential to enable the design of heterogeneous fronthaul networks combining wireless and wired links, and reducing drastically both the required capacity (to 40%) and the total cost of ownership (to 35%).