Software-defined networking in cellular radio access networks: potential and challenges

TLDR

Software‑defined networking has improved performance and management in wired networks, raising the question of whether its principles can similarly benefit cellular radio access networks. This article investigates whether SDN concepts can be applied to RANs and explores the potential and challenges of programming the fronthaul network in a cloud‑based RAN. The authors note that control‑data plane separation already exists in RANs through self‑organizing networking, and that cloud RANs enable fine‑timescale optimizations such as coordinated multi‑point transmission by decoupling processing from transmission, while also considering fronthaul programming. They argue that software‑defined fronthaul (SDF) can orchestrate novel applications and should be a key focus for RAN optimization.

Abstract

Software-defined networking has brought both performance and management benefits to wired networks. It is natural to wonder if SDN principles also apply and can deliver similar benefits to RANs. It is this question that we intend to address in this article. We first highlight that the core SDN principle of decoupling control and data planes already exists in RANs in the form of self-organizing networking solutions, which can optimize RAN performance at coarse timescales. In addition to control/data plane separation, we also elaborate how fine timescale optimizations such as coordinated multi-point transmission are made practical with the help of cloud RANs (C-RANS), which offer the notion of processing decoupled from transmission within the data plane. We finally discuss the potential and challenges related to a less explored application of SDN in the RAN, which is programming the fronthaul network in a C-RAN. We argue that this novel notion of software-defined fronthaul, SDF, has the power to orchestrate novel applications and thus should be a key area of focus in RAN optimization.