Test And Performance Analysis of an O-RAN Network

The evolution of Radio Access Network (RAN) has been driven by the increasing demand for high performance, reliability and cost efficient wireless communication systems, contributing to continuous advancements in telecommunications technologies. Traditional RAN architecture relies completely on proprietary hardware and software, limiting the interoperability and increasing operational costs. Open Radio Access Network (O-RAN) aims to bypass these limitations by introducing an open, modular and virtualized architecture, allowing a multi vendor scenario. Integrating artificial intelligence (AI), cloud native technologies and standardized open interfaces, O-RAN allows also operators to optimize the network while reducing operating costs. In this work the ORAN network is provided by the O-RAN alliance, which disaggregates the traditional base station into three elements: the Centralized Unit (CU), the Distributed Unit (DU) and the Radio Unit (RU). This approach enables virtualization, scalability and the integration of AI optimization. The deployment of O-RAN on a cloud-native platform, OpenShift, enables orchestration, automation, and seamless integration with software-defined networking (SDN). This thesis focuses on implementing and experimentally evaluating a 5G O-RAN network, using OpenAirInterface (OAI) for the gNB and Open5GS for the core network. The experimental goals include assessing device compatibility by testing multiple commercial devices and evaluating connectivity and performance under different conditions. The tests aim to measure key signal indicators such as Reference Signal Received Power (RSRP), Signal-to-Noise Ratio and Block Error Rate (BLER) to analyze the performance of the O-RAN architecture. Further analysis explores network behavior under single and multi device scenarios, with a focus on MAC (Medium Access Control) resource allocation and scheduling performance.