Towards a Scalable 5G RAN Central Unit
Cuidi Wei, A.C. Kak, Nakjung Choi, Timothy Wood · 2023
The radio access network (RAN) connects mobile users to the core network. As the complexity of the RAN has increased with the shift towards 5G and beyond, the need for flexible deployment options has become a key requirement, as evidenced by the rise of disaggregated RAN architectures. Since a disaggregated RAN deployment may utilize a single central unit (CU) for multiple distributed units (DUs), it is critical that the CU not become a performance bottleneck in terms of the aggregate traffic that can be processed. To that end, in this work we focus on improving the scalability of the CU by addressing the performance limitations in a reference open source solution, the OpenAirInterface (OAI) 5G platform. We demonstrate a roughly 5X improvement in scalability by using the Data Plane Development Kit (DPDK) to provide a kernel bypass solution for the OAI 5G CU. While this provides a significant performance boost, we believe that even greater improvements may be possible with further architecture refinements or the use of accelerator platforms like programmable network cards or switches.