Assessing The TCO For Virtualized RAN And Centralized RAN
This paper is a companion to “How much can operators save with a Cloud RAN? A TCO model for virtualized and distributed RAN,” which compared the TCO for DRAN and vRAN (Cloud RAN).
In this paper, we look at two remote-BBU topologies, Centralized RAN and Virtualized RAN, using that same TCO model. We then compare them to the DRAN TCO. In a vRAN, the vBBUs can be co-located with the vEPC, and operators need only mobile fronthaul to connect the RAN to the EPC. In a Centralized RAN the physical BBUs are remote from the radio but located closer to the RAN. To reach the EPC, a fronthaul link and a BH link are necessary. In a DRAN, the BBU is at the cell site, so a backhaul link connects to the EPC from the RAN.
Mobile networks are becoming virtualized end to end – from the core up to and including the radio access network (RAN). As mobile operators embark on the transition, we see a variety of approaches, which vary in the depth and speed of the virtualization process as well as in the chosen topology. A critical decision is where to place functions physically within a virtualized network.
Which functions should operators keep in a centralized data center and which ones should they move toward the edge? What RAN functionality should operators keep at the cell site, and what at the data center? And with multi-access edge computing (MEC) and other edge computing initiatives, where should the edge equipment go? Depending on capacity, coverage, synchronization and latency requirements, operators can use the core, RAN and edge computing equipment at different network locations to optimize performance and cost-efficiency.